feat(dashboard): show SKH01 sleeve in 4-sleeve portfolio view

active_sleeves() already feeds the per-sleeve table & combined metrics, so SKH01
appears automatically. Manual touch-ups: title/docstring -> +SKH01; position label
is now sleeve-aware (the None fallback used to mislabel every pos-fn-less sleeve as
XS01's "book 19 gambe" — now XS01/SKH01/VRP01 get correct labels); footer note adds
SKH01 (quasi-orthogonal @25%, FULL Sharpe 1.68->2.13, DD 14->8%, research/forward-monitor).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

.env.ibgw.example

@@ -0,0 +1,7 @@
+# Credenziali IB Gateway PAPER per la ricerca dati (account paper, es. DUQ513966).
+# COPIA questo file in  .env.ibgw  (gitignored) e riempi i valori REALI.
+#   cp .env.ibgw.example .env.ibgw && chmod 600 .env.ibgw && nano .env.ibgw
+# NON committare mai .env.ibgw. Sono credenziali del CONTO PAPER (nessun denaro reale),
+# l'API e' comunque READ_ONLY (solo dati storici, nessun ordine).
+TWS_USERID=il_tuo_username_paper
+TWS_PASSWORD=la_tua_password_paper

.gitignore

@@ -6,6 +6,8 @@ build/
 .venv/
 .env
 !.env.example
+.env.ibgw
+!.env.ibgw.example
 .vscode/
 .idea/
 .DS_Store
@@ -63,3 +65,4 @@ scripts/research/blind/leaderboard.json
 
 # forward-monitor runtime state (regenerable, forward-only)
 data/paper_prevday/
+data/paper_combo/

CLAUDE.md

@@ -51,11 +51,23 @@ Prima ondata di ricerca onesta su BTC/ETH certificati (5 track, harness condivis
   monitor forward. NB il gate concentra XS nei regimi dispersi (2025-26 = hold-out alta-dispersione).
   Ricerca `scripts/portfolio/{xsec_research,xsec_blend,xsec_dispgate}.py`. Diari `2026-06-19-hyperliquid-xsec`
   / `-xsec-blend` / `-xsec-dispgate` / `-xsec-universe-expansion` / `-trend-multiasset`.
-- **PORTAFOGLIO ATTIVO = TP01 (55%) + XS01 (25%) + VRP01 (20%)** (`src/portfolio/sleeves.active_sleeves`):
+- **PORTAFOGLIO ATTIVO = TP01 (41.25%) + XS01 (18.75%) + VRP01 (15%) + SKH01 (25%)** (`src/portfolio/sleeves.active_sleeves`):
   TP01+XS01 combinato **FULL Sharpe 1.55, HOLD-OUT 1.55, DD 4.4%**. Aggiunto **VRP01** (options
   short-vol, sotto): TP01+VRP01 da solo fa FULL Sh 1.30→1.44 / HOLD 0.31→0.40 a peso 20% (3-way da
-  validare locale con dati HL). Report `scripts/portfolio/run_portfolio.py`. Sleeve a date d'inizio
-  diverse → outer-join con pesi rinormalizzati (TP01 da solo 2019-20, VRP dal 2021, blend pieno dal 2024).
+  validare locale con dati HL). **Aggiunto SKH01-V2-DD @25% effettivo (2026-06-23, sotto):** i tre
+  preesistenti scalati nel restante 0.75 (rapporto 55:25:20). Il portafoglio a **4 sleeve** fa
+  **FULL Sharpe 1.68→2.13, HOLD-OUT 1.63→2.30, DD full 14.3%→7.8%** (Skyhook è quasi-ortogonale,
+  corr ~0.09). Report `scripts/portfolio/run_portfolio.py`. Sleeve a date d'inizio
+  diverse → outer-join con pesi rinormalizzati (TP01/SKH01 dal 2019, VRP dal 2021, XS dal 2024).
+- **SKH01-V2-DD "Skyhook" — DIVERSIFICATORE quasi-ortogonale (research)** — `src/strategies/skyhook.SKH01_V2_DD`,
+  sleeve `src/portfolio/sleeves._skyhook_returns`. Sistema dual-TF (segnale 690m / exec 230m) regime
+  (BuzVola/BuzVolume tipo-Chande) AND pattern (Donchian breakout), NON trend-follower, L/S. Vincitrice
+  di 2 onde multi-agente (la 2ª = DD-reduction): exit a **percentuale fissa ASIMMETRICA** (long sl4%/tp10%,
+  short sl2%/tp8% più stretto) → standalone **maxDD BTC 21% / ETH 27% (<30%)**, minFull +0.99, minHold
+  +1.26, causale (0/400), fee-surviving 0.40%RT. Marginal vs TP01 **ADDS** (corr 0.09, has_insample_edge,
+  robust_oos multicut 7/7, is_hedge=False); blend 0.75·TP01+0.25·SKH **hold-out 0.31→1.17**. Verificato
+  leak-free + 2 scettici. **CAVEAT:** equity daily-step (Sharpe lens), ETH DD margine sottile, book 230m
+  (costi ribilanciamento da verificare a deploy) → research win, forward-monitor. Diario `2026-06-23-skyhook.md`.
 - **VRP01 Options Short-Vol — DIVERSIFICATORE da FinanceOld/OptionsAgent** — `src/portfolio/sleeves._vrp_combo_returns`.
   Put credit spread settimanale (vendi put -0.28, compra put -0.10) gated su IV-rank. Idee portate da
   `../FinanceOld/OptionsAgent` (Bear Call Spread + gate d'ingresso). Migliora il lead VRP nudo

docker-compose.yml

@@ -13,3 +13,20 @@ services:
       # token mainnet (sola lettura) per lo "Shadow live": conto/posizioni reali sulla dashboard.
       # Montato a runtime (NON nell'immagine: .env.mainnet e' dockerignored). Solo letture, nessun ordine.
       - ./.env.mainnet:/app/.env.mainnet:ro
+
+  # IB Gateway (PAPER) per la RICERCA DATI Interactive Brokers — replica il setup provato di BuzWay
+  # (scout). IBC fa login automatico headless; nessuna GUI desktop. API READ-ONLY (solo dati storici,
+  # MAI ordini). Bind SOLO su 127.0.0.1 -> non esposto in rete. Credenziali in .env.ibgw (gitignored).
+  # host 4002 -> container 4004 (socat paper), esattamente come nel connect("127.0.0.1", 4002).
+  ib-gateway:
+    image: ghcr.io/gnzsnz/ib-gateway:stable
+    container_name: pythagoras-ibgw
+    restart: unless-stopped
+    env_file: .env.ibgw
+    environment:
+      TRADING_MODE: paper
+      READ_ONLY_API: "yes"          # SOLO dati: nessun ordine possibile via API
+      TWOFA_TIMEOUT_ACTION: restart
+      TIME_ZONE: Europe/Rome
+    ports:
+      - "127.0.0.1:4002:4004"       # gateway paper (socat) raggiungibile solo da localhost dell'host

docs/diary/2026-06-21-prevday-bootstrap.md

@@ -0,0 +1,68 @@
+# PREVDAY block-bootstrap — coda-fortuna vs persistente (blocker #2/#3)
+
+**Data:** 2026-06-21 (chiude la trilogia: fill-haircut → turnover/hedge → bootstrap)
+**Script:** `scripts/research/intraday/prevday_bootstrap.py`
+**Esito:** PREVDAY-full **non** è più coda-fortuna di TP01 e l'edge è **bootstrap-robusto** (full
+99% / hold-out 93% dei resample con uplift>0). MA la gamba short (= tutto il valore) è
+**tail-dipendente** (top-5 giorni = 130% del suo netto). PREVDAY = tail-hedge legittimo dal payoff
+grumoso. Resta forward-monitor.
+
+## Chiarimento di scope
+
+Il "top-5 giorni = 76-83% del PnL" del diario intraday era sulle GAMBE REVERT del combo a 5 segnali
+(vol_event/volume_spike/gap_fill), poi SCARTATE. Il sopravvissuto è PREVDAY (breakout-continuation).
+Qui si testa PREVDAY STESSO — e la sua gamba SHORT, che (prevday_turnover) è l'intero valore di
+portafoglio. Block bootstrap circolare (blocchi 20g, B=3000) per preservare autocorrelazione/regime.
+
+## [A] Concentrazione del PnL nei top-K giorni
+
+| serie | n | totRet | top5 | top10 | top20 | giorni→50% gain |
+|-------|--:|-------:|-----:|------:|------:|----------------:|
+| PREVDAY full | 2869 | +182% | 22% | 36% | 59% | 411 (14.3%) |
+| **PREVDAY short-only** | 2869 | **+28%** | **130%** | 218% | 345% | 312 (10.9%) |
+| PREVDAY long-only | 2869 | +154% | 18% | 30% | 49% | 287 (10.0%) |
+| TP01 (riferimento) | 2657 | +116% | 19% | 33% | 55% | 213 (8.0%) |
+
+- **PREVDAY-full NON è più coda-fortuna di TP01**: top5 22% vs 19%, e per il 50% del guadagno serve
+  *più* tempo (14.3% dei giorni vs 8.0% → più distribuito). Il tail-luck del diario era sulle gambe
+  revert scartate, non su PREVDAY.
+- **Gamba short tail-dipendente:** top5 = **130% del netto** → togliendo i 5 giorni migliori la short
+  va in perdita (gli altri 2864 giorni nettano −8%). Sono i giorni-crash dove la short paga.
+
+## [B] Circular block bootstrap (20g, B=3000)
+
+| campione | PREVDAY Sharpe (mediana [5°,95°], %>0) | blend 80/20 uplift (mediana [5°,95°], %>0, %>+0.10) |
+|----------|----------------------------------------|------------------------------------------------------|
+| full (2018-08→2026-06) | +1.24 [+0.64,+1.80] 100% | +0.28 [+0.09,+0.47] 99% / 93% |
+| hold-out (2025+) | +1.27 [−0.01,+2.46] 95% | +0.53 [−0.05,+1.21] 93% / 88% |
+| short-only hold-out | +1.12 [−0.32,+2.41] 90% | +0.53 [−0.08,+1.31] 92% / 87% |
+
+- **Full sample: edge robustissimo** — 99% dei resample dà uplift>0 (mediana +0.28). Non è "un blocco
+  fortunato".
+- **Hold-out: regge con coda più larga** (5° pctl appena negativo: hold-out corto ~536g + short
+  tail-dipendente), ma 93% dei resample >0, 88% >+0.10.
+
+## Verdetto blocker #2/#3
+
+- **#3 tail-luck — DECLASSATO per PREVDAY-full, CONFERMATO per la gamba short.** La strategia intera
+  non è più concentrata di TP01 (che già deployamo); il motore di valore (la short) sì: vive su <10
+  giorni-crash/anno. Bootstrap-robusto (non un singolo blocco), ma il forward sarà GRUMOSO, non un
+  liscio +0.56/periodo.
+- **#2 null-corr-zero — RIDIMENSIONATO.** L'uplift è genuinamente positivo (93-99% dei resample), non
+  rumore; il punto era di *efficienza relativa* (rende meno di un ipotetico asset perfettamente
+  scorrelato), non di esistenza dell'edge.
+
+## Sintesi della trilogia (fill-haircut + turnover/hedge + bootstrap)
+
+PREVDAY, dopo tre attacchi avversariali:
+1. **Eseguibile alla taglia reale** ($600): haircut di fill +0.01 (blocker #4 smontato).
+2. **Già a turnover efficiente**: ridurlo erode l'edge; nessuna ottimizzazione (config congelata).
+3. **È un HEDGE, non alpha**: tutto il valore è la gamba short → tail-hedge di regime-down, additivo
+   alla flat-stance di TP01 (blocker #1 inchiodato).
+4. **Edge bootstrap-robusto** ma **payoff grumoso** (il valore è in pochi giorni-crash) (blocker #3
+   declassato sul full, confermato sulla short; #2 ridimensionato).
+
+→ **Candidato tail-hedge legittimo**, non sleeve-alpha. Resta in FORWARD-MONITOR: la domanda forward
+non è più "è eseguibile / è overfit", ma **"la gamba short continua a pagare nei prossimi crash fuori
+da 2022 e 2025-26?"**. Se sì → si valuta come overlay di tail-hedge (peso piccolo, atteso payoff
+lumpy); se no → era beta-corto del regime down 2025-26.

docs/diary/2026-06-21-prevday-fill-haircut.md

@@ -0,0 +1,67 @@
+# PREVDAY fill-haircut a basso capitale — il blocker d'esecuzione è BENIGNO (1/4 smontato)
+
+**Data:** 2026-06-21 (follow-up di `2026-06-21-intraday-microstructure.md`)
+**Script:** `scripts/research/intraday/fill_haircut.py`
+**Esito:** l'haircut del fill reale a $600 è **+0.01 Sharpe** (trascurabile). Lo scettico
+d'esecuzione (blocker #4) è **benigno**. Gli altri 3 blocker (hedge / null-corr-zero / tail-luck)
+restano → PREVDAY resta in **forward-monitor, non deploy**.
+
+## Domanda
+
+Lo scettico d'esecuzione dell'onda intraday aveva segnalato: il vol-target di PREVDAY fa ~8500
+ribilanciamenti/anno per gamba, 97-98% < $1 di nozionale a $600; a quel capitale (min_order $5) NON
+puoi piazzarli, quindi il libro MODELED (ribilanciamento continuo, frictionless) è una finzione e lo
+Sharpe modellato è gonfiato. Il forward-monitor traccia MODELED-$2000 vs REAL-$600 per misurarlo nei
+mesi a venire — qui lo stimiamo SUBITO su tutto lo storico, replicando la STESSA logica dei due libri
+di `paper_prevday.py` ma sull'intero path 1h (2019-03 → 2026-06, 63.732 barre).
+
+Due libri identici tranne il fill:
+- **MODELED**: ribilancia ad ogni barra (fee proporzionale su ogni |Δ|).
+- **REAL-$C**: salta i ribilanciamenti con nozionale `|Δpos|·leg_cap < $5` (posizione stale →
+  tracking error, ma niente fee sui trade infinitesimi). Sweep C ∈ {600, 2000, 20000}.
+
+## Risultati
+
+| libro | FULL Sh | HOLD Sh | CAGR | DD | rebal/yr | skip% | fee-drag/yr |
+|-------|---------|---------|------|----|---------:|------:|------------:|
+| MODELED ($∞) | +1.23 | +1.27 | +24.3% | −27% | 17.484 | 0.0% | 2.49% |
+| REAL-$20k | +1.23 | +1.27 | +24.4% | −27% | 3.747 | 78.6% | 2.47% |
+| REAL-$2000 | +1.23 | +1.27 | +24.4% | −27% | 677 | 96.1% | 2.42% |
+| REAL-$600 | +1.22 | +1.26 | +24.2% | −27% | 277 | 98.4% | 2.39% |
+
+**HAIRCUT $600 (MODELED − REAL): FULL Sharpe +0.01, HOLD-OUT +0.01.**
+
+Domanda-soldi (l'uplift del blend regge col fill reale?):
+
+| PV | w | FULL (uplift) | HOLD (uplift) |
+|----|---|---------------|---------------|
+| MODELED | 20% | 1.58 (+0.28) | 0.86 (+0.56) |
+| MODELED | 30% | 1.65 (+0.36) | 1.08 (+0.78) |
+| **REAL-$600** | 20% | 1.58 (+0.28) | 0.86 (**+0.55**) |
+| **REAL-$600** | 30% | 1.65 (+0.35) | 1.08 (**+0.77**) |
+
+(TP01 solo: FULL +1.30, HOLD +0.31.) L'uplift hold-out sopravvive **quasi intatto**.
+
+## Lettura
+
+Saltare il **98.4%** dei micro-ribilanciamenti a $600 non costa quasi nulla perché quei trade sono
+*individualmente infinitesimi*: sia la fee risparmiata sia il tracking-error introdotto sono
+trascurabili. Il PnL è dominato dai ~50 flip di direzione/anno + la deriva lenta del vol-target, che
+il libro $600 cattura comunque sui movimenti grandi (la fee-drag passa solo da 2.49% a 2.39%). La
+"finzione della fee sub-dollaro" è quindi **benigna**: non gonfia lo Sharpe modellato (MODELED e
+REAL-$600 coincidono a ±0.01). NB: lo Sharpe **non si degrada** scendendo di capitale → l'edge
+modellato di PREVDAY è eseguibile alla taglia reale; il blocker era altrove.
+
+## Conseguenza sul verdetto
+
+Dei 4 blocker che tenevano PREVDAY fuori dal deploy, il **#4 (fill a basso capitale) è SMONTATO**.
+Restano in piedi i 3 strutturali (dall'onda intraday, non rivalutati qui):
+1. **hedge-shaped** — l'uplift viene dai regimi TP01-down (uplift +0.79 TP01-down vs +0.20 TP01-up);
+2. **fallisce il null a corr-zero** — uplift pre-2025 al 20-24° pctl del null di un asset random
+   scorrelato (aggiunge MENO del rumore);
+3. **tail-luck** — top-5 giorni = 76-83% del PnL delle gambe revert, <10 eventi/anno.
+
+PREVDAY resta il lead **più solido sull'esecuzione** di tutta la ricerca post-reset (il dubbio più
+"fisico" è caduto), ma **forward-monitor, non deploy**, finché il track record forward non scioglie
+hedge/coda/null. Lezione harness: `eval_weights_smallcap` (il gate min-order) va sempre eseguito
+PRIMA di scartare un lead per "fill irreale" — qui avrebbe evitato di sopravvalutare il blocker #4.

docs/diary/2026-06-21-prevday-overlay-portfolio.md

@@ -0,0 +1,68 @@
+# PREVDAY come overlay di tail-hedge sul portafoglio — simulazione d'impatto (NON deploy)
+
+**Data:** 2026-06-21 (segue la trilogia fill-haircut / turnover-hedge / bootstrap)
+**Script:** `scripts/portfolio/prevday_overlay.py`
+**Esito:** a peso 10%, PREVDAY taglia il maxDD FULL del portafoglio **14.3% → 9.9% (−31%)** e alza
+l'hold-out Sharpe **1.66 → 1.97 (+0.31)**. 10% è vicino all'ottimo di DD. MA è tutto IN-SAMPLE: il
+prize si materializza solo SE l'edge di PREVDAY persiste forward. PREVDAY resta FORWARD-MONITOR.
+
+## Setup
+
+Simulazione che NON tocca il registry di produzione: prende il portafoglio attivo (TP01 55% + XS01
+25% + VRP01 20%), riscala i tre sleeve a (1−W) mantenendone le proporzioni, e aggiunge PREVDAY a
+peso W. Sweep W ∈ {0,5,10,15,20%}. PREVDAY = libro 1h breakout-continuation, parametri congelati,
+50/50 BTC+ETH, fee 0.10% RT. Outer-join del portafoglio: PREVDAY dal 2018, VRP01 dal 2021, XS01 dal
+2024 → nel 2019-20 PREVDAY pesa di fatto >W (solo TP01 accanto); nell'hold-out 2025+ (tutti e 4
+attivi) pesa esattamente ~W → **l'HOLD-OUT è il confronto pulito a "10%"**.
+
+## Sweep peso overlay
+
+| peso PREVDAY | FULL Sharpe | FULL DD | HOLD Sharpe | HOLD DD | HOLD ret |
+|--------------|------------:|--------:|------------:|--------:|---------:|
+| BASELINE (55/25/20) | 1.68 | 14.3% | 1.66 | 3.4% | +16.7% |
+| 5% | 1.80 | 11.1% | 1.83 | 3.3% | +17.8% |
+| **10%** | **1.88** | **9.9%** | **1.97** | 3.3% | **+19.0%** |
+| 15% | 1.93 | 10.3% | 2.06 | 3.3% | +20.2% |
+| 20% | 1.95 | 10.6% | 2.09 | 3.4% | +21.4% |
+
+(PREVDAY standalone: FULL Sh 1.23 / DD 26.7%; HOLD Sh 1.28 / DD 10.8%.)
+
+## Lettura a 10%
+
+- **FULL: Sharpe +0.20 (1.68→1.88), maxDD 14.3%→9.9% (−4.4pp ≈ −31%).** Comportamento da tail-hedge:
+  la gamba short ammortizza i crash storici (2019-21, 2022).
+- **HOLD-OUT: Sharpe +0.31 (1.66→1.97), ret +16.7%→+19.0%, DD 3.4%→3.3% (già bassissimo).** Nel
+  regime recente il beneficio è rendimento/Sharpe, non taglio DD.
+- **10% ≈ ottimo di DD.** Oltre, lo Sharpe sale ancora (1.93→1.95) ma il maxDD FULL smette di
+  scendere (10.3→10.6%): stai solo aggiungendo rischio direzionale short. Argomento per ~10% in
+  chiave hedge (massimizza il taglio-coda per unità di rischio aggiunto).
+
+## Per anno (baseline → overlay 10%)
+
+| anno | ret | DD |
+|------|-----|----|
+| 2019 | +11.3 → +15.2 | 10.3 → 8.2 |
+| 2020 | +51.1 → +53.1 | 8.4 → 6.3 |
+| 2021 | +32.5 → +28.3 | 5.2 → 4.3 |
+| 2022 | −3.0 → −1.6 | 3.7 → 3.0 |
+| 2023 | +11.2 → +11.4 | 9.2 → 9.9 |
+| 2024 | +24.4 → +25.7 | 3.9 → 3.4 |
+| 2025 | +12.0 → +12.0 | 3.4 → 3.3 |
+| 2026 | +4.2 → +6.2 | 2.6 → 2.2 |
+
+Migliora o pareggia quasi ovunque; costa solo nel toro 2021 (premio d'assicurazione atteso per un
+hedge) e leggermente sul DD 2023; paga nel bear 2022 e nel 2026.
+
+## Verdetto
+
+L'overlay 10% è **attraente in simulazione** — taglia il drawdown FULL di ~31% e alza l'hold-out
+Sharpe +0.31, con 10% vicino all'ottimo di DD. Ma:
+1. **È in-sample.** I guadagni assumono che l'edge di PREVDAY persista — il forward-monitor esiste
+   proprio per verificarlo. Questa simulazione quantifica il PRIZE, non lo prova.
+2. **Outer-join:** il taglio-DD storico è gonfiato dal peso effettivo >10% nel 2019-20; il read
+   pulito a 10% è l'hold-out (prize = Sharpe +0.31).
+3. Incidentale: il 3-way TP01+XS01+VRP01 baseline qui fa FULL 1.68 / HOLD 1.66.
+
+**Azione: nessuna.** PREVDAY resta FORWARD-MONITOR (registry di produzione invariato). Quando il
+track record forward avrà ~2-3 mesi, ri-valutare l'overlay 10% con la stessa metrica (taglio-DD +
+hold-out Sharpe) su dati VERAMENTE fuori campione. Lo script è il riferimento per quel confronto.

docs/diary/2026-06-21-prevday-turnover-and-hedge.md

@@ -0,0 +1,64 @@
+# PREVDAY — la fee viene dai FLIP (no free lunch sul turnover) + è un HEDGE, non alpha
+
+**Data:** 2026-06-21 (follow-up di `2026-06-21-prevday-fill-haircut.md`)
+**Script:** `scripts/research/intraday/prevday_turnover.py`
+**Esito:** (1) ridurre il turnover di PREVDAY erode l'edge — la config congelata è già efficiente.
+(2) Il test long-only inchioda il blocker #1: **tutto il valore di portafoglio è la gamba SHORT** →
+PREVDAY è un **hedge di regime-down**, non alpha. Resta forward-monitor.
+
+## Premessa (da fill_haircut)
+
+Il libro REAL-$600 salta il 98.4% dei ribilanciamenti del vol-target e la fee-drag scende solo
+2.49% → 2.39%/anno. Quindi la fee (~2.6%/anno) NON viene dal churn sub-dollaro ma dai **~70 flip di
+direzione/anno**. Un deadband d'esecuzione è inutile; l'unica leva è ridurre i flip a LIVELLO DI
+SEGNALE. Qui sweep delle leve (buffer, anchor, min-hold) + long-only vs long-short. Libro MODELED
+(l'haircut di fill è +0.01, irrilevante). Metrica che conta = **uplift hold-out del blend 80/20**.
+
+## (1) Turnover-reduction — no free lunch
+
+| config | flip/yr | fee/yr | FULL Sh | HOLD Sh | DD | corrTP | blend HOLD upl |
+|--------|--------:|-------:|--------:|--------:|---:|-------:|---------------:|
+| **BASE** (anchor=1, k=0.30, LS) | 70 | 2.59% | +1.23 | +1.27 | −27% | +0.15 | **+0.56** |
+| k=0.50 | 48 | 1.86% | +1.23 | +0.99 | −15% | +0.20 | +0.40 |
+| k=0.75 | 32 | 1.31% | +1.06 | +0.13 | −16% | +0.27 | +0.00 |
+| k=1.00 | 23 | 1.01% | +0.88 | +0.72 | −22% | +0.36 | +0.22 |
+| anchor=2 | 39 | 1.55% | +0.89 | +0.54 | −22% | +0.25 | +0.20 |
+| anchor=3 | 27 | 1.14% | +0.67 | −0.18 | −22% | +0.29 | −0.12 |
+| anchor=5 | 15 | 0.75% | +1.15 | +0.70 | −19% | +0.41 | +0.25 |
+| min_hold=24h | 70 | 2.59% | +1.22 | +1.37 | −32% | +0.15 | **+0.60** |
+| min_hold=72h | 65 | 2.39% | +0.86 | +0.67 | −33% | +0.12 | +0.27 |
+| combo-LT (k.75+anc2+24h) | 16 | 0.79% | +0.79 | +0.69 | −20% | +0.34 | +0.24 |
+
+- **Allargare buffer/anchor taglia fee e turnover ma l'uplift cala monotonicamente** (k: 0.56→0.40→
+  0.00). Anchor multi-giorno tutto peggio → conferma il "anchor=1 only" del diario. I flip SONO
+  l'edge: meno flip = meno edge.
+- **min_hold=24h** è l'unico ritocco "quasi gratis" (uplift +0.56→+0.60 a parità di fee) ma
+  **peggiora il DD −27%→−32%** → non vale cambiare una strategia congelata in forward-monitor.
+- **Verdetto: la config base è già sulla frontiera efficiente turnover↔edge. Si lascia congelata.**
+
+## (2) Long-only vs long-short — il blocker #1 inchiodato
+
+| | FULL Sh | HOLD Sh | corrTP | blend HOLD upl | fee/yr |
+|--|--------:|--------:|-------:|---------------:|-------:|
+| **long-only** (no short) | **+1.55** | +0.52 | **+0.64** | **+0.09** | 1.30% |
+| long-short (BASE) | +1.23 | +1.27 | +0.15 | +0.56 | 2.59% |
+
+La versione **long-only ha Sharpe standalone più ALTO** (1.55 vs 1.23) ma è **correlata +0.64 a TP01
+e non aggiunge quasi nulla al blend** (+0.09). **Tutto il valore di portafoglio viene dalla gamba
+SHORT:** la short *abbassa* lo Sharpe standalone (shortare crypto nel toro 2019-24 perde) ma fornisce
+**tutta** la decorrelazione (corrTP 0.64→0.15) e l'uplift hold-out (0.09→0.56).
+
+→ **PREVDAY non è alpha: è strutturalmente un HEDGE di crash/regime-down.** Costa nel toro, paga
+nell'orso (2022, 2025-26 down/chop). È additivo a TP01, che va *flat* nel risk-off ma non *short*.
+Questo conferma e affina il blocker #1 dell'onda intraday ("l'uplift viene dai regimi TP01-down"):
+non è solo conditional sui regimi down, è **interamente la gamba short** = una scommessa direzionale
+che i ribassi continuino.
+
+## Conseguenza sul verdetto
+
+- Niente da ottimizzare: la config congelata è già efficiente; nessun cambio.
+- **Riframing utile:** se PREVDAY un giorno avrà un ruolo, è come **overlay di tail-hedge** (non
+  sleeve-alpha), additivo alla difensività di TP01. Ma resta soggetto agli altri due blocker
+  (fallisce il null a corr-zero; tail-luck: top-5 giorni = 76-83% del PnL delle gambe revert).
+- **Forward-monitor invariato.** Il test forward decisivo: la gamba short continua a pagare fuori da
+  2022 e 2025-26? Se sì → candidato tail-hedge; se no → era regime-luck.

docs/diary/2026-06-22-crypto-lead-50agent-sweep.md

@@ -0,0 +1,62 @@
+# 2026-06-22 — Sweep 65-agenti: crypto -> mercati IB (mercati × timing × anni)
+
+## Obiettivo (goal utente)
+Usare >=50 agenti per prendere l'anticipazione crypto->equity e trovare la MIGLIORE soluzione,
+provando diversi mercati e timing, su piu' anni.
+
+## Setup
+- **Dati**: universo IB esteso a **26 ETF certificati** (azioni US/settori/intl/bond/credito/oro/
+  commodity/REIT), cache su disco (`fetch_ib_equities.py` + BROAD2). Crypto BTC/ETH 1h (Deribit).
+- **Harness onesto** (`crypto_lead_harness.py`): per ogni sessione equity, lead = crypto nella
+  finestra equity-CHIUSO [P 21:00 -> D 13:00 UTC] (overnight; il weekend e' il caso lungo). Predice
+  gap/intraday/full. Metriche: corr, **t incrementale vs sessione equity precedente**, Sharpe
+  eseguibile (sign(lead)*predict, net costi) FULL/IS/OOS, **hit per-anno**.
+- **Workflow** (`wf_crypto_lead.js`): grid 416 config (2 lead × 26 mercati × 2 giorni × 2 predict ×
+  2 finestre). **52 agenti sweep** -> **12 agenti verifica avversariale** (stress 10bps + OOS 2024+ +
+  multi-anno) -> **1 sintesi**. Totale **65 agenti**, 1.7M token.
+
+## Risultato
+
+### Fenomeno fortissimo: crypto overnight -> GAP di apertura equity
+Cluster coerente in cima, TUTTI predict=gap/overnight, su ogni target risk-on:
+| lead->target | t-incr | Sh OOS@4bps | @10bps | OOS-recente | anni+ |
+|---|---|---|---|---|---|
+| ETH->IWM gap | 17.1 | 2.49 | 1.96 | 2.41 | 7/8 |
+| ETH->QQQ gap | 17.9 | 2.36 | 1.83 | 2.31 | 7/8 |
+| ETH->XLK gap | 17.4 | 2.40 | 1.93 | 2.30 | 7/8 |
+| **BTC->QQQ gap** | 15.0 | 2.31 | 1.78 | 2.16 | **9/9** |
+| BTC->SPY gap | 14.4 | 2.14 | 1.69(lf) | 2.03 | 9/9 |
+
+Statisticamente schiacciante (t 14-18, sopra Bonferroni su 416 test), regge stress costi e OOS
+recente, **positivo 8-9 anni su 8-9**.
+
+### Ma DUE killer (i verificatori avversariali concordi)
+1. **NON tradabile via ETF**: il gap e' gia' prezzato all'open dell'ETF -> serve un FUTURE indice
+   tenuto overnight (MNQ/MES/M2K). A $0.5-2k il margin overnight di anche un micro consuma il
+   capitale e rischia la liquidazione su un gap avverso -> **fuori portata per costruzione**.
+2. **E' RISK-BETA, non alpha**: la finestra-lead crypto e' quasi CONTEMPORANEA al gap (stesso shock
+   macro overnight, equity chiuso). t enorme = co-movimento risk-on/off, non ETH/BTC che *anticipa*.
+   Firma: la forza e' negli anni alta-vol (2022 hit 0.71-0.75), piatta negli anni calmi (2019/21/23).
+   corr ~0.37 -> beta implicito ~37%, alpha residuo piccolo.
+
+### L'unico tradabile via ETF e' troppo debole
+ETH->XLE intraday 6h (compri XLE al day-open, chiudi +6h): Sh OOS 0.48@4bps **-> 0.15@10bps** (annuo
+4.1%->1.0%), t-incr 2.38 **sotto Bonferroni** (~3.5 su 416 test). Edge netto onesto ~ZERO.
+
+## Verdetto (sintesi multi-agente)
+**Nessun edge proprietario deployabile a basso capitale.** Il fenomeno crypto->equity-overnight e'
+statisticamente reale e robustissimo su 9 anni, ma e' (a) risk-beta condiviso, non anticipazione
+sfruttabile, e (b) catturabile solo con futures overnight, fuori dal nostro capitale. L'unica
+versione ETF-eseguibile e' dentro il rumore da multiple-testing. Coerente col soffitto del progetto:
+"niente di nuovo regge" alla verifica onesta.
+
+**Migliore soluzione (come FENOMENO da forward-monitor, non deploy):** BTC->QQQ gap overnight — la
+storia piu' lunga (9/9 anni), lead noto prima dell'open. Da monitorare; deployabile solo con capitale
+~>$20-30k su micro-futures indice e con i costi notturni modellati.
+
+## Lezione
+Anche con 65 agenti e una ricerca esaustiva su mercati/timing/anni, la disciplina onesta
+(tradabilita' al capitale reale + multiple-testing + beta-vs-alpha) riduce un "Sharpe 2.5 su 9 anni"
+a un non-edge per noi. Il valore della ricerca: aver QUANTIFICATO e CLASSIFICATO il fenomeno
+(risk-beta overnight) invece di scambiarlo per alpha.
+Artefatti: `crypto_lead_harness.py`, `wf_crypto_lead.js`.

docs/diary/2026-06-22-crypto-macro-leadlag.md

@@ -0,0 +1,53 @@
+# 2026-06-22 — Crypto × mercati IB: correlazioni e ANTICIPAZIONI (lead-lag)
+
+## Obiettivo
+Cercare correlazioni e soprattutto ANTICIPAZIONI tra crypto e mercati IB: un mercato fa capire
+l'andamento dell'altro? Dati: cache su disco (BTC/ETH Deribit 1h->1d UTC; ETF eq_* con OPEN). Nessun
+IB online. Disciplina: attenzione ai tranelli di timing daily (crypto chiude 00:00 UTC, US equity
+21:00 -> lag-0 contaminato), test del segno + OOS + multiple-testing.
+Script: `crypto_macro_leadlag.py`, `crypto_weekend_signal.py`.
+
+## (1) Correlazione contemporanea
+Crypto = asset RISK-ON: corr BTC/ETH ~ **+0.32/0.37** con SPY/QQQ/IWM, **+0.25/0.28** con HYG
+(credito), **+0.13** GLD, **~-0.02** TLT (bond). Atteso.
+
+## (2) Lead-lag giornaliero: NIENTE
+corr(BTC_{t-k}, ETF_t) ha picco a **k=0** (~0.32) e crolla a rumore (±0.05) per |k|>=1. Al daily
+**nessuno anticipa l'altro** (ne' crypto->equity ne' viceversa). Honest negative.
+
+## (3) EFFETTO WEEKEND: anticipazione PULITA, significativa, OOS-robusta
+La crypto si muove Sab+Dom (azionario chiuso) -> quel movimento e' info PRIOR al lunedi'.
+- **Anticipa il GAP del lunedi'**: corr +0.22/+0.24 (SPY/QQQ/IWM/HYG), hit 59-62%, e **si RAFFORZA
+  OOS (2022+): +0.30/+0.36**. Coerente su 4 ETF (non cherry-pick).
+- Intraday del lunedi' (open->close) piu' debole ma presente (corr 0.10-0.15, OOS 0.18-0.22).
+
+### Validazione avversariale
+- **(A) INCREMENTALE vs venerdi'**: regressione `Mon ~ weekend_crypto + friday_eq`. Coeff weekend
+  crypto significativo ovunque (QQQ gap **t=+4.7**, intr t=+2.9; SPY +4.4/+2.0; IWM +4.7/+2.7);
+  friday_eq NON significativo. -> e' info CRYPTO-SPECIFICA del weekend, non momentum equity.
+- **(B) TRADABILE** (osservo weekend crypto Dom 24:00, entro Monday OPEN, esco CLOSE, net 4bps):
+  | ETF | hit | Sharpe FULL / IS / OOS22+ | long-flat OOS | ann |
+  |---|---|---|---|---|
+  | QQQ | 60% | 1.46 / 1.61 / 1.33 | **1.91** | ~+9-11% |
+  | SPY | 60% | 0.96 / 0.91 / 1.01 | 1.70 | ~+5% |
+  | IWM | 56% | 0.89 / 0.73 / 1.04 | 1.07 | ~+6% |
+
+## Verdetto
+**Trovata UNA anticipazione reale**: il weekend crypto anticipa il lunedi' azionario (massimo su QQQ,
+risk-on/tech). Significativa (t>4 sul gap), incrementale al venerdi', tradabile net costi, **regge e
+si rafforza OOS**, coerente su piu' ETF. Meccanismo economico sensato: crypto = proxy 24/7 del
+risk-sentiment; nel weekend l'equity e' chiuso e lunedi' "recupera" la direzione crypto.
+
+### Caveat onesti
+- **Capacita' bassa**: ~52 lunedi'/anno, intraday -> ~+9%/yr sul capitale impiegato il lunedi', non
+  una macchina da compounding. E' un segnale TATTICO, non un cornerstone.
+- Il GAP (t=4.7) e' piu' forte dell'intraday (t=2.9) ma per catturarlo serve entrare PRIMA del Monday
+  open -> via **futures indice IB (MNQ/MES, aperti la domenica sera)**: enhancement eseguibile da
+  validare (cattura gap+sessione).
+- Multiple-testing 3 ETF x 2 target: ma TUTTI significativi e coerenti -> effetto ampio, non fortuna.
+- Niente IB online qui (cache); per il deploy servirebbe il feed crypto live la domenica sera.
+
+## Prossimo (se si procede)
+Validare la variante FUTURES (MNQ domenica sera -> cattura il gap del lunedi') e il sizing a basso
+capitale; eventualmente paper-trade. E' la prima ANTICIPAZIONE cross-mercato trovata: crypto come
+lead di sentiment sul lunedi' equity.

docs/diary/2026-06-22-deployable-combo-tp01-gtaa.md

@@ -0,0 +1,40 @@
+# 2026-06-22 — Combo DEPLOYABLE: TP01 (Deribit) + GTAA (IB)
+
+## Perche'
+Il combo crypto-pieno (TP01+XS01+VRP01)+GTAA diversificava (Sharpe 1.81), ma XS01/VRP01 sono
+STAT-MODE (non eseguibili a $600). Validazione del combo ONESTO/eseguibile: solo le gambe deployable
+a basso capitale — TP01 (gia' armato live su Deribit) + GTAA vt12 (eseguibile su IB, frazioni,
+switch mensile). `eq_tp01_gtaa_combo.py`. TP01 compoundato sul calendario giorni-di-borsa.
+
+## Risultati (finestra comune 2019-03 .. 2026-06, ~7y)
+| | Sharpe | CAGR | volAnn | maxDD |
+|---|---|---|---|---|
+| TP01 (crypto, Deribit) | 1.25 | 16.4% | 12.9% | 14% |
+| GTAA vt12 (equity, IB) | 1.12 | 6.0% | 5.3% | 8% |
+| **blend 50/50** | **1.48** | 11.3% | 7.5% | **8%** |
+| blend 40/60 (best cap-mix) | 1.52 | 10.2% | 6.6% | 8% |
+| risk-parity (29c/71e) | 1.52 | 9.1% | 5.9% | 8% |
+
+**corr TP01<->GTAA = +0.21**. Il blend (1.48-1.52) batte entrambe le gambe (best solo 1.25),
+maxDD 14%->8%. **DIVERSIFICA anche da deployable.**
+
+## Caveat onesti
+- Per-anno 50/50: 2019 2.11, 2020 2.51, 2021 1.66, **2022 -2.64**, 2023 1.40, 2024 1.73, 2025 0.98,
+  2026 0.94. Anni boom iniziali gonfiano lo Sharpe assoluto; il **2022 e' negativo** (trend whipsaw
+  su entrambe le gambe nel bear). Recenti ~0.95. -> il numero robusto e' il GUADAGNO da
+  diversificazione (+0.27 Sharpe del blend vs solo), non il livello assoluto.
+- **Costo deployability**: crypto-pieno+GTAA = 1.81 vs deployable = 1.48. I ~0.33 di Sharpe persi sono
+  cio' che XS01/VRP01 darebbero se eseguibili (servirebbe ~20k).
+- **Cross-venue** Deribit+IB: due conti, capitale split. Entrambe switch mensile/basso turnover,
+  frazionabili a $0.5-2k.
+
+## Verdetto
+Combo deployable VALIDO: due trend difensivi scorrelati (corr 0.21) su mercati diversi -> Sharpe
+~1.5 / maxDD ~8%, meglio di ciascuna gamba. E' il candidato concreto per un paper-trade cross-venue.
+NON risolve EUR50/g (resta capitale), ma e' la migliore configurazione rischio-aggiustata
+EFFETTIVAMENTE eseguibile trovata finora. Lezione cross-mercato confermata: il salto di qualita' non
+e' un nuovo alpha ma un SECONDO mercato scorrelato.
+
+## Prossimo (se si procede)
+Paper-trade della gamba GTAA su IB (forward-only, come paper_trend per TP01), per validare
+l'esecuzione cross-venue a rischio zero prima di qualunque capitale reale.

docs/diary/2026-06-22-equities-gtaa-and-crypto-combo.md

@@ -0,0 +1,57 @@
+# 2026-06-22 — EQ-GTAA01 (trend multi-asset) + COMBO cross-mercato equity×crypto
+
+## (1) EQ-GTAA01 — trend difensivo multi-asset (GTAA)
+
+EQ-TREND01 (trend su SPY) taglia il DD. Diversificare le SORGENTI di trend (azioni US/tech/small +
+bond + oro + high-yield) migliora il rischio-aggiustato. `eq_gtaa_trend.py`: ogni asset gestito col
+proprio trend long-flat (TSMOM multi-orizzonte), equal-weight tra gli asset disponibili (outer-join,
+cash dove off/assente). Universo SPY/QQQ/IWM/TLT/GLD/HYG. Causale, netto fee, OOS 2015+.
+
+| strategia | CAGR | Sharpe (pre15/OOS) | maxDD | corr SPY |
+|---|---|---|---|---|
+| SPY buy&hold | 9.7% | 0.58 (0.45/0.82) | 55% | — |
+| EW statico (no trend) | 9.4% | 0.59 | 62% | 0.89 |
+| SPY-trend mono | 5.5% | 0.56 (/0.78) | 30% | 0.72 |
+| **GTAA lf vt12%** | 3.8% | **0.64** (0.53/**0.89**) | **15%** | **0.64** |
+| GTAA vt12 (6-asset, 2016+) | 5.6% | **1.08** | **8%** | 0.60 |
+
+DD nei bear (GTAA vs SPY): dot-com 32%/49% · GFC **14%/55%** · COVID **10%/34%** · 2022 11%/24%.
+Marginale vs SPY: corr 0.64; 50/50 uplift +0.041 FULL / **+0.086 OOS** (meglio del mono-SPY). Plateau
+stabile (Sh 0.55-0.61, DD 25-35%). **Migliore sleeve equity**: Sharpe più alto, maxDD bassissimo
+(8-15%), corr SPY più bassa (0.64) = diversificatore migliore. Tradeoff: CAGR molto più basso
+(fortemente difensiva). Caveat: la finestra 6-asset (Sh 1.08) è tutta OOS ma un solo regime (toro).
+
+## (2) COMBO cross-mercato — equity-trend × crypto
+
+La via che alza il Sharpe COMPLESSIVO senza nuovo alpha: combinare due book scorrelati.
+`eq_crypto_combo.py`: crypto = portafoglio attivo TP01+XS01+VRP01 (`StrategyPortfolio.combined_daily`,
+rinormalizzato); equity = GTAA lf vt12%. Crypto compoundato sul calendario giorni-di-borsa (cattura
+i weekend). Finestra comune = era crypto (2019-03 .. 2026-06, 1827 giorni di borsa).
+
+| | Sharpe | CAGR | volAnn | maxDD |
+|---|---|---|---|---|
+| crypto TP01+XS01+VRP01 | 1.60 | 18.7% | 11.1% | 14% |
+| equity GTAA vt12 | 1.12 | 6.0% | 5.3% | 8% |
+| **blend 50/50** | **1.81** | 12.4% | 6.6% | **7%** |
+| risk-parity (32c/68e) | 1.78 | 10.1% | 5.5% | 8% |
+
+**Correlazione crypto↔equity = +0.167** (bassissima). Il blend 50/50 fa **Sharpe 1.81 > di ciascuno**
+(crypto 1.60, equity 1.12), **maxDD dimezzato 14%→7%**. VERDETTO: DIVERSIFICA (blend > miglior solo
+di +0.21 Sharpe). È il guadagno STRUTTURALE: due fonti di rischio scorrelate alzano il Sharpe
+complessivo senza cercare un nuovo edge.
+
+### Caveat onesti
+- **Finestra crypto corta (~7y) e favorevole**: il crypto Sharpe 1.60 e' alto (regime toro + XS01/VRP01
+  STAT-MODE a storia corta). Gli SHARPE ASSOLUTI sono ottimistici. Ma il PUNTO della diversificazione
+  (corr 0.17, blend > solo, DD dimezzato) è robusto al livello assoluto.
+- **Cross-venue**: crypto su Deribit, equity su IB → due conti, due percorsi d'esecuzione. A $0.5-2k
+  totali, ogni sleeve è minuscola. La parte equity (GTAA) e la TP01 sono entrambe eseguibili a basso
+  capitale; XS01/VRP01 restano STAT-MODE (il blend "reale" deployable è ~TP01 + GTAA).
+
+## Lettura strategica
+Il fronte equity da' due cose: (a) una sleeve difensiva robusta (GTAA, maxDD ~10%), (b) — piu'
+importante — un DIVERSIFICATORE quasi-scorrelato al crypto che alza il Sharpe del portafoglio
+complessivo da ~1.6 a ~1.8. Non risolve €50/g (resta capitale), ma e' il primo miglioramento
+STRUTTURALE del rischio-aggiustato complessivo trovato in tutta la ricerca post-reset, ed e' del tipo
+giusto (diversificazione vera, non alpha fittizio). Prossimo: validare il combo deployable TP01+GTAA
+(solo le due gambe eseguibili), e valutare l'operativita' cross-venue.

docs/diary/2026-06-22-equities-sector-momentum.md

@@ -0,0 +1,45 @@
+# 2026-06-22 — Fronte EQUITY aperto + EQ-MOM01 (momentum settoriale): NON batte SPY
+
+## Apertura fronte (branch research/equities-ib)
+
+Le 4 ondate crypto hanno esaurito gli angoli su BTC/ETH (soffitto ~1.3). L'unico modo di superarlo è
+un **mercato diverso**. Aperto il fronte azioni/ETF via IB (paper, `gnzsnz/ib-gateway`, read-only).
+
+**Dati certificati + cache su disco** (`fetch_ib_equities.py` → `data/raw/eq_*.parquet`, ADJUSTED_LAST
+div+split, gitignored = cache locale; loader `eqlib.py` con lru_cache → ricerca legge da disco, MAI
+da IB). Universo: 9 SPDR settoriali classici dal **1998 (27.5y)** + XLRE(2015)/XLC(2018) + SPY(1996,
+30y)/QQQ/IWM/GLD/HYG/TLT. Tutti integri (monotoni, no dup, no spike>50%, 0 gap lunghi).
+NB bug timestamp risolto: `pd.Timestamp` a risoluzione µs → salvati in secondi, corretti a ms.
+
+## EQ-MOM01 — momentum cross-sectional settoriale
+
+Costruzione causale (`eq_sector_momentum.py`): ogni 21g, momentum = blend lookback [63,126,252]g con
+skip-21 (12-1 classico), z-score cross-sectional. long-only top-k (full-invested, confronto
+like-for-like con SPY) e long-short (dollar-neutral, test alpha puro). Netto fee, hold-out OOS 2015+.
+
+### Risultati (9 settori, 1998-2026)
+| strategia | CAGR | Sharpe (pre15/OOS15+) | maxDD | corr SPY |
+|---|---|---|---|---|
+| **SPY buy&hold** | 8.2% | **0.51** (0.31/0.82) | 55% | — |
+| EW 9 settori | 8.9% | 0.56 (0.44/0.76) | 53% | 0.96 |
+| MOM long top-3 | 7.7% | 0.50 (0.32/0.76) | 47% | 0.85 |
+| MOM long vol-target 15% | 7.3% | 0.52 | 39% | 0.75 |
+| **MOM long-short top-3** | −0.9% | **−0.08** (−0.19/0.08) | 32% | −0.20 |
+
+### Verdetto: NESSUN edge vs SPY
+- **Long-short Sharpe −0.08** → l'alpha cross-sectional di momentum settoriale è **morto** su 27 anni
+  (decadimento post-2000 noto in letteratura). Niente alpha market-neutral.
+- **Long-only ≈ SPY**: corr 0.85, **uplift marginale ~0.00** (blend 75/25 +0.012 FULL / +0.001 OOS;
+  50/50 +0.015 / −0.010). È un SPY a beta più basso, non un edge. Plateau stabile ma sempre ~0.50
+  (vs SPY 0.51); sugli 11 settori (2018+) fa peggio (0.69 vs 0.82). Fee-robusto (ma niente da salvare).
+- L'unico beneficio (maxDD 55%→39%) è del **vol-target**, non del momentum (lo daresti a SPY stesso).
+
+## Lezione (coerente col progetto)
+Il momentum **relative-value** è morto anche in equity, come nel crypto (ortho wave). Il baseline
+equity da battere è SPY buy&hold (Sharpe ~0.51 full / 0.82 OOS), ostico come il toro crypto.
+
+## Prossimo angolo plausibile (NON ancora testato)
+L'analogo equity di TP01 (l'unica cosa che ha retto nel crypto = trend DIFENSIVO): **time-series
+trend su SPY long-flat/long-bonds** — non per battere il CAGR ma per **tagliare il 55% di drawdown**
+restando vicino al ritorno. È il punto dove vive il valore robusto in equity (e dove il cross-section
+NON guarda). Da provare con lo stesso gauntlet: marginale vs SPY, OOS lungo, plateau.

docs/diary/2026-06-22-equities-spy-trend.md

@@ -0,0 +1,48 @@
+# 2026-06-22 — EQ-TREND01: trend DIFENSIVO su SPY = edge difensivo REALE (analogo di TP01)
+
+## Contesto
+Il momentum cross-sectional settoriale è morto (EQ-MOM01: long-short Sharpe −0.08, long-only ≈ SPY).
+Ma nel crypto l'unica cosa che ha retto NON era relative-value: era **TP01**, un trend DIFENSIVO che
+taglia il drawdown. L'equity ha lo stesso buco: SPY buy&hold Sharpe ~0.54 ma maxDD **55%**.
+
+## Costruzione (causale, stile TP01)
+`eq_spy_trend.py`. TSMOM multi-orizzonte [21,63,126,252]g, target = frazione di orizzonti in
+trend-up (allocazione graduale 0..1), opz. vol-target. Posizione decisa a ≤i-1, tenuta da i. Netto
+fee. Varianti: long-flat (cash in risk-off), long-bonds (TLT, solo 2016+), SMA-200 (Faber). Dati da
+cache eqlib (ADJUSTED, nessun IB). Periodo 1997-2026, OOS 2015+.
+
+## Risultati
+| strategia | CAGR | Sharpe (pre15/OOS) | maxDD | in-mkt |
+|---|---|---|---|---|
+| SPY buy&hold | 9.0% | 0.54 (0.38/0.82) | 55% | 99% |
+| **SMA-200 (Faber)** | 7.0% | **0.65** (0.52/0.88) | **29%** | 76% |
+| TSMOM lf cap1.0 | 5.7% | 0.57 (0.44/0.78) | 30% | 92% |
+| TSMOM lf vt15% | 5.7% | 0.62 (0.51/0.78) | **25%** | 92% |
+
+**Drawdown nei bear (TSMOM vs SPY):** dot-com 26%/49% · GFC **19%/55%** · COVID 17%/34% · 2022 16%/24%.
+
+**Plateau** (long-flat): ogni config Sharpe 0.56-0.65 (> SPY 0.54), maxDD 25-31% (~metà di SPY).
+SMA-200 il più semplice E il migliore (Sh 0.65, OOS 0.88, DD 29%). **Fee-robusto** (Sh 0.48 a
+0.10%/lato), basso turnover.
+
+**Marginale vs SPY:** corr 0.73. blend 50/50 uplift +0.035 FULL / +0.031 OOS (modesto positivo);
+100% trend uplift −0.012 / −0.041 (nel toro recente la difesa costa).
+
+## Verdetto: edge DIFENSIVO reale (non alpha) — analogo di TP01
+- ✅ Sharpe 0.54→0.62/0.65, **maxDD dimezzato** (55%→~27%, nei bear lenti più che dimezzato),
+  plateau robusto, fee-robusto, **eseguibile a $0.5-2k** (switch mensile SPY/cash).
+- ⚠️ NON genera ritorno (CAGR −2/3pp): è risk-management, come TP01.
+- ⚠️ I tagli grossi (dot-com/GFC) sono IN-SAMPLE; l'OOS 2015-26 è quasi tutto toro → lì ha seguito
+  SPY a beta minore (ma COVID, OOS, dimezzato). La difesa "serve" nei bear, rari nell'OOS.
+- ⚠️ long-bonds (TLT) non convince (TLT distrutto 2022).
+
+## Lettura strategica
+Primo positivo del fronte equity, e dello stesso TIPO che ha retto nel crypto: trend difensivo, non
+relative-value. Conferma la lezione cross-mercato: **il valore robusto è nel ridurre il rischio
+(trend long-flat), non nel battere il buy&hold**. Da solo non risolve €50/g (problema di capitale).
+
+## Prossimo angolo plausibile
+**Trend multi-asset / GTAA** sull'universo ETF in cache (SPY/QQQ/IWM + TLT/GLD/HYG): un portafoglio
+di trend long-flat su classi d'attivo diverse di solito batte il trend mono-SPY sul rischio-aggiustato
+(diversificazione dei trend). + domanda cross-mercato: la sleeve equity-trend DIVERSIFICA il
+portafoglio crypto (TP01+XS01+VRP01)? (esecuzione split Deribit+IB).

docs/diary/2026-06-22-funding-carry-hl.md

@@ -0,0 +1,96 @@
+# 2026-06-22 — Funding-CARRY cross-sectional su Hyperliquid (FC01): LEAD fragile, NON regge
+
+## Contesto
+
+Onda "nuova ricerca mirata" (l'utente ha chiesto di cercare un angolo non coperto dalle due grandi
+ondate — sweep 104-ipotesi e ortho relative-value, entrambe esaurite sul *prezzo* BTC/ETH). L'unico
+meccanismo con una **fonte di ritorno diversa** non ancora testato su dati certi è il **carry da
+funding**: incassare il cashflow dei perp stando delta-neutral.
+
+### Scan di fattibilità dati (prima di tutto, lezione v2.0.0)
+- **Funding price-clock** (drift attorno agli stamp 00/08/16) sul feed Deribit certificato →
+  già testato nell'onda intraday (`agent_03_funding_clock_15m`) = **FAIL** ("il funding è un cashflow
+  perp-vs-spot; il prezzo index non ha drift tradabile attorno allo stamp al netto del trend").
+- **Funding carry su Deribit** (dove eseguiamo) → ccxt `fetch_funding_rate_history` = **0 righe**
+  (bloccato), Cerbero MCP espone solo `get_historical` (candele), endpoint funding = 404.
+- **Funding carry su Hyperliquid** → API pubblica `/info {"type":"fundingHistory"}` = **disponibile**,
+  oraria, tokenless, serie native dal 2023-05. HL è già l'universo certificato di XS01.
+
+### Dato scaricato e certificato
+`scripts/research/fetch_hl_funding.py` (backoff anti-429) → **19 major** (gli stessi di XS01),
+`data/raw/hlfund_<sym>_1h.parquet`. Certificazione: cadenza ~1h, **0 gap**, copertura 98-100%,
+funding annualizzato per asset da **APT +1.0%** a **NEAR +21.6%** (mediana ~+11.7%). Pochi `cap_hit`
+(ore con |funding|>0.06%/h) su INJ/TIA/SEI, plausibili in alt ad alta vol. Dato pulito.
+
+## Ipotesi e costruzione (FC01)
+
+Book dollar-neutral che **SHORTA i k perp ad alto funding** e **COMPRA i k a basso** → incassa il
+premio (chi è long paga il funding). Ritorno perp per un long = `price_ret − funding`. Causale come
+XS01: ogni H=10 giorni, segnale = media causale del funding giornaliero realizzato sugli ultimi L
+giorni (shift 1), rank cross-section, vol-target 20%, fee 0.05%/lato sul turnover.
+`scripts/research/funding_carry_hl.py`. Domanda chiave: **edge reale e ORTOGONALE a XS01**, o XS01
+travestito? (gli alt ad alto funding sono spesso i pompati = quelli che XS01 *compra*; qui li
+*shortiamo* → potenziale anti-correlazione, oppure il carry domina).
+
+## Risultati
+
+### Premio reale ma direzione-dipendente
+`carry` (short alto-funding) batte sistematicamente `anti` (long alto-funding, sempre molto negativo)
+→ **il premio di funding esiste**: shortare i perp ad alto funding paga, in aggregato.
+
+### Ma il book NON regge il gauntlet (19 asset, 2024-2026, 904g)
+- **Standalone base (L=7 k=5): FULL Sharpe −0.12, in-sample 0.44, HOLD −0.50, DD 28.6%**, −2.8%/anno.
+  Decadimento netto: 2024 **+0.44** → 2025 −0.06 → 2026 **−1.42**.
+- Correlazioni: TP01 −0.02, **XS01 −0.19** (ortogonale, come da ipotesi — NON è XS01 travestito),
+  VRP01 +0.05.
+- **`marginal_vs_tp01` = DILUTES**: `has_insample_edge=False` (in-sample 0.44 < 0.5),
+  `multicut_persistent=False`, blend w25 uplift FULL −0.21 / HOLD −0.39.
+- **Non aggiunge a XS01**: uplift w25 FULL −0.04 / HOLD −0.19.
+
+### Il colpo di grazia: FRAGILITÀ all'universo
+Un preview su 17 asset (mancavano NEAR e AAVE) dava FULL **+0.62**, ADDS, +0.22 uplift — un PASS
+tentatore. Sui 19 completi: **DILUTES**. Jackknife lascia-fuori-uno (base L=7 k=5):
+
+```
+19 asset:  FULL -0.12  HOLD -0.50
+-SUI   FULL -0.39      ...      -BTC  FULL +0.17
+-SEI   FULL -0.31               -AAVE FULL +0.26
+-BNB   FULL -0.29               -NEAR FULL +0.30
+=> FULL oscilla in [-0.39, +0.30] togliendo UN solo asset (range 0.70), attraversa lo zero.
+```
+
+Togliere **NEAR o AAVE** (i due assenti nel preview) **recupera il segno** → il preview era fortunato
+*proprio* perché quei due non c'erano ancora. **Un edge robusto non cambia segno per un singolo nome.**
+Le poche celle "buone" del plateau (es. L=7 k=3: HOLD 0.91) hanno **in-sample debole + hold-out forte**
+= la firma del hold-out-luck che la metodologia indurita uccide.
+
+## Perché fallisce (meccanismo)
+
+Tensione fondamentale **carry vs momentum**: il funding-carry shorta i forti (alto funding = domanda
+long aggressiva), ma in un mercato alt toro i forti **continuano a correre** (NEAR/AAVE: alto funding
+*e* prezzo su → shortarli perde più del premio incassato). Il premio di funding è reale in aggregato,
+ma il book cross-sectional equal-weight top-k è dominato da pochi nomi a funding estremo che *anche*
+trendano, e su 2.5 anni / 19 nomi questo basta a ribaltare il segno.
+
+## Verdetto
+
+**FC01 NON è uno sleeve.** Né deploy (è STAT-MODE: 10 gambe market-neutral, non eseguibile a $600),
+né lead affidabile: fragile all'universo (sign-flip su un nome), DILUTES vs TP01, non aggiunge a XS01,
+in-sample edge < 0.5, niente persistenza multi-cut, decadimento 2026. Conferma — di nuovo — il
+soffitto del progetto: promettente su un sottoinsieme fortunato, collassa sotto il gauntlet onesto.
+**Win metodologico:** lo scorer indurito + il jackknife d'universo hanno intercettato un falso
+positivo che il preview a 17 asset avrebbe promosso.
+
+## Lascito / lavoro futuro (NON inseguire ora)
+- I 19 parquet funding (`hlfund_*`) restano certificati per ricerca futura. Il fetcher NON va in cron
+  (FC01 fallito → niente da monitorare in forward).
+- Idee se mai si tornasse sul carry (NON ora): (a) **gate sul LIVELLO** di funding (short solo quando
+  estremo, regime-filter alla VRP01 IV-rank) invece dello short-top-k incondizionato; (b) cap sul peso
+  per-nome / neutralizzazione momentum per togliere il dominio NEAR/AAVE. Entrambe rischiano
+  overfitting su storia corta — soglia di prova alta.
+
+## Nota IB (thread parallelo, stessa sessione)
+Esplorato come fonte per il **basis CME crypto** (cugino eseguibile del carry). Gateway paper
+`gnzsnz/ib-gateway` su `127.0.0.1:4002` (read-only, `docker-compose.yml`), sonda `ib_probe.py`.
+Esito dati: **backtest del basis NON fattibile** (ContFuture back-adjusted; contratti scaduti = 1
+barra). IB resta valido per esecuzione/forward, non per scoprire l'edge. Dettagli nel corpo sessione.

docs/diary/2026-06-23-combo-paper-deploy.md

@@ -0,0 +1,33 @@
+# 2026-06-23 — Combo DEPLOYABLE in PAPER: TP01 (Deribit) + GTAA (IB), cross-venue
+
+## Decisione
+Dopo aver esaurito (onestamente) la ricerca di nuovi edge e anticipazioni cross-mercato, l'unica cosa
+VERA e deployabile e' la DIVERSIFICAZIONE: TP01 (crypto, Deribit) + GTAA (equity, IB), corr ~0.21 ->
+blend Sharpe ~1.5, maxDD dimezzato (diari 2026-06-22-deployable-combo). Si va in PAPER cross-venue.
+
+## Costruito
+- **`src/portfolio/gtaa.py`** — GTAA come sleeve di prima classe: trend difensivo long-flat TSMOM
+  [21/63/126/252g], vol-target 12%, EW su SPY/QQQ/IWM/TLT/GLD/HYG. Espone `gtaa_returns()` (Sharpe
+  full 0.64, 7542 barre 1996+) e `gtaa_weights()` (pesi ETF CORRENTI azionabili). Legge cache eq_*.
+- **`scripts/live/paper_combo.py`** — paper-tracker FORWARD-ONLY del blend 50/50 TP01+GTAA (crypto
+  compoundato sul grid giorni-di-borsa). Stato in data/paper_combo/. Mostra le posizioni azionabili
+  su entrambi i venue. SOLO le gambe eseguibili (XS01/VRP01 STAT-MODE esclusi).
+- **`fetch_ib_equities.py --only SPY,QQQ,...`** — refresh mirato dei 6 ETF GTAA (per il cron).
+- **`scripts/cron_daily.sh`** — aggiunto: up gateway IB (idempotente) -> refresh ETF GTAA ->
+  avanza paper_combo. Dipendenza cross-venue gestita (gateway paper sempre-up, restart unless-stopped).
+
+## Stato iniziale (2026-06-23)
+Paper combo init a 2000, forward da 2026-06-22. Posizioni azionabili:
+- TP01 (Deribit): BTC/ETH 0.0x (flat, TSMOM risk-off — coerente col live).
+- GTAA (IB): SPY 13% / QQQ 8% / IWM 9% / TLT 17% / GLD 2% / HYG 17% / cash 34% (difensivo).
+Catena end-to-end testata: gateway -> refresh ETF -> avanza paper. OK.
+
+## Onesta'
+- E' PAPER (rischio zero). Valida l'OPERATIVITA' cross-venue prima di capitale reale.
+- Sharpe atteso ~1.5 e' ottimistico (finestra crypto corta/favorevole); il dato robusto e' la
+  diversificazione (corr 0.21, DD dimezzato), non il livello assoluto.
+- A capitale reale e' un portafoglio su DUE conti (Deribit ~$600 + IB); GTAA frazionabile a basso
+  capitale, TP01 gia' armato. Prossimo passo eventuale: dashboard del combo + (molto dopo) capitale.
+
+## Prossimo
+Lasciar girare il paper forward (cron giornaliero) e ricontrollare l'equity tra qualche settimana.

docs/diary/2026-06-23-crossmarket-beyond-sp500.md

@@ -0,0 +1,38 @@
+# 2026-06-23 — Cross-market crypto-lead OLTRE l'SP500: bond, commodity, indici esteri -> niente
+
+## Obiettivo
+Estendere il test "crypto anticipa il mercato?" oltre SP500/azionario USA: commodity, bond, indici
+ESTERI (Europa/Asia, fasi orarie diverse = il caso a priori piu' favorevole a un lead vero).
+
+## Dati (IB, orari, cache fut_*_1h)
+ES/NQ/RTY (gia'); + ZN (T-note 10y), ESTX50 (Euro Stoxx50), DAX, NKD (Nikkei). Storia ~2-2.4y (2024+).
+Commodity GC/CL/HG: VUOTE (market-data subscription COMEX/NYMEX mancante sul paper) -> non testate.
+
+## Test (`fut_leadlag_generic.py`): crypto[T-8h->T] -> future[T->T+6h], non-sovrapposto, controllo=moto proprio future
+- ES/NQ/RTY: nessun edge (gia' noto).
+- ZN (bond): NEGATIVO (0/3 anni).
+- NKD (Nikkei): debole (t_crypto 0.2, Sharpe 0.66 ~ overnight drift, non crypto).
+- **ESTX50 / DAX: forte all'apparenza** — BTC->T0h: t_crypto ~7.8, Sharpe 2.5, ann ~22%, 3/3 anni.
+
+## Ma e' un ARTEFATTO DI CONFINE UTC (deep-dive `eu_overnight_deepdive.py`)
+- **Picco a coltello a T=00:00 UTC**: t/Sharpe salgono T20->T0 (2.5->7.8 / 0.24->2.45) e CROLLANO a
+  T=1h (t 1.3, Sharpe -0.09). Un lead vero non e' a coltello su una sola ora.
+- **GAP test**: inserendo 1h tra fine-segnale (00:00) e inizio-cattura, l'effetto MUORE
+  (Sharpe 2.45 -> -0.52, t 7.8 -> 1.6).
+- **Singola ora**: T=0h/H=1h (cattura 00:00->01:00) Sharpe +2.93 (t 8.7); T=1h/H=1h (01:00->02:00)
+  Sharpe -1.02. L'INTERO "edge" e' la barra di confine 00:00->01:00.
+- vs SEMPRE-LONG: always-long overnight e' negativo (-0.8/-1.2), quindi non e' overnight-drift; ma
+  l'uplift del crypto e' tutto nella barra di confine.
+-> Firma esatta di `day_boundary_robust` (CLAUDE.md): effetto che vive/muore spostando il confine del
+   giorno UTC di poche ore = etichettatura/contaminazione, NON anticipazione economica.
+
+## Verdetto
+NIENTE di tradabile oltre l'SP500 nemmeno. Su TUTTI i mercati il legame crypto->X e' o co-movimento
+contemporaneo (risk-beta) o artefatto di confine. L'anticipazione crypto->altri-mercati sfruttabile
+NON esiste su dati onesti (finestre non-sovrapposte + boundary-robust + gap). Conferma definitiva del
+soffitto del progetto, ora anche cross-mercato.
+
+## Cosa resta di valore (immutato)
+La diversificazione TP01(crypto)+GTAA(equity), corr 0.21 -> Sharpe portafoglio ~1.5, DD dimezzato.
+Quello e' strutturale e deployabile; l'anticipazione cross-mercato no.
+Script: fetch_ib_futures.py (multi-exchange), fut_leadlag_generic.py, eu_overnight_deepdive.py.

docs/diary/2026-06-23-crypto-overnight-lookahead.md

@@ -0,0 +1,29 @@
+# 2026-06-23 — "Monitor Deribit / trade IB": il gap crypto->equity e' LOOK-AHEAD
+
+## Idea testata (utente)
+Guardare crypto live su Deribit (24/7) e tradare l'indice su IB sul segnale del gap overnight.
+
+## Trappola trovata
+Il segnale crypto [P 21:00 -> D 13:00 UTC] e il "gap" equity [P close -> D open 13:30] coprono QUASI
+LE STESSE ORE. Condizionare il gap sul crypto-overnight = correlare due ritorni dello STESSO intervallo
+notturno -> look-ahead. All'entrata (D 13:00, pre-open) il gap e' GIA' avvenuto: non catturabile.
+
+## Prova (net 2bps, sqrt(252))
+| target | OVERLAP gap (look-ahead) | TRADABILE intraday (post-entrata) |
+|---|---|---|
+| SPY | Sharpe 3.60 (OOS 5.23) | -0.03 (OOS 0.12) |
+| QQQ | Sharpe 4.01 (OOS 5.47) | 0.25 (OOS 0.43) |
+| IWM | Sharpe 3.98 (OOS 5.72) | 0.15 (OOS 0.44) |
+
+Lo "Sharpe 5" e' artefatto. L'edge REALE tradabile via ETF (intraday, entri all'open) ~0, muore a costi.
+NB: anche i Sharpe "gap" del workflow 65-agenti erano (a) look-ahead di overlap e (b) sotto-annualizzati
+(sqrt(52) invece di sqrt(252)); il verdetto "non deployabile" resta, rafforzato.
+
+## Cosa resta possibile (non testato, serve dato)
+L'unica versione onesta dell'idea: entrare a META' notte via FUTURES IB e vedere se crypto [P21:00->T]
+predice il future indice [T->open] su finestre NON sovrapposte (crypto come sensore di rischio piu'
+veloce). Richiede dati INTRADAY dei futures (ES/NQ/RTY), non in cache -> data step se si vuole indagare.
+
+## Lezione
+Un risultato "troppo bello" (Sharpe 5) e' un test di disciplina: era overlap di finestre. Catturato.
+Script: crypto_overnight_equity.py (versione artefatto), crypto_overnight_honest.py (decomposizione).

docs/diary/2026-06-23-futures-overnight-leadlag.md

@@ -0,0 +1,37 @@
+# 2026-06-23 — "Monitor Deribit / trade IB" su futures: test ONESTO non-sovrapposto -> edge ~0
+
+## Idea
+Monitorare crypto live (Deribit 24/7) ed entrare sul FUTURE indice IB (ES/NQ/RTY, tradato di notte)
+a meta' notte, catturando il moto SUCCESSIVO -> finestre NON sovrapposte (no look-ahead, vs il "gap"
+che era contemporaneo al segnale).
+
+## Dati
+`fetch_ib_futures.py` -> data/raw/fut_{es,nq,rty}_1h.parquet (ContFuture orario, UTC). ES 3y (2023-06+),
+NQ 2.75y, RTY 2.3y. (NB: ContFuture NON accetta endDateTime -> chiamata singola "4 Y" = ~3y max orari.)
+
+## Test (`fut_overnight_leadlag.py`)
+entrata a T (ora UTC notturna): segnale = crypto[P21:00->T]; controllo = future[P21:00->T] (moto
+PROPRIO del future); cattura = future[T->open 13:00]. Incrementale: crypto predice la cattura OLTRE il
+moto proprio del future? Trade: sign(crypto[P21:00->T]) * future[T->open], net 2bps. T in {0,3,6,9}h.
+
+## Risultati
+| future | miglior Sharpe (trade crypto) | t_crypto incrementale | esito |
+|---|---|---|---|
+| ES (S&P500) | ~0 / negativo (-0.03..-0.93) | 0..1.5 | NESSUN edge |
+| NQ (Nasdaq) | 0.41 (T=3h) | 0.5 (debole) | momentum del future, non crypto |
+| RTY (Russell) | 0.40-0.77 | 2.0-2.7 (BTC->RTY) | soffio debole, non robusto |
+
+- **SP500: NIENTE.** Il crypto della prima notte non predice l'ES della seconda. Il "Sharpe 5" del gap
+  era interamente look-ahead (finestre sovrapposte): catturato e ucciso.
+- **RTY (small-cap)** e' l'unico con t_crypto incrementale ~2-2.7 e crypto che AGGIUNGE oltre il moto
+  proprio del future (futOwn Sharpe negativo). MA: Sharpe 0.4-0.5 modesto, 24 config (multiple-testing),
+  storia 2.3y, per-anno INCOERENTE (BTC->RTY T=3h: 2024 +0.99 / 2025 +0.52 / 2026 -0.31).
+
+## Verdetto
+L'idea "monitor Deribit / trade IB" NON da' un edge tradabile, men che meno su SP500. Il forte
+fenomeno crypto<->equity e' CO-MOVIMENTO contemporaneo (risk-beta overnight), non anticipazione: quando
+si impone una finestra causale non-sovrapposta, l'edge svanisce (efficienza di mercato). L'unico
+residuo (crypto->small-cap overnight) e' debole, borderline su multiple-testing e instabile per anno
+-> forward-monitor al piu', NON deploy. Coerente col soffitto del progetto.
+
+Script: fetch_ib_futures.py, fut_overnight_leadlag.py. (look-ahead documentato: 2026-06-23-crypto-overnight-lookahead.md)

docs/diary/2026-06-23-skyhook.md

@@ -0,0 +1,141 @@
+# 2026-06-23 — SKH01 "Skyhook": porting onesto del sistema ES dual-timeframe su BTC/ETH
+
+Branch: `strategy_skyhook`. Engine: `src/strategies/skyhook.py`. Harness: `scripts/research/skyhook/skyhooklib.py`.
+Test: `tests/test_skyhook.py` (5 pass). Ricerca: `scripts/research/skyhook/{sweep,grid,check_v1}.py` + `runs/`.
+
+## Il brief
+
+Sistema "Skyhook" (origine ES / E-mini S&P, genetico, a doppio timeframe), da portare su crypto:
+- **data2 = 690 min (segnale)**, **data1 = 230 min (esecuzione)**. NB **690 = 3 × 230**.
+- NON trend-follower: entra **solo** quando coincidono (a) un **regime** di volatilità/volume e
+  (b) un **pattern** di breakout.
+- Pipeline per barra: indicatori (BuzVola su ATR, BuzVolume su volume, tipo-Chande 0-100) →
+  fasce regime → pattern (Donchian/breakout su data2) → composer (regime AND pattern) →
+  ingresso (max 1/giorno, stop-and-reverse) → uscite (time-based asimmetrico uscitalong=24 /
+  uscitashort=18 + stop/profit).
+- Ancore demo: trend lineare → **BuzVola=50** (vol steady → neutro), **BuzVolume=100** (volume in rampa).
+
+## Ricostruzione (fedele + onesta)
+
+- **Resample dal feed 5m certificato** con `origin='epoch'`: 230 min = 46×5m, 690 min = 138×5m,
+  e i confini 690 sono un **sottoinsieme** dei confini 230 → una barra HTF chiude esattamente su
+  una chiusura LTF. Merge HTF→LTF causale: `merge_asof` backward sulla **chiusura HTF** (≤ chiusura
+  LTF), così una barra HTF è usata solo quando è davvero chiusa. (~2287 barre/anno LTF, ~762 HTF.)
+- **BuzVola / BuzVolume = `chande01`** (Chande Momentum Oscillator normalizzato 0-100): serie
+  steady → 50, rampa-su → 100, rampa-giù → 0. Le ancore demo sono soddisfatte a livello di
+  indicatore (è la lettura fedele: "vol steady → neutro"). NB: l'EMA-ATR su un *linspace* sintetico
+  dà 100 per drift di warm-up/floating-point, non per comportamento reale — su BTC reale BuzVola
+  oscilla intorno a 50 (EMA-ATR vs SMA-ATR corr 0.90).
+- **Pattern** = Donchian breakout leak-free (shift(1)) su HTF, `ptn_n` barre (default 13 da 13/13/1).
+- **Regime** = bande-soglia tunabili su BuzVola/BuzVolume (i magici interi 4/3/2 - 4/2/2 non sono
+  nel brief; ricostruiti come `[vola_lo,vola_hi]` × `[vol_lo,vol_hi]`).
+- **Composer** = regime AND pattern. **Ingressi** ≤1/giorno (prima barra qualificante).
+- **Uscite**: time-based asimmetrico (`uscitalong`/`uscitashort` barre LTF) + hard stop/profit. Lo
+  "stop 2000 / profit 5000" in $ del sistema ES → **multipli di ATR LTF** (scale-free): default
+  `sl_atr=2.0`, `tp_atr=5.0` (~ rapporto 40:100 pt ES), con modalità `pct` alternativa.
+- Engine espresso come **entries `{dir,tp,sl,max_bars}`** per `backtest_signals` (motore onesto del
+  progetto: TP/SL intrabar, max_bars, non-overlap). Causalità verificata con prefix-recompute
+  (0 mismatch).
+
+## Baseline → V1 (lever scout + grid, inline, veloce)
+
+- **Baseline** (default 13/13, sl2/tp5, vola[35,95], vol_lo50): causale, fee-surviving, FULL Sharpe
+  BTC +0.91 / ETH +0.64, ma **HOLD-OUT debole** (BTC −0.09 / ETH +0.17) → FAIL del gate onesto.
+- **Lever scout** (`sweep.py`): gli **short servono** (long_only → HOLD −0.52); il **regime gate
+  conta** (togliere la banda vola → HOLD −0.80); il **floor di volume** a 50 *frenava* l'hold-out
+  (vol_lo=40 o 0 → PASS); **breakout più lento** (ptn_n=55) e **stop più larghi** (sl2.5/tp6)
+  alzano l'hold-out.
+- **Grid combinato** (`grid.py`): vincitrice **SKH01-V1**
+  `SkyhookParams(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35, vola_hi=95, vol_lo=0.0)`:
+  - **min-asset FULL +0.69, HOLD-OUT +0.64** (BTC 0.64 / ETH 0.64), **PASS**, fee-surviving a 0.30%RT.
+  - BTC FULL +0.69/+275% DD49% ; ETH FULL +1.01/+871% DD31% ; entrambi HOLD-OUT positivi.
+  - **Marginal vs TP01 = ADDS** e regge i gate induriti: **corr 0.06** (ortogonale, NON trend-beta),
+    `has_insample_edge=True` (Sharpe in-sample standalone 1.15), `is_hedge=False`, multi-cut
+    persistente. Blend **0.75·TP01 + 0.25·SKH01: HOLD Sharpe 0.31 → 0.74 (+0.44), DD 11.9%**;
+    blend 50/50 HOLD 0.88, DD 17.8%.
+  - Unico sub-gate fallito: `clean_year_uplift` +0.014 (sotto 0.02) → `earns_slot=False` per un pelo,
+    nonostante tutto il resto sia forte. **Debolezza principale: DD standalone alto (40-49%).**
+
+→ SKH01 è un **diversificatore quasi-ortogonale** reale (non un TP01 travestito): da solo è
+volatile, ma come sleeve al 25% migliora moltissimo l'hold-out del portafoglio a DD bassissimo.
+
+## Onda 1 (`skyhook-improve`, 30 agenti) — winner intermedio
+
+Famiglie: param (RR, ptn_n, regime bands, exit bars, chande, local), regime-redef (percentile,
+realized-vol, vol-expansion, LTF), pattern (confirmation, ROC, Keltner, NR, dual), exit + overlay,
+ognuna verificata da 2 scettici. Risultato: **winner intermedio**
+`SkyhookParams(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16, vola_lo=35, vola_hi=95, vol_lo=0)`
+— **minFull +0.83, minHold +0.81** (vs V1 +0.69/+0.64), causale, fee-surviving 0.30%RT, marginal
+**ADDS** (corr 0.05, has_insample_edge, robust_oos, multicut, clean_year_uplift +0.37), blend w25
+uplift_hold +0.58. **MA standalone maxDD ancora 34% (BTC) / 31% (ETH) → l'unico goal mancato era il DD<30%.**
+
+## Onda 2 (`skyhook-improve-v2`, 14 famiglie DD-reduction) — SKH01-V2-DD vince
+
+Obiettivo: tagliare il **DD standalone <30%** tenendo hold-out + `earns_slot`, e alzare l'uplift di
+portafoglio. 14 famiglie (ensemble param/struct, vol-target, DD kill-switch, RR/stop grid, regime
+tight, percentile, vol-expansion, breakout confirmation, dual-TF, asimmetria L/S, cadenza, chande,
+Keltner), ognuna verificata da 2 scettici avversariali (window-luck/multicut/jackknife +
+causalità/fee/plateau/overfit). Esito: **il winner intermedio cade.** Nuovo campione **SKH01-V2-DD**
+(famiglia ASYM_LS, `src/strategies/skyhook.py:SKH01_V2_DD`, run `runs/SKH2_ASYM_LS.py`):
+
+- **Config:** stesso SEGNALE del winner (`ptn_n=45, vola_lo=35, vola_hi=95, vol_lo=0, exit-bars 24/16`)
+  ma EXIT commutati da ATR a **percentuale fissa ASIMMETRICA** — long `sl=4% / tp=10%`, short
+  `sl=2% (più stretto) / tp=8%`. Motivazione meccanica: in crypto lo short si fa steamrollare da uno
+  spike vola e lo stop-ATR si allarga lasciando correre la perdita → il %-SL stretto sullo short
+  **cappa la perdita per-trade** che FORMA il maxDD. (Implementato come override per-direzione nel
+  motore, backward-compatible: campi `*_short=None` → comportamento simmetrico invariato.)
+- **Numeri veri (verificati indipendentemente via `sk.study(SKH01_V2_DD)`):** standalone maxDD
+  **BTC 21.4% / ETH 27.4%** (<30% ✓, vs 34.4/30.5 del winner) — **goal RAGGIUNTO**; minFull **+0.99**,
+  minHold **+1.26**; causalità **0/400** entrambi gli asset; fee@0.30%RT BTC +1.05 / ETH +0.80
+  (positiva anche a 0.40%). Marginal vs TP01 **ADDS** (corr 0.09, has_insample_edge, is_hedge=False,
+  robust_oos, multicut, clean_year_uplift +0.57). **Blend 0.75·TP01 + 0.25·SKH: uplift_hold +0.87**
+  (vs +0.58 del winner); **blend 50/50: full 1.84 / hold 1.59 / DD 10.7%**. earns_slot=True,
+  beats_winner=True. **Plateau reale** (i vicini Spct_mb14/16 sl2% tengono DD 27-28%), non knife-edge.
+  Entrambi gli scettici: holds_up=True, confidence high, killer_finding=null.
+
+**Top-3 dell'onda 2 (criteri onesti):**
+
+| # | Famiglia | maxDD (BTC/ETH) | minHold | w25 uplift_hold | Verifica |
+|---|---|---|---|---|---|
+| **1** | **ASYM_LS → SKH01-V2-DD** | 27.4% (21.4/27.4) | +1.26 | **+0.87** | 2/2 high, killer=null ✅ |
+| 2 | ENS_STRUCT (3-regime ensemble) | **22.9%** (21.2/22.8) | +1.00 | +0.67 | 2/2 high — ma 3 motori da eseguire |
+| 3 | TPSL_DD (%-SL/TP hard) | 28.0% (28/25.5) | +1.11 | +0.75 | 1/1 (rate-limit) — caveat hedge-like |
+
+**Lezioni anti-DD:**
+- **Ha funzionato (STRUTTURA dell'exit, non i parametri):** cambiare il MECCANISMO di uscita — %-SL
+  hard, asimmetria L/S, o ensemble di exit/regime diversi (decorrelazione). Il DD del winner nasce
+  dalla coda intra-trade negli spike ATR; il %-SL la cappa.
+- **NON ha funzionato (la leva non raggiunge il DD vincolante):** DD kill-switch entry-only (sopprime
+  solo le NUOVE entry, non chiude il trade aperto che forma il maxDD → floor 33-36%); vol-target
+  causale (DD<30 e uplift≥0.55 mutuamente esclusivi; cap>1 PEGGIORA il DD levereggiando nel pre-crash);
+  cadenza/FREQ (accorciare gli hold short fa esplodere ETH a 50-66%); dual-TF (LTF è resample dello
+  stesso prezzo → quasi-tautologico, DD invariato).
+- **Bocciato dagli scettici come overfit:** PATTERN_CONF (sub-30 solo a vola_lo=45, knife-edge: sl_atr
+  ±0.5 → ETH 40-47%; la conferma "close_loc" da sola NON taglia il DD). Esempio canonico del perché
+  serviva la doppia verifica.
+- **Non promuovibili:** PCTL_DD (numeri spettacolari ma **0 verifiche**, le 2 sono morte per rate-limit
+  → forward-monitor, non fidato); ENS_PARAM / TPSL_DD (battono i gate ma uplift recency/hedge-loaded,
+  concentrato nei regimi TP01-down → forward-monitor).
+
+**Promozione (questa sessione):** `SKH01_V2_DD` canonico nel motore + override exit-short
+asimmetrici (backward-compatible, V1/winner invariati) + 3 test nuovi (8/8 pass).
+
+**Sleeve cablato @0.25 effettivo** (`src/portfolio/sleeves.skyhook_sleeve` → `active_sleeves`): i tre
+sleeve preesistenti scalati nel restante 0.75 mantenendo il rapporto 55:25:20 → **TP01 41.25% / XS01
+18.75% / VRP01 15% / SKH01 25%**. Report del portafoglio (4 sleeve, `run_portfolio.py`):
+
+| | FULL Sharpe | FULL DD | HOLD-OUT Sharpe | HOLD-OUT DD |
+|---|---|---|---|---|
+| 3 sleeve (TP01+XS01+VRP01) | 1.68 | 14.3% | 1.63 | 3.4% |
+| **+ SKH01 @25%** | **2.13** | **7.8%** | **2.30** | 3.5% |
+| Δ | **+0.45** | **−6.5pt** | **+0.67** | ~0 |
+
+→ aggiungere Skyhook **alza lo Sharpe full +0.45 e DIMEZZA il DD full (14.3→7.8%)**, e alza l'hold-out
++0.67 a DD invariato. Portafoglio combinato: FULL Sh 2.13 / ret +365% / DD 7.8%, HOLD Sh 2.30 / DD 3.5%,
+positivo ogni anno (2019-26, DD annuo ≤7.8%) vs buy&hold 50/50 FULL Sh 0.93 / DD 76%.
+
+**Caveat onesti / NON deploy:** è un portafoglio di **ricerca** (peso fisso, no costi di ribilanciamento
+reale a $600; lo Sharpe daily-step di Skyhook è la convenzione del lens). ETH DD standalone 27.4% ha
+margine sottile vs 30%. Prima di un eventuale deploy: ri-verificare la causalità sul **codice di
+esecuzione reale** (qui è l'harness di ricerca) e i costi del book a 230m (ribilanciamento più frequente
+del resto). XS01/VRP01 restano STAT-MODE/lead. Per ora: research win + sleeve cablato, forward-monitor.

docs/diary/2026-06-23-tail-hedge-lab.md

@@ -0,0 +1,66 @@
+# 2026-06-23 — Tail-hedge / protezione DD del combo (incl. OPZIONI): vince la guardia-drawdown
+
+## Obiettivo (goal utente)
+Trovare uno sleeve/overlay da AGGIUNGERE al combo (TP01+GTAA) per proteggere il drawdown e gli anni
+tipo 2022. Valutare anche le opzioni.
+
+## Diagnosi del rischio (decisiva)
+Il MaxDD del combo (1x) e' **8.4%** e il 2022 fu **-4.4%**: NON un crash, un **grind lento** (peggior
+giorno -2.8%, peggior mese 2022 -1.6%). Il doppio trend (TP01+GTAA long-flat) gia' taglia i crash
+veloci. Il tail residuo = (a) whipsaw da mercato choppy (2022), (b) gap/crash overnight LATENTE (TP01
+non reagisce intraday, non nel campione storico), (c) la LEVA.
+
+## Candidati testati (`tail_hedge_lab.py`)
+| protezione | MaxDD | 2022 | Sharpe | CAGR |
+|---|---|---|---|---|
+| combo baseline | 8.4% | -4.4% | 1.48 | 11.3% |
+| **+ guardia-DD -4%** | **5.8%** | **-1.8%** | 1.38 | 9.2% |
+| + vol-target 5% | 8.4% | -5.9% | 1.46 | 8.0% |
+| + opzioni (put/put-spread, budget 3%/y) | 8.4% | -4.4% | 1.48 | 11.3% |
+
+### OPZIONI (valutate): NON adatte al 2022
+- Put-spread/long-put LONG su indice 50/50 BTC/ETH (mirror di VRP01), premio BS su DVOL reale, payoff
+  sul path. Strike corretti (compra -0.30delta, vende -0.10delta).
+- Sempre-on costa **~50%/anno** di premio -> con budget 3%/anno size ~0.06-0.10x = effetto ~nullo.
+- Nel 2022 (grind, niente crash settimanali) **sanguinano** (scadono inutili) -> Δ2022 ~0.
+- Pagano SOLO nei crash secchi: stress -30% overnight -> put paga **+25% netto**, put-spread +3.8%.
+- Verdetto: assicurazione BLACK-SWAN cara e fuori-bersaglio per il grind. Utile solo come piccola
+  copertura del gap overnight latente, NON come fix del 2022.
+
+### GUARDIA DRAWDOWN: centra il rischio
+De-risk (esposizione 1.0->0.4) quando il DD da picco supera -4%, ri-rischia a -1.6%. Targetizza il
+grind: MaxDD 8.4%->5.8%, 2022 -4.4%->-1.8%, ogni anno DD intra <=5.4%. Costo: Sharpe 1.48->1.38,
+CAGR -2.1pp (de-risca sui cali, perde rimbalzo — prezzo onesto della protezione).
+
+### A LEVA (dove il tail morde)
+guard applicato pre-leva: 2x 2022 -15.6%->-10.9%, MaxDD 28%->24%; 3x resta MARGIN-CALL (DD 39%>=33%).
+-> la protezione rende il 2x sopportabile; il 3x va evitato comunque.
+
+## Raccomandazione
+AGGIUNGERE una **guardia-drawdown a livello di portafoglio** al combo (overlay, niente premio):
+e' la protezione che colpisce il rischio reale (grind/2022) a costo Sharpe minimo. Le opzioni NO come
+fix del 2022; eventualmente una micro-allocazione deep-OTM come assicurazione black-swan separata.
+Vol-target NON aiuta (il 2022 non e' uno spike di vol).
+
+## Onesta'
+- Il guard e' REATTIVO (de-risca dopo l'inizio del DD, restituisce un po' di rimbalzo) -> costa CAGR.
+- Trade: -2.1pp CAGR per dimezzare il MaxDD e azzerare quasi il 2022. Sensato se la priorita' e' il DD.
+- Parametri (-4% trigger) semplici; il meccanismo (non la soglia esatta) e' la sostanza.
+Script: tail_hedge_lab.py.
+
+## Aggiornamento — protezioni CLASSICHE (stop-loss), goal "prova anche SL" (`stops_lab.py`)
+Confronto equo (trigger/re-entry sul NAV di mercato, non sull'equity congelata):
+| protezione | Sharpe | MaxDD | 2022 | CAGR | in-mkt |
+|---|---|---|---|---|---|
+| baseline | 1.48 | 8.4% | -4.4% | 11.3% | 100% |
+| **soft-guard -4% (0.4x)** | **1.38** | **5.8%** | **-1.8%** | 9.2% | 100% |
+| trail-stop -4% (uscita tot.) | 1.07 | 7.5% | +0.0% | 6.6% | 42% |
+| trail-stop -6% (re:newhigh) | 1.34 | 6.6% | -2.1% | 9.0% | 72% |
+| trail-stop -8% | 1.41 | 8.3% | -4.2% | 10.0% | 87% |
+| stop mensile -5% | 1.48 | 8.4% | -4.4% | 11.3% | 100% (mai scatta) |
+| vol-stop (>90pctl) | 1.48 | 8.4% | -4.4% | 10.4% | 100% |
+
+VERDETTO: lo SL classico funziona solo a -6% (e resta inferiore al soft-guard); a -4% fa WHIPSAW
+(Sh 1.48->1.07, fuori mercato 58%) perche' l'uscita TOTALE viene choppata nel grind. Il soft-guard
+alla stessa soglia non whippa (de-risk parziale 0.4x). Stop mensile/vol inutili (bersaglio sbagliato).
+Conferma: per un DD da grind, de-risk PARZIALE > stop-loss duro. Soft-guard -4% confermato come scelta.

scripts/cron_daily.sh

@@ -11,5 +11,10 @@ mkdir -p logs
   uv run python scripts/live/paper_portfolio.py                       # avanza paper TP01+XS01
   uv run python scripts/live/paper_prevday.py                         # forward-monitor lead prevday-breakout (PAPER, non deploy)
   uv run python scripts/live/live_execute.py --execute                # TP01 LIVE su Deribit (gated da config/live.json)
+  # --- COMBO cross-venue (PAPER): refresh ETF IB (GTAA) + avanza paper TP01+GTAA ---
+  docker compose up -d ib-gateway >/dev/null 2>&1                      # gateway IB paper (idempotente)
+  for i in $(seq 1 25); do (echo > /dev/tcp/127.0.0.1/4002) >/dev/null 2>&1 && break; sleep 6; done
+  uv run --with ib_async python scripts/research/fetch_ib_equities.py --only SPY,QQQ,IWM,TLT,GLD,HYG  # ETF GTAA freschi
+  uv run python scripts/live/paper_combo.py                           # avanza paper combo (forward-only)
   echo "===== done $(date -u '+%H:%M:%SZ') ====="
 } >> logs/cron_daily.log 2>&1

scripts/live/paper_combo.py

@@ -0,0 +1,117 @@
+"""PAPER COMBO — forward-only del combo cross-venue TP01 (Deribit) + GTAA (IB), NUDO vs PROTETTO.
+
+Le due gambe eseguibili a basso capitale (XS01/VRP01 STAT-MODE esclusi), scorrelate (corr ~0.21) ->
+blend Sharpe ~1.5, DD dimezzato. Traccia FORWARD-ONLY DUE versioni in parallelo:
+  * NUDO     = blend 50/50 TP01+GTAA
+  * PROTETTO = stesso blend + GUARDIA-DRAWDOWN -4% (de-risk a 0.4x quando il DD da picco supera -4%,
+               ri-rischia a -1.6%). Backtest: MaxDD 8.4%->5.8%, 2022 -4.4%->-1.8%, Sharpe 1.48->1.38
+               (diario 2026-06-23-tail-hedge-lab). Le opzioni NON aiutano il grind del 2022 -> escluse.
+Crypto compoundato sul grid giorni-di-borsa. NESSUNA esecuzione reale. Mostra posizioni azionabili.
+
+  uv run python scripts/live/paper_combo.py [--status|--reset]
+"""
+from __future__ import annotations
+import sys, json
+from pathlib import Path
+PROJECT_ROOT = Path(__file__).resolve().parents[2]
+sys.path.insert(0, str(PROJECT_ROOT))
+import numpy as np, pandas as pd
+from src.portfolio.sleeves import _tp01_returns, _tp01_positions
+from src.portfolio.gtaa import gtaa_returns, gtaa_weights
+
+STATE_DIR = PROJECT_ROOT / "data" / "paper_combo"
+STATE = STATE_DIR / "state.json"
+EQ = STATE_DIR / "equity.csv"
+INITIAL = 2000.0
+W_CRYPTO = 0.5
+DD_TRIGGER = 0.04   # guardia-drawdown della versione PROTETTA
+
+
+def combo_daily(wc: float = W_CRYPTO) -> pd.Series:
+    tp = _tp01_returns()
+    if tp.index.tz is None:
+        tp.index = tp.index.tz_localize("UTC")
+    eq = gtaa_returns().dropna()
+    grid = eq.index[eq.index >= tp.index[0]]
+    cum = (1.0 + tp).cumprod()
+    tpg = (cum.reindex(cum.index.union(grid)).ffill().reindex(grid)).pct_change()
+    J = pd.concat({"c": tpg, "e": eq.reindex(grid)}, axis=1).dropna()
+    return (wc * J["c"] + (1 - wc) * J["e"]).dropna()
+
+
+def apply_dd_guard(r: pd.Series, trigger: float = DD_TRIGGER) -> pd.Series:
+    """De-risk a 0.4x quando il DD da picco > trigger; ri-rischia a 1.0x quando < 0.4*trigger."""
+    rv = r.values; n = len(rv); eq = np.cumprod(1 + rv); pk = np.maximum.accumulate(eq)
+    expo = np.ones(n); on = True
+    for i in range(1, n):
+        ddi = (pk[i - 1] - eq[i - 1]) / pk[i - 1] if pk[i - 1] > 0 else 0.0
+        if ddi > trigger: on = False
+        if ddi < trigger * 0.4: on = True
+        expo[i] = 1.0 if on else 0.4
+    return pd.Series(expo * rv, index=r.index)
+
+
+def both_daily():
+    naked = combo_daily()
+    return naked, apply_dd_guard(naked)
+
+
+def load():
+    return json.loads(STATE.read_text()) if STATE.exists() else None
+
+
+def save(st):
+    STATE_DIR.mkdir(parents=True, exist_ok=True)
+    STATE.write_text(json.dumps(st, indent=2))
+
+
+def advance():
+    naked, guard = both_daily()
+    st = load()
+    if st is None or "equity_g" not in st:                # init (o migrazione a doppia versione)
+        last = str(naked.index[-1])
+        st = dict(start=last, last=last, initial=INITIAL, n_days=0, w_crypto=W_CRYPTO, dd_trigger=DD_TRIGGER,
+                  equity=INITIAL, peak=INITIAL, max_dd=0.0,
+                  equity_g=INITIAL, peak_g=INITIAL, max_dd_g=0.0)
+        save(st); EQ.write_text("date,nudo,protetto\n" + f"{last},{INITIAL},{INITIAL}\n")
+        return st
+    last = pd.Timestamp(st["last"])
+    nn = naked[naked.index > last]; gg = guard[guard.index > last]
+    if len(nn):
+        e = st["equity"]; pk = st["peak"]; dd = st["max_dd"]
+        eg = st["equity_g"]; pkg = st["peak_g"]; ddg = st["max_dd_g"]; lines = []
+        for d in nn.index:
+            e *= (1 + float(nn[d])); pk = max(pk, e); dd = max(dd, (pk - e) / pk if pk > 0 else 0)
+            eg *= (1 + float(gg[d])); pkg = max(pkg, eg); ddg = max(ddg, (pkg - eg) / pkg if pkg > 0 else 0)
+            lines.append(f"{d},{e:.4f},{eg:.4f}")
+        st.update(equity=e, peak=pk, max_dd=dd, equity_g=eg, peak_g=pkg, max_dd_g=ddg,
+                  last=str(nn.index[-1]), n_days=st["n_days"] + len(nn))
+        save(st)
+        with open(EQ, "a") as f:
+            f.write("\n".join(lines) + "\n")
+    return st
+
+
+def main():
+    a = sys.argv[1:]
+    if "--reset" in a:
+        for f in (STATE, EQ):
+            f.unlink(missing_ok=True)
+        print("paper combo azzerato.")
+    st = load() if "--status" in a else advance()
+    if st is None or "equity_g" not in st:
+        st = advance()
+    days = (pd.Timestamp(st["last"]) - pd.Timestamp(st["start"])).days
+    rn = st["equity"] / st["initial"] - 1; rg = st["equity_g"] / st["initial"] - 1
+    gw = gtaa_weights(); asof = gw.pop("_asof", "?"); cash = gw.pop("_cash", None)
+    print("PAPER COMBO — TP01 (Deribit) + GTAA (IB), forward-only, blend 50/50")
+    print(f"  start {st['start'][:10]} -> last {st['last'][:10]} ({days}g, {st['n_days']} barre)")
+    print(f"  NUDO     : eq {st['equity']:.2f}  ret {rn*100:+.2f}%  maxDD {st['max_dd']*100:.1f}%")
+    print(f"  PROTETTO : eq {st['equity_g']:.2f}  ret {rg*100:+.2f}%  maxDD {st['max_dd_g']*100:.1f}%  (guardia-DD -{st.get('dd_trigger',DD_TRIGGER)*100:.0f}%)")
+    print(f"  --- POSIZIONI AZIONABILI ---")
+    print(f"  TP01 (Deribit): {_tp01_positions()}")
+    print(f"  GTAA (IB, asof {asof}): " + ", ".join(f"{k} {v:.0%}" for k, v in gw.items() if v) + f"  | cash {cash:.0%}")
+
+
+if __name__ == "__main__":
+    main()

scripts/portfolio/prevday_overlay.py

@@ -0,0 +1,115 @@
+"""SIMULAZIONE — PREVDAY come overlay di tail-hedge sul portafoglio attivo (NON deploy).
+
+PREVDAY (src/strategies/prevday_breakout) resta in FORWARD-MONITOR. Qui misuriamo SOLO, in
+simulazione, cosa farebbe al portafoglio live (TP01 55% + XS01 25% + VRP01 20%) aggiungerlo come
+overlay a peso W, riscalando i tre sleeve esistenti a (1-W) e tenendo le loro proporzioni. La
+trilogia (fill-haircut/turnover/bootstrap) ha stabilito che PREVDAY e' un HEDGE di regime-down
+(tutto il valore = gamba short) eseguibile a taglia reale: l'overlay si giudica sul TAGLIO DEL
+DRAWDOWN del portafoglio, non sul ritorno.
+
+NB outer-join: PREVDAY parte dal 2018, XS01 dal 2024, VRP01 dal 2021. I pesi sono rinormalizzati
+ogni giorno fra i soli sleeve con dato -> nel 2019-20 (solo TP01+PREVDAY) PREVDAY pesa di piu' del
+target W; nell'hold-out 2025+ (tutti e 4 attivi) pesa esattamente ~W. Per questo l'HOLD-OUT e' il
+confronto piu' pulito a "peso 10%".
+
+    uv run python scripts/portfolio/prevday_overlay.py
+"""
+from __future__ import annotations
+import sys
+from pathlib import Path
+
+PROJECT_ROOT = Path(__file__).resolve().parents[2]
+sys.path.insert(0, str(PROJECT_ROOT))
+import numpy as np
+import pandas as pd
+
+from src.backtest.harness import load
+from src.strategies import prevday_breakout as pb
+from src.portfolio.portfolio import StrategyPortfolio, Sleeve, metrics, HOLDOUT
+from src.portfolio.sleeves import _tp01_returns, _xsec_returns, _vrp_combo_returns
+
+ASSETS = ("BTC", "ETH")
+FEE_SIDE = 0.0005
+BASE_W = dict(TP01=0.55, XS01=0.25, VRP01=0.20)   # proporzioni dei tre sleeve attivi
+HEADLINE = 0.10
+
+
+def _prevday_returns() -> pd.Series:
+    """Rendimenti netti per-barra (1h) del libro PREVDAY 50/50 BTC+ETH (parametri congelati)."""
+    series = {}
+    for a in ASSETS:
+        df = load(a, "1h").reset_index(drop=True)
+        c = df["close"].values.astype(float)
+        r = np.zeros(len(c)); r[1:] = c[1:] / c[:-1] - 1.0
+        tgt = np.nan_to_num(pb.target(df), nan=0.0)
+        held = np.zeros(len(tgt)); held[1:] = tgt[:-1]
+        net = held * r - FEE_SIDE * np.abs(np.diff(tgt, prepend=tgt[0]))
+        series[a] = pd.Series(net, index=pd.to_datetime(df["datetime"], utc=True))
+    J = pd.concat(series, axis=1, join="inner").fillna(0.0)
+    return pd.Series(0.5 * J.iloc[:, 0].values + 0.5 * J.iloc[:, 1].values, index=J.index)
+
+
+def build_portfolio(series_cache: dict, w_prev: float) -> StrategyPortfolio:
+    """Portafoglio coi 3 sleeve riscalati a (1-w_prev) + PREVDAY a w_prev (0 = baseline)."""
+    sl = [
+        Sleeve("TP01_trend_1d", BASE_W["TP01"] * (1 - w_prev), lambda s=series_cache["TP01"]: s),
+        Sleeve("XS01_xsec_hl", BASE_W["XS01"] * (1 - w_prev), lambda s=series_cache["XS01"]: s),
+        Sleeve("VRP01_shortvol", BASE_W["VRP01"] * (1 - w_prev), lambda s=series_cache["VRP01"]: s),
+    ]
+    if w_prev > 0:
+        sl.append(Sleeve("PREVDAY_hedge", w_prev, lambda s=series_cache["PREVDAY"]: s))
+    return StrategyPortfolio(sl)
+
+
+def line(label, m):
+    return (f"  {label:<22s} Sh {m['sharpe']:>5.2f} | ret {m['ret']*100:>+8.1f}% "
+            f"CAGR {m['cagr']*100:>+6.1f}% | DD {m['maxdd']*100:>5.1f}% | n {m['n']}")
+
+
+def main():
+    print("=" * 92)
+    print("  PREVDAY OVERLAY (simulazione, NON deploy) — tail-hedge sul portafoglio TP01+XS01+VRP01")
+    print("=" * 92)
+    print("  Precalcolo sleeve...", flush=True)
+    cache = dict(TP01=_tp01_returns(), XS01=_xsec_returns(),
+                 VRP01=_vrp_combo_returns(), PREVDAY=_prevday_returns())
+
+    print(f"\n  PREVDAY standalone (per riferimento):")
+    from src.portfolio.portfolio import to_daily
+    pvd = to_daily(cache["PREVDAY"])
+    print(line("PREVDAY full", metrics(pvd)))
+    print(line("PREVDAY hold-out", metrics(pvd[pvd.index >= HOLDOUT])))
+
+    print(f"\n  SWEEP PESO OVERLAY (FULL | HOLD-OUT) — headline {HEADLINE*100:.0f}%:")
+    print(f"  {'peso PREVDAY':<14s} {'FULL Sharpe':>11s} {'FULL DD':>9s} | {'HOLD Sharpe':>11s} {'HOLD DD':>9s} {'HOLD ret':>9s}")
+    rows = {}
+    for w in (0.0, 0.05, 0.10, 0.15, 0.20):
+        bt = build_portfolio(cache, w).backtest()
+        rows[w] = bt
+        tag = "BASELINE" if w == 0 else f"{w*100:.0f}%"
+        star = "  <-- headline" if abs(w - HEADLINE) < 1e-9 else ""
+        print(f"  {tag:<14s} {bt['full']['sharpe']:>11.2f} {bt['full']['maxdd']*100:>8.1f}% | "
+              f"{bt['holdout']['sharpe']:>11.2f} {bt['holdout']['maxdd']*100:>8.1f}% "
+              f"{bt['holdout']['ret']*100:>+8.1f}%{star}")
+
+    base, ov = rows[0.0], rows[HEADLINE]
+    print(f"\n  DETTAGLIO a {HEADLINE*100:.0f}% vs BASELINE:")
+    print(line("BASELINE   FULL", base['full']));   print(line(f"OVERLAY{HEADLINE*100:.0f}% FULL", ov['full']))
+    print(line("BASELINE   HOLD", base['holdout'])); print(line(f"OVERLAY{HEADLINE*100:.0f}% HOLD", ov['holdout']))
+    dSh = ov['holdout']['sharpe'] - base['holdout']['sharpe']
+    dDD = (ov['holdout']['maxdd'] - base['holdout']['maxdd']) * 100
+    print(f"\n  >> HOLD-OUT: ΔSharpe {dSh:+.2f}  |  ΔmaxDD {dDD:+.1f}pp  "
+          f"(tail-hedge = ci aspettiamo DD piu' basso)")
+
+    print(f"\n  PER ANNO (baseline -> overlay {HEADLINE*100:.0f}%): ret%  /  DD%")
+    yb, yo = base['yearly'], ov['yearly']
+    for y in sorted(set(yb) | set(yo)):
+        b = yb.get(y, {}); o = yo.get(y, {})
+        print(f"    {y}: ret {b.get('ret',0)*100:>+7.1f}% -> {o.get('ret',0)*100:>+7.1f}%   "
+              f"DD {b.get('dd',0)*100:>5.1f}% -> {o.get('dd',0)*100:>5.1f}%")
+    print("=" * 92)
+    print("  Nota: PREVDAY resta FORWARD-MONITOR. Questa e' una simulazione di impatto, non un deploy.")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/combo_leverage_sim.py

@@ -0,0 +1,74 @@
+"""Simulazione LEVA 1x/2x/3x su COMBO (TP01+GTAA) e TP01-solo, da $2k e $5k.
+
+Leva modellata onestamente: ritorno_giorno = L*r - (L-1)*financing/252 (costo del nozionale preso a
+prestito ~8%/anno blended: perp funding crypto + margin IB). MaxDD calcolato sul PERCORSO LEVATO REALE
+(non scalato: il compounding peggiora il DD oltre ×L). Check RUINA/margin-call: se l'equity tocca la
+soglia di liquidazione (perdita cumulata >= 1/L del picco -> margin call).
+
+CLAUDE.md: la leva NON e' la scorciatoia; raddoppia (e oltre) il drawdown. Caso base = 1x.
+"""
+import sys
+from pathlib import Path
+import numpy as np, pandas as pd
+ROOT = Path(__file__).resolve().parents[2]
+sys.path.insert(0, str(ROOT)); sys.path.insert(0, str(ROOT / "scripts" / "research"))
+from combo_yearly_report import combo_daily
+from src.portfolio.sleeves import _tp01_returns
+
+FIN = 0.08   # costo finanziamento annuo sul nozionale preso a prestito (perp funding + margin IB), blended
+
+
+def lever(ret: pd.Series, L: float) -> pd.Series:
+    return L * ret - (L - 1) * FIN / 252.0
+
+
+def analyze(ret: pd.Series, L: float, cap0: float):
+    r = lever(ret.dropna().sort_index(), L)
+    curve = cap0 * np.cumprod(1 + r.values)
+    peak = np.maximum.accumulate(curve)
+    dd = (peak - curve) / peak
+    maxdd = float(np.max(dd))
+    # margin call: perdita dal picco >= 1/L (a leva L, un drawdown del sottostante di 1/L azzera il margine)
+    ruin = bool(np.any(dd >= 1.0 / L - 1e-9)) if L > 1 else False
+    yrs = (r.index[-1] - r.index[0]).days / 365.25
+    cagr = (curve[-1] / cap0) ** (1 / yrs) - 1 if yrs > 0 and curve[-1] > 0 else -1
+    sh = float(r.mean() / r.std() * np.sqrt(252)) if r.std() > 0 else 0
+    worst_y = min((np.prod(1 + r[r.index.year == y].values) - 1) for y in sorted(set(r.index.year)))
+    return dict(L=L, final=float(curve[-1]), cagr=cagr, maxdd=maxdd, sharpe=sh, ruin=ruin,
+                worst_y=float(worst_y), perday=(curve[-1] - cap0) / yrs / 365)
+
+
+def main():
+    print("=" * 92)
+    print("  LEVA su COMBO vs TP01-solo — percorso reale (fin 8%/anno sul prestito), 2019-2026 (~7.3y)")
+    print("=" * 92)
+    strat = {"COMBO TP01+GTAA": combo_daily(), "TP01 solo (crypto)": _tp01_returns()}
+    for nm, r in strat.items():
+        if r.index.tz is None:
+            r.index = r.index.tz_localize("UTC")
+    for cap0 in (2000.0, 5000.0):
+        print(f"\n  ##### capitale iniziale ${cap0:,.0f} #####")
+        print(f"  {'strategia':20}{'leva':>5}{'CAGR':>8}{'MaxDD':>8}{'Sharpe':>8}{'pegg.anno':>10}{'$/giorno':>10}{'eq fine':>12}{'  RUINA?':>9}")
+        for nm, r in strat.items():
+            for L in (1, 2, 3):
+                a = analyze(r, L, cap0)
+                flag = "MARGIN-CALL" if a["ruin"] else "ok"
+                print(f"  {nm:20}{L:>4}x{a['cagr']*100:>7.1f}%{a['maxdd']*100:>7.1f}%{a['sharpe']:>8.2f}"
+                      f"{a['worst_y']*100:>9.1f}%{('$'+format(a['perday'],',.0f')):>10}{('$'+format(a['final'],',.0f')):>12}{flag:>11}")
+    # per-anno del COMBO a 2x e 3x da 2k (dettaglio)
+    print(f"\n  ##### COMBO per-anno a leva, da $2.000 #####")
+    cd = combo_daily()
+    for L in (2, 3):
+        print(f"\n  --- COMBO {L}x ---")
+        print(f"  {'anno':6}{'PnL %':>9}{'MaxDD %':>9}{'eq fine':>11}")
+        eq = 2000.0; r = lever(cd, L)
+        for y in sorted(set(r.index.year)):
+            ry = r[r.index.year == y]
+            if len(ry) < 5: continue
+            eq0 = eq; curve = eq0 * np.cumprod(1 + ry.values); peak = np.maximum.accumulate(curve)
+            ddp = float(np.max((peak - curve) / peak)); eq = float(curve[-1])
+            print(f"  {y:<6}{(eq/eq0-1)*100:>+8.1f}%{ddp*100:>8.1f}%{('$'+format(eq,',.0f')):>11}")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/combo_protected_yearly.py

@@ -0,0 +1,90 @@
+"""Resoconto anno-per-anno CON PROTEZIONE (soft-guard DD -4%) — combo e singoli.
+
+Mette combo (TP01+GTAA 50/50), TP01 e GTAA sulla STESSA griglia giorni-di-borsa (come dentro al
+combo), applica la guardia-DD -4% a ciascuna serie (de-risk 1.0->0.4 a -4% dal picco, ri-rischia a
+-1.6%), e per ogni anno riporta: NL (net liquidation da $2000), DD intra-anno, rendimento (=CAGR
+1y), Sharpe. Riga TOT con CAGR e Sharpe complessivi.
+"""
+import sys
+from pathlib import Path
+import numpy as np, pandas as pd
+
+ROOT = Path(__file__).resolve().parents[2]
+sys.path.insert(0, str(ROOT))
+from src.portfolio.sleeves import _tp01_returns
+from src.portfolio.gtaa import gtaa_returns
+
+INITIAL = 2000.0
+ANN = np.sqrt(252.0)
+DD_TRIG = 0.04
+
+
+def dd_guard(r, dd_trigger=DD_TRIG):
+    """De-risk: esposizione 1.0->0.4 se DD da picco > dd_trigger; ri-rischia a dd_trigger*0.4."""
+    r = r.values; n = len(r); eq = np.cumprod(1 + r); pk = np.maximum.accumulate(eq)
+    expo = np.ones(n); on = True
+    for i in range(1, n):
+        ddi = (pk[i - 1] - eq[i - 1]) / pk[i - 1]
+        if ddi > dd_trigger: on = False
+        if ddi < dd_trigger * 0.4: on = True
+        expo[i] = 1.0 if on else 0.4
+    return pd.Series(expo * r, index=_idx)  # set below
+
+
+def legs_on_grid(wc=0.5):
+    """TP01(crypto, compoundato sul grid) e GTAA(equity) sulla stessa griglia giorni-di-borsa."""
+    tp = _tp01_returns()
+    if tp.index.tz is None:
+        tp.index = tp.index.tz_localize("UTC")
+    eq = gtaa_returns().dropna()
+    grid = eq.index[eq.index >= tp.index[0]]
+    cum = (1 + tp).cumprod()
+    tpg = cum.reindex(cum.index.union(grid)).ffill().reindex(grid).pct_change()
+    J = pd.concat({"c": tpg, "e": eq.reindex(grid)}, axis=1).dropna()
+    combo = wc * J["c"] + (1 - wc) * J["e"]
+    return combo, J["c"], J["e"]
+
+
+def sh(r): r = r.dropna().values; return float(np.mean(r) / np.std(r) * ANN) if len(r) > 5 and np.std(r) > 0 else 0.0
+def maxdd(curve): pk = np.maximum.accumulate(curve); return float(np.max((pk - curve) / pk)) if len(curve) else 0.0
+
+
+def yearly(ret, label):
+    ret = ret.dropna().sort_index()
+    print(f"\n  ===== {label} (guardia-DD -4%) =====")
+    print(f"  {'anno':6}{'NL inizio':>11}{'NL fine':>11}{'rend%':>9}{'DD%':>8}{'Sharpe':>9}")
+    eq = INITIAL
+    for y in sorted(set(ret.index.year)):
+        r = ret[ret.index.year == y]
+        if len(r) < 5: continue
+        eq0 = eq
+        curve = eq0 * np.cumprod(1 + r.values)
+        eq = float(curve[-1])
+        print(f"  {y:<6}{eq0:>11,.0f}{eq:>11,.0f}{(eq/eq0-1)*100:>+8.1f}%{maxdd(curve)*100:>7.1f}%{sh(r):>9.2f}")
+    yrs = (ret.index[-1] - ret.index[0]).days / 365.25
+    cagr = (eq / INITIAL) ** (1 / yrs) - 1 if yrs > 0 else 0
+    full_curve = INITIAL * np.cumprod(1 + ret.values)
+    print(f"  {'TOT':<6}{INITIAL:>11,.0f}{eq:>11,.0f}{(eq/INITIAL-1)*100:>+8.1f}%{maxdd(full_curve)*100:>7.1f}%{sh(ret):>9.2f}"
+          f"   |  CAGR {cagr*100:+.1f}%  ({yrs:.1f}y)")
+
+
+def main():
+    global _idx
+    print("=" * 78)
+    print("  RESOCONTO PROTETTO (soft-guard DD -4%) — da $2.000, anno per anno")
+    print("  Tutte e tre sulla griglia giorni-di-borsa del combo (dal 2019), esposizione 1x.")
+    print("=" * 78)
+    combo, tp, g = legs_on_grid()
+    for ret, lbl in [(combo, "COMBO TP01+GTAA 50/50"), (tp, "solo TP01 (crypto)"), (g, "solo GTAA (equity)")]:
+        _idx = ret.index
+        yearly(dd_guard(ret), lbl)
+    # confronto NON protetto (baseline) in coda, una riga TOT per riferimento
+    print("\n  --- riferimento NON protetto (baseline, TOT) ---")
+    for ret, lbl in [(combo, "COMBO"), (tp, "TP01"), (g, "GTAA")]:
+        yrs = (ret.index[-1] - ret.index[0]).days / 365.25
+        eqf = INITIAL * np.prod(1 + ret.values)
+        print(f"  {lbl:6} CAGR {((eqf/INITIAL)**(1/yrs)-1)*100:>+5.1f}%  DD {maxdd(INITIAL*np.cumprod(1+ret.values))*100:>4.1f}%  Sharpe {sh(ret):.2f}")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/combo_yearly_report.py

@@ -0,0 +1,99 @@
+"""Resoconto anno-per-anno della strategia combo (TP01+GTAA) + componenti, da $2.000.
+
+Per anno: PnL ($ e %), MaxDD (intra-anno), NumTrades, equity di fine anno (compounding da 2k).
+Combo = blend 50/50 TP01(Deribit) + GTAA(IB) (crypto compoundato su grid giorni-di-borsa).
+NumTrades: TP01 = cambi di target BTC/ETH (>0.05); GTAA = ribilanci MENSILI per-gamba (>2%).
+Onesto: il combo parte dal 2019 (crypto). GTAA-solo dato anche su 10y come contesto.
+"""
+import sys
+from pathlib import Path
+import numpy as np, pandas as pd
+
+ROOT = Path(__file__).resolve().parents[2]
+sys.path.insert(0, str(ROOT)); sys.path.insert(0, str(ROOT / "scripts" / "live"))
+from src.data.downloader import load_data
+from src.strategies.trend_portfolio import TrendPortfolio, CANONICAL, resample_1d
+from src.portfolio.sleeves import _tp01_returns
+from src.portfolio.gtaa import gtaa_returns, _exposure, _close, EQ_UNIVERSE
+INITIAL = 2000.0
+
+
+def tp01_trades_per_year():
+    tp = TrendPortfolio(**CANONICAL); cnt = {}
+    for a in ("BTC", "ETH"):
+        df = resample_1d(load_data(a, "1h")); tgt = tp.target_series(df)
+        idx = pd.DatetimeIndex(pd.to_datetime(df["datetime"]))
+        chg = pd.Series(np.abs(np.diff(tgt, prepend=tgt[0])) > 0.05, index=idx)
+        for y, c in chg.groupby(idx.year).sum().items():
+            cnt[int(y)] = cnt.get(int(y), 0) + int(c)
+    return cnt
+
+
+def gtaa_trades_per_year():
+    # pesi giornalieri -> ribilancio MENSILE realistico -> conta gambe cambiate >2%
+    W = {}
+    for a in EQ_UNIVERSE:
+        ex = _exposure(_close(a)) / len(EQ_UNIVERSE)
+        W[a] = ex
+    Wd = pd.concat(W, axis=1).dropna()
+    Wm = Wd.resample("ME").last()                       # peso a fine mese
+    chg = (Wm.diff().abs() > 0.02).sum(axis=1)          # gambe ribilanciate quel mese
+    return chg.groupby(chg.index.year).sum().astype(int).to_dict()
+
+
+def yearly(ret: pd.Series, trades: dict, label: str, start_capital=INITIAL):
+    ret = ret.dropna().sort_index()
+    print(f"\n  ===== {label} =====")
+    print(f"  {'anno':6}{'eq inizio':>12}{'PnL $':>12}{'PnL %':>9}{'MaxDD %':>9}{'NumTrades':>11}{'eq fine':>12}")
+    eq = start_capital
+    for y in sorted(set(ret.index.year)):
+        r = ret[ret.index.year == y]
+        if len(r) < 5:
+            continue
+        eq0 = eq
+        curve = eq0 * np.cumprod(1 + r.values)
+        peak = np.maximum.accumulate(curve)
+        dd = float(np.max((peak - curve) / peak)) if len(curve) else 0.0
+        eq = float(curve[-1])
+        pnl = eq - eq0
+        nt = trades.get(y, None)
+        print(f"  {y:<6}{eq0:>12,.0f}{pnl:>+12,.0f}{(eq/eq0-1)*100:>+8.1f}%{dd*100:>8.1f}%"
+              f"{(str(nt) if nt is not None else '—'):>11}{eq:>12,.0f}")
+    tot = eq / start_capital - 1
+    yrs = (ret.index[-1] - ret.index[0]).days / 365.25
+    cagr = (eq / start_capital) ** (1 / yrs) - 1 if yrs > 0 else 0
+    sh = float(r.mean()) if False else float(ret.mean() / ret.std() * np.sqrt(252))
+    print(f"  {'TOT':<6}{start_capital:>12,.0f}{eq-start_capital:>+12,.0f}{tot*100:>+8.1f}%"
+          f"{'':>9}{sum(v for v in trades.values()) if trades else 0:>11}{eq:>12,.0f}")
+    print(f"  -> da ${start_capital:,.0f} a ${eq:,.0f} in {yrs:.1f}y | CAGR {cagr*100:+.1f}% | Sharpe {sh:.2f}")
+
+
+def combo_daily(wc=0.5):
+    tp = _tp01_returns()
+    if tp.index.tz is None:
+        tp.index = tp.index.tz_localize("UTC")
+    eq = gtaa_returns().dropna()
+    grid = eq.index[eq.index >= tp.index[0]]
+    cum = (1 + tp).cumprod()
+    tpg = (cum.reindex(cum.index.union(grid)).ffill().reindex(grid)).pct_change()
+    J = pd.concat({"c": tpg, "e": eq.reindex(grid)}, axis=1).dropna()
+    return (wc * J["c"] + (1 - wc) * J["e"]).dropna()
+
+
+def main():
+    print("=" * 80)
+    print("  RESOCONTO STRATEGIA — da $2.000, anno per anno")
+    print("=" * 80)
+    tpt = tp01_trades_per_year(); gtt = gtaa_trades_per_year()
+    combo_tr = {y: tpt.get(y, 0) + gtt.get(y, 0) for y in set(tpt) | set(gtt)}
+    # COMBO (la strategia deployata)
+    yearly(combo_daily(), combo_tr, "COMBO TP01+GTAA 50/50 (deployabile, dal 2019)")
+    # componenti
+    tp = _tp01_returns(); tp.index = tp.index.tz_localize("UTC") if tp.index.tz is None else tp.index
+    yearly(tp, tpt, "solo TP01 (crypto, Deribit)")
+    g = gtaa_returns(); g10 = g[g.index >= (g.index[-1] - pd.Timedelta(days=3660))]
+    yearly(g10, gtt, "solo GTAA (equity, IB) — ULTIMI 10 ANNI")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/crypto_lead_harness.py

@@ -0,0 +1,126 @@
+"""HARNESS parametrizzato — anticipazione crypto -> mercato (lead-lag eseguibile, onesto).
+
+Generalizza l'effetto weekend: la finestra-LEAD e' l'intervallo in cui l'equity e' CHIUSO e la crypto
+no (prev close 21:00 UTC -> next open 13:30 UTC). Il weekend e' il caso lungo (Ven 21:00 -> Lun 13:30).
+Per ogni sessione equity D (con sessione precedente P):
+  lead = crypto return su [lead_start, D 13:00]   (lead_start = P 21:00 se hours='overnight', else D13:00-hours)
+  predict target: gap = open[D]/close[P]-1 ; intraday = close[D]/open[D]-1 ; full = close[D]/close[P]-1
+  control = rendimento sessione precedente equity (close[P]/close[P_prev]-1)  -> test INCREMENTALE
+Filtro giorni: all | mon (solo lunedi'/weekend) | nonmon.
+
+Output JSON per config: n, corr, beta+t-stat del lead AL NETTO del control (incrementale), Sharpe
+settimanale/annualizzato del trade eseguibile (sign(lead)*predict - costo) FULL/IS/OOS(2022+),
+hit-rate, e PER-ANNO (hit e mean) per la robustezza multi-anno.
+
+  uv run python scripts/research/crypto_lead_harness.py --configs '[{"lead":"BTC","target":"QQQ","day":"mon","predict":"intraday","hours":"overnight"}]'
+Dati: cache su disco (crypto 1h, ETF eq_*). Nessun IB online. Vettoriale, veloce.
+"""
+import sys, json, argparse
+from pathlib import Path
+import numpy as np, pandas as pd
+
+ROOT = Path(__file__).resolve().parents[2]
+sys.path.insert(0, str(ROOT)); sys.path.insert(0, str(ROOT / "scripts" / "research"))
+from src.data.downloader import load_data
+import eqlib
+
+OOS_DEFAULT = "2022-01-01"
+OPEN_H = 13   # ~apertura US 13:30 UTC -> uso barra 13:00 (info nota prima dell'open per il lead)
+CLOSE_H = 21  # ~chiusura US 21:00 UTC
+
+_CRYPTO = {}
+def crypto_hourly(asset):
+    if asset not in _CRYPTO:
+        s = load_data(asset, "1h").set_index("datetime")["close"].astype(float)
+        full = pd.date_range(s.index[0].floor("h"), s.index[-1].ceil("h"), freq="h", tz="UTC")
+        _CRYPTO[asset] = s.reindex(s.index.union(full)).ffill().reindex(full)
+    return _CRYPTO[asset]
+
+
+def at(series, ts):
+    try:
+        return float(series.asof(ts))
+    except Exception:
+        return np.nan
+
+
+def evaluate(cfg, cost_rt=0.0004, oos=OOS_DEFAULT):
+    OOS = pd.Timestamp(oos, tz="UTC")
+    lead = cfg["lead"]; tgt = cfg["target"]; day = cfg.get("day", "all")
+    predict = cfg.get("predict", "intraday"); hours = cfg.get("hours", "overnight")
+    bc = crypto_hourly(lead)
+    try:
+        oc = eqlib.load_eq(tgt)["open"].astype(float); cc = eqlib.load_eq(tgt)["close"].astype(float)
+    except Exception as e:
+        return {**cfg, "err": f"no data {tgt}"}
+    idx = cc.index
+    rows = []
+    for j in range(2, len(idx)):
+        D = idx[j]; P = idx[j-1]; Pp = idx[j-2]
+        if day == "mon" and D.weekday() != 0: continue
+        if day == "nonmon" and D.weekday() == 0: continue
+        d_open = D.normalize() + pd.Timedelta(hours=OPEN_H)
+        p_close = P.normalize() + pd.Timedelta(hours=CLOSE_H)
+        lead_start = p_close if hours == "overnight" else d_open - pd.Timedelta(hours=int(hours))
+        c1 = at(bc, d_open); c0 = at(bc, lead_start)
+        if not (np.isfinite(c1) and np.isfinite(c0) and c0 > 0): continue
+        ld = c1 / c0 - 1.0
+        gap = oc[D] / cc[P] - 1.0
+        intr = cc[D] / oc[D] - 1.0
+        full = cc[D] / cc[P] - 1.0
+        ctrl = cc[P] / cc[Pp] - 1.0
+        rows.append((D, ld, gap, intr, full, ctrl))
+    if len(rows) < 60:
+        return {**cfg, "err": f"n={len(rows)}"}
+    D_ = pd.DataFrame(rows, columns=["d", "lead", "gap", "intraday", "full", "ctrl"]).set_index("d")
+    y = D_[predict].values; x = D_["lead"].values; ctrl = D_["ctrl"].values
+
+    def z(a):
+        sd = a.std(); return (a - a.mean()) / sd if sd > 0 else a * 0
+    corr = float(np.corrcoef(x, y)[0, 1])
+    # incrementale vs control (OLS standardizzato)
+    X = np.column_stack([np.ones(len(y)), z(x), z(ctrl)])
+    beta, *_ = np.linalg.lstsq(X, z(y), rcond=None)
+    resid = z(y) - X @ beta
+    dof = max(len(y) - 3, 1)
+    se = np.sqrt(np.sum(resid**2) / dof * np.diag(np.linalg.inv(X.T @ X)))
+    t_inc = float(beta[1] / se[1]) if se[1] > 0 else 0.0
+    # trade eseguibile: long-short e long-flat su segno del lead, intraday/predict, net costi
+    sign = np.sign(x)
+    def sharpe(r):
+        r = r[np.isfinite(r)]
+        return float(np.mean(r) / np.std(r) * np.sqrt(52)) if len(r) > 5 and np.std(r) > 0 else 0.0
+    ls = sign * y - cost_rt
+    lf = np.where(x > 0, y, 0.0) - np.where(x > 0, cost_rt, 0.0)
+    yrs = D_.index.year.values
+    def per_year(r):
+        out = {}
+        for yv in sorted(set(yrs)):
+            m = yrs == yv
+            if m.sum() >= 8:
+                out[int(yv)] = round(float(np.mean(np.sign(x[m]) == np.sign(y[m]))), 2)
+        return out
+    is_m = D_.index < OOS; oos_m = D_.index >= OOS
+    py = per_year(ls)
+    return {**cfg, "n": len(D_), "corr": round(corr, 3), "t_incremental": round(t_inc, 2),
+            "hit": round(float(np.mean(sign == np.sign(y))), 3),
+            "sh_ls_full": round(sharpe(ls), 2), "sh_ls_is": round(sharpe(ls[is_m]), 2),
+            "sh_ls_oos": round(sharpe(ls[oos_m]), 2),
+            "sh_lf_full": round(sharpe(lf), 2), "sh_lf_oos": round(sharpe(lf[oos_m]), 2),
+            "ann_ls_pct": round(float(np.nanmean(ls) * 52 * 100), 1),
+            "years_pos": int(sum(1 for v in py.values() if v > 0.5)), "years_tot": len(py),
+            "per_year_hit": py}
+
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument("--configs", required=True)
+    ap.add_argument("--cost", type=float, default=0.0004)
+    ap.add_argument("--oos", default=OOS_DEFAULT)
+    args = ap.parse_args()
+    cfgs = json.loads(args.configs)
+    print(json.dumps([evaluate(c, cost_rt=args.cost, oos=args.oos) for c in cfgs]))
+
+
+if __name__ == "__main__":
+    main()

scripts/research/crypto_macro_leadlag.py

@@ -0,0 +1,120 @@
+"""CRYPTO x MERCATI IB — correlazioni e ANTICIPAZIONI (lead-lag).
+
+Obiettivo: la crypto (24/7) anticipa i mercati IB (azioni/bond/oro/credito), o viceversa?
+Disciplina onesta: i tranelli di timing daily sono enormi (crypto chiude 00:00 UTC, US equity 21:00
+UTC -> il lag-0 e' contaminato), quindi (1) allineo i rendimenti sullo STESSO intervallo
+(compounding crypto sul grid giorni-di-borsa), (2) guardo i lag >=1 giorno, (3) test del segno con
+hit-rate e split in-sample/OOS, (4) flag multiple-testing.
+
+Ipotesi piu' pulita = EFFETTO WEEKEND: la crypto si muove Sab+Dom (azionario chiuso) -> quel
+movimento e' informazione PRIOR al lunedi'. Predice il gap/intraday del lunedi' azionario?
+Uso gli OPEN dei parquet eq_ (Monday open noto alle 13:30 UTC, weekend crypto noto alle 00:00 UTC).
+
+DATI: cache su disco (BTC/ETH Deribit 1h->1d UTC; ETF IB eq_*). Nessun IB online.
+"""
+import sys
+from pathlib import Path
+import numpy as np, pandas as pd
+
+ROOT = Path(__file__).resolve().parents[2]
+sys.path.insert(0, str(ROOT)); sys.path.insert(0, str(ROOT / "scripts" / "research"))
+from src.data.downloader import load_data
+import eqlib
+
+ETFS = ["SPY", "QQQ", "IWM", "TLT", "GLD", "HYG"]
+
+
+def crypto_daily_close(asset="BTC") -> pd.Series:
+    df = load_data(asset, "1h").set_index("datetime")["close"].astype(float)
+    return df.resample("1D").last().dropna()   # close ~00:00 UTC del giorno dopo
+
+
+def _ret(s):
+    return s.pct_change()
+
+
+def _corr_lags(x: pd.Series, y: pd.Series, lags=range(-5, 6)):
+    """corr(x_{t-k}, y_t): k>0 => x ANTICIPA y di k giorni. Allineati sullo stesso grid."""
+    J = pd.concat({"x": x, "y": y}, axis=1, join="inner").dropna()
+    out = {}
+    for k in lags:
+        out[k] = round(float(J["x"].shift(k).corr(J["y"])), 3)
+    return out
+
+
+def main():
+    print("=" * 98)
+    print("  CRYPTO x MERCATI IB — correlazioni & anticipazioni (lead-lag)")
+    print("=" * 98)
+    btc = crypto_daily_close("BTC"); eth = crypto_daily_close("ETH")
+    btc_r = _ret(btc); eth_r = _ret(eth)
+    # equity close + grid giorni-di-borsa
+    eq_close = {s: eqlib.load_eq(s)["close"].astype(float) for s in ETFS}
+    eq_open = {s: eqlib.load_eq(s)["open"].astype(float) for s in ETFS}
+    grid = eq_close["SPY"].index
+    grid = grid[grid >= btc.index[0]]
+    # crypto compoundato sul grid giorni-di-borsa (stesso intervallo dell'equity ret)
+    def to_grid(s):
+        cum = (1 + _ret(s)).cumprod()
+        return (cum.reindex(cum.index.union(grid)).ffill().reindex(grid)).pct_change()
+    btc_g = to_grid(btc); eth_g = to_grid(eth)
+
+    print(f"  overlap dal {grid[0].date()} ({len(grid)} giorni di borsa)\n")
+
+    print("  --- (1) CORRELAZIONE CONTEMPORANEA (stesso intervallo; lag0 contaminato da timing) ---")
+    print(f"    {'ETF':5} {'corr BTC':>9} {'corr ETH':>9}")
+    for s in ETFS:
+        er = _ret(eq_close[s]).reindex(grid)
+        cb = round(float(pd.concat([btc_g, er], axis=1).dropna().corr().iloc[0, 1]), 3)
+        ce = round(float(pd.concat([eth_g, er], axis=1).dropna().corr().iloc[0, 1]), 3)
+        print(f"    {s:5} {cb:>9} {ce:>9}")
+
+    print("\n  --- (2) LEAD-LAG BTC vs ETF: corr(BTC_{t-k}, ETF_t), k>0 = BTC ANTICIPA ---")
+    print(f"    {'ETF':5} " + " ".join(f"k{ k:+d}" for k in range(-3, 4)) + "   picco")
+    for s in ETFS:
+        er = _ret(eq_close[s]).reindex(grid)
+        cl = _corr_lags(btc_g, er, range(-3, 4))
+        peak = max(cl, key=lambda k: abs(cl[k]))
+        row = " ".join(f"{cl[k]:+.2f}" for k in range(-3, 4))
+        tag = f"k={peak:+d} ({'BTC->ETF' if peak>0 else 'ETF->BTC' if peak<0 else 'contemp'})"
+        print(f"    {s:5} {row}   {tag}")
+
+    print("\n  --- (3) EFFETTO WEEKEND: crypto Sab+Dom -> lunedi' azionario (anticipazione pulita) ---")
+    # weekend crypto = close(Dom 00:00 lun) / close(Ven) - 1 ; calcolato su crypto daily (calendario)
+    cal = pd.date_range(btc.index[0], btc.index[-1], freq="D", tz="UTC")
+    bc = btc.reindex(cal).ffill()
+    for s in ["SPY", "QQQ", "IWM", "HYG"]:
+        oc = eq_open[s]; cc = eq_close[s]
+        rows = []
+        for mon in grid:
+            if mon.weekday() != 0:      # solo lunedi'
+                continue
+            fri = mon - pd.Timedelta(days=3)
+            if fri not in cc.index:      # venerdi' non di borsa (festa) -> salta
+                continue
+            wk = float(bc.get(mon, np.nan) / bc.get(fri + pd.Timedelta(days=0), np.nan) - 1) if fri in bc.index else np.nan
+            # weekend crypto: da venerdi 00:00(close ven) a lunedi 00:00 -> usa bc[fri]..bc[mon]
+            wk = float(bc.loc[mon] / bc.loc[fri] - 1) if (mon in bc.index and fri in bc.index) else np.nan
+            gap = float(oc.loc[mon] / cc.loc[fri] - 1) if (mon in oc.index and fri in cc.index) else np.nan
+            intr = float(cc.loc[mon] / oc.loc[mon] - 1) if mon in oc.index else np.nan
+            rows.append((mon, wk, gap, intr))
+        D = pd.DataFrame(rows, columns=["mon", "wk", "gap", "intr"]).dropna().set_index("mon")
+        if len(D) < 50:
+            print(f"    {s}: pochi dati ({len(D)})"); continue
+        def stat(col):
+            c = float(D["wk"].corr(D[col]))
+            hit = float((np.sign(D["wk"]) == np.sign(D[col])).mean())
+            return c, hit
+        cg, hg = stat("gap"); ci, hi = stat("intr")
+        # OOS: split 2022
+        Dh = D[D.index >= pd.Timestamp("2022-01-01", tz="UTC")]
+        cg_o = float(Dh["wk"].corr(Dh["gap"])); ci_o = float(Dh["wk"].corr(Dh["intr"]))
+        print(f"    {s}: n={len(D)} | weekend-crypto -> Mon GAP corr {cg:+.2f} hit {hg*100:.0f}% (OOS22+ {cg_o:+.2f}) "
+              f"| Mon INTRADAY corr {ci:+.2f} hit {hi*100:.0f}% (OOS {ci_o:+.2f})")
+
+    print("\n  NB: lag-0/contemporanea contaminata dal timing (crypto chiude 00:00, equity 21:00 UTC).")
+    print("      Il GAP del lunedi' e' il test pulito (weekend crypto = info prior all'apertura).")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/crypto_overnight_equity.py

@@ -0,0 +1,88 @@
+"""DERIBIT-monitor -> IB-trade: il segnale crypto overnight AGGIUNGE al long-overnight indice?
+
+Idea utente: guardare crypto live su Deribit (24/7) e tradare l'indice su IB. Il GAP di apertura =
+movimento OVERNIGHT dei futures (MES/MNQ/M2K, tradati di notte) -> catturabile, net ~2bps.
+
+DOMANDA DECISIVA (test onesto): l'azionario ha gia' un OVERNIGHT PREMIUM noto (il drift positivo
+notturno). Quindi "long indice overnight" rende di suo. Il segnale crypto MIGLIORA quel baseline,
+o e' solo overnight-premium + beta? Confronto:
+  A) ALWAYS-LONG overnight (incondizionato) = cattura il premio notturno puro
+  B) LONG se crypto-overnight>0, else FLAT = usa il crypto come filtro
+  C) LONG/SHORT sul segno del crypto = il segnale pieno
+La metrica chiave e' B,C VS A (l'uplift del crypto sul puro premio notturno), non A in assoluto.
+
+Dati: cache su disco (crypto 1h Deribit; ETF eq_* = proxy del future indice). net 2bps RT (futures).
+"""
+import sys
+from pathlib import Path
+import numpy as np, pandas as pd
+
+ROOT = Path(__file__).resolve().parents[2]
+sys.path.insert(0, str(ROOT)); sys.path.insert(0, str(ROOT / "scripts" / "research"))
+import eqlib
+from crypto_lead_harness import crypto_hourly, at, OPEN_H, CLOSE_H
+
+OOS = pd.Timestamp("2022-01-01", tz="UTC")
+COST = 0.0002   # ~2bps RT micro-future (commissione + spread)
+
+
+def _sh(r):
+    r = np.asarray(r, float); r = r[np.isfinite(r)]
+    return float(np.mean(r) / np.std(r) * np.sqrt(252)) if len(r) > 5 and np.std(r) > 0 else 0.0
+
+
+def build(target="SPY", lead="BTC"):
+    bc = crypto_hourly(lead)
+    oc = eqlib.load_eq(target)["open"].astype(float); cc = eqlib.load_eq(target)["close"].astype(float)
+    idx = cc.index; rows = []
+    for j in range(1, len(idx)):
+        D = idx[j]; P = idx[j-1]
+        d_open = D.normalize() + pd.Timedelta(hours=OPEN_H)
+        p_close = P.normalize() + pd.Timedelta(hours=CLOSE_H)
+        c1 = at(bc, d_open); c0 = at(bc, p_close)
+        if not (np.isfinite(c1) and np.isfinite(c0) and c0 > 0):
+            continue
+        crypto = c1 / c0 - 1.0
+        overnight = oc[D] / cc[P] - 1.0       # gap = ritorno overnight del future indice (catturabile)
+        rows.append((D, crypto, overnight))
+    return pd.DataFrame(rows, columns=["d", "crypto", "on"]).set_index("d")
+
+
+def main():
+    print("=" * 92)
+    print("  CRYPTO overnight -> LONG INDICE OVERNIGHT (Deribit-monitor / IB-trade): aggiunge al premio?")
+    print("=" * 92)
+    print(f"  net {COST*1e4:.0f}bps RT (micro-future). overnight = open[D]/close[P]-1 (= move notturno future).\n")
+    for tgt in ("SPY", "QQQ", "IWM"):
+        for lead in ("BTC", "ETH"):
+            D = build(tgt, lead)
+            up = D[D["crypto"] > 0]["on"]; dn = D[D["crypto"] <= 0]["on"]
+            # strategie
+            A = D["on"].values - COST                                   # always-long
+            B = np.where(D["crypto"] > 0, D["on"], 0.0) - np.where(D["crypto"] > 0, COST, 0.0)  # long se crypto su
+            C = np.sign(D["crypto"]) * D["on"] - COST                   # long/short
+            shA, shB, shC = _sh(A), _sh(B), _sh(C)
+            # OOS
+            m = D.index >= OOS
+            print(f"  {tgt} <- {lead}: n={len(D)} | notti crypto-SU {len(up)} ret medio {up.mean()*1e4:+.1f}bps "
+                  f"| crypto-GIU {len(dn)} ret medio {dn.mean()*1e4:+.1f}bps  (spread {(up.mean()-dn.mean())*1e4:+.1f}bps)")
+            print(f"        Sharpe: A always-long {shA:.2f} | B long-if-cryptoUp {shB:.2f} | C long/short {shC:.2f}  "
+                  f"-> UPLIFT crypto B-A {shB-shA:+.2f}, C-A {shC-shA:+.2f}")
+            print(f"        OOS22+: A {_sh(A[m]):.2f} | B {_sh(B[m]):.2f} | C {_sh(C[m]):.2f} | "
+                  f"ann: A {np.nanmean(A)*252*100:+.1f}% B {np.nanmean(B)*252*100:+.1f}% C {np.nanmean(C)*252*100:+.1f}%")
+        print()
+
+    # focus SPY-BTC: per-anno A vs C, e sketch deploy
+    D = build("SPY", "BTC")
+    A = D["on"].values - COST; C = np.sign(D["crypto"]) * D["on"] - COST
+    print("  --- SPY<-BTC per-anno: Sharpe A(always-long) vs C(crypto long/short) ---")
+    for y in sorted(set(D.index.year)):
+        mm = D.index.year == y
+        if mm.sum() >= 40:
+            print(f"    {y}: A {_sh(A[mm]):+.2f}  C {_sh(C[mm]):+.2f}  (n={mm.sum()})")
+    print("\n  NB: se C-A ~ 0, il crypto NON aggiunge al premio overnight (e' solo il premio + beta).")
+    print("      Se C-A >> 0 e A gia' alto, il crypto e' un filtro REALE sul rischio notturno.")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/crypto_overnight_honest.py

@@ -0,0 +1,71 @@
+"""ONESTA' DI TIMING: il 'Sharpe 5' del crypto->overnight equity e' look-ahead?
+
+Test: separare la versione OVERLAP (segnale e ritorno coprono le stesse ore = look-ahead) dalla
+versione TRADABILE (segnale finisce all'ENTRATA, catturo solo il moto SUCCESSIVO).
+  segnale crypto = BTC [P 21:00 -> D 13:00 UTC]  (noto solo a D 13:00, poco prima dell'open 13:30)
+  - OVERLAP   capture = gap   = open[D]/close[P]-1   [P 21:00 -> D 13:30]  <-- sovrappone il segnale
+  - TRADABILE capture = intra = close[D]/open[D]-1   [D 13:30 -> D 21:00]  <-- DOPO l'entrata, pulito
+Se OVERLAP >> TRADABILE, il numero alto e' artefatto: all'open il gap e' gia' avvenuto, non lo catturi.
+Annualizzazione CORRETTA sqrt(252) (giornaliero), non sqrt(52).
+"""
+import sys
+from pathlib import Path
+import numpy as np, pandas as pd
+
+ROOT = Path(__file__).resolve().parents[2]
+sys.path.insert(0, str(ROOT)); sys.path.insert(0, str(ROOT / "scripts" / "research"))
+import eqlib
+from crypto_lead_harness import crypto_hourly, at, OPEN_H, CLOSE_H
+
+OOS = pd.Timestamp("2022-01-01", tz="UTC"); COST = 0.0002
+
+
+def _sh(r):
+    r = np.asarray(r, float); r = r[np.isfinite(r)]
+    return float(np.mean(r) / np.std(r) * np.sqrt(252)) if len(r) > 5 and np.std(r) > 0 else 0.0
+
+
+def build(target="SPY", lead="BTC"):
+    bc = crypto_hourly(lead)
+    oc = eqlib.load_eq(target)["open"].astype(float); cc = eqlib.load_eq(target)["close"].astype(float)
+    idx = cc.index; rows = []
+    for j in range(1, len(idx)):
+        D = idx[j]; P = idx[j-1]
+        c1 = at(bc, D.normalize() + pd.Timedelta(hours=OPEN_H))      # crypto a D 13:00 (entrata)
+        c0 = at(bc, P.normalize() + pd.Timedelta(hours=CLOSE_H))     # crypto a P 21:00
+        if not (np.isfinite(c1) and np.isfinite(c0) and c0 > 0):
+            continue
+        crypto = c1 / c0 - 1.0
+        gap = oc[D] / cc[P] - 1.0          # OVERLAP col segnale
+        intra = cc[D] / oc[D] - 1.0        # DOPO l'entrata (tradabile)
+        rows.append((D, crypto, gap, intra))
+    return pd.DataFrame(rows, columns=["d", "crypto", "gap", "intra"]).set_index("d")
+
+
+def main():
+    print("=" * 90)
+    print("  TIMING ONESTO: OVERLAP (look-ahead) vs TRADABILE — crypto overnight -> equity")
+    print("=" * 90)
+    print("  segnale noto a D13:00. OVERLAP=gap [copre il segnale]. TRADABILE=intraday [dopo entrata]. net 2bps, sqrt(252)\n")
+    for tgt in ("SPY", "QQQ", "IWM"):
+        D = build(tgt, "BTC")
+        for name, col in (("OVERLAP gap (look-ahead)", "gap"), ("TRADABILE intraday", "intra")):
+            C = np.sign(D["crypto"]) * D[col] - COST
+            m = D.index >= OOS
+            print(f"  {tgt} {name:26}: Sharpe long/short {_sh(C):.2f}  (OOS {_sh(C[m]):.2f})  ann {np.nanmean(C)*252*100:+.1f}%")
+        print()
+    print("  --- VERDETTO ---")
+    D = build("QQQ", "BTC")
+    cg = _sh(np.sign(D["crypto"]) * D["gap"] - COST)
+    ci = _sh(np.sign(D["crypto"]) * D["intra"] - COST)
+    print(f"  QQQ: gap(overlap) Sharpe {cg:.2f}  vs  intraday(tradabile) Sharpe {ci:.2f}")
+    print(f"  -> il moto che il crypto 'predice' avviene NELLA finestra del segnale (overnight), non dopo.")
+    print(f"     All'entrata (D13:00, pre-open) il gap e' gia' realizzato -> NON catturabile con l'ETF.")
+    print(f"     Cio' che resta da catturare (intraday) ha Sharpe ~{ci:.1f}: l'edge tradabile e' quello.")
+    print(f"  Per catturare PARTE dell'overnight servirebbe entrare a META' notte via FUTURES IB e")
+    print(f"  testare se il crypto [P21:00->T] predice il future [T->open] (finestre NON sovrapposte):")
+    print(f"  richiede dati intraday dei futures indice (ES/NQ), che NON abbiamo in cache -> data step.")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/crypto_weekend_signal.py

@@ -0,0 +1,97 @@
+"""WEEKEND CRYPTO -> LUNEDI' AZIONARIO — validazione avversariale dell'anticipazione.
+
+L'analisi lead-lag ha trovato UNA anticipazione pulita: il movimento crypto del weekend (Sab+Dom,
+azionario chiuso) anticipa il lunedi' azionario (gap corr ~0.24, OOS piu' forte). Prima di crederci,
+due test scettici:
+  (A) INCREMENTALE: aggiunge info OLTRE il rendimento del VENERDI'? (o e' solo momentum equity?)
+      regressione  Mon ~ weekend_crypto + friday_equity ; il coeff del crypto resta significativo?
+  (B) TRADABILE: segnale eseguibile = osservo weekend crypto (noto Dom 24:00 UTC), entro al Monday
+      OPEN, esco al Monday CLOSE. Net di costi (4 bps RT ETF). Sharpe/hit/OOS vs sempre-long lunedi'.
+
+DATI: cache su disco (BTC Deribit 1h->1d; ETF eq_* con OPEN). Nessun IB online.
+"""
+import sys
+from pathlib import Path
+import numpy as np, pandas as pd
+
+ROOT = Path(__file__).resolve().parents[2]
+sys.path.insert(0, str(ROOT)); sys.path.insert(0, str(ROOT / "scripts" / "research"))
+from src.data.downloader import load_data
+import eqlib
+
+OOS = pd.Timestamp("2022-01-01", tz="UTC")
+COST_RT = 0.0004   # 4 bps round-trip ETF (entry open + exit close)
+
+
+def _sh_weekly(r):
+    r = np.asarray(pd.Series(r).dropna(), float)
+    return float(np.mean(r) / np.std(r) * np.sqrt(52)) if len(r) > 2 and np.std(r) > 0 else 0.0
+
+
+def build(asset_etf="QQQ"):
+    btc = load_data("BTC", "1h").set_index("datetime")["close"].astype(float).resample("1D").last()
+    cal = pd.date_range(btc.index[0], btc.index[-1], freq="D", tz="UTC")
+    bc = btc.reindex(cal).ffill()
+    oc = eqlib.load_eq(asset_etf)["open"].astype(float)
+    cc = eqlib.load_eq(asset_etf)["close"].astype(float)
+    rows = []
+    for mon in cc.index:
+        if mon.weekday() != 0:
+            continue
+        fri = mon - pd.Timedelta(days=3)
+        thu = mon - pd.Timedelta(days=4)
+        if fri not in cc.index or fri not in bc.index or mon not in bc.index:
+            continue
+        wk = bc.loc[mon] / bc.loc[fri] - 1.0                        # weekend crypto (Ven 00:00 -> Lun 00:00)
+        fri_eq = (cc.loc[fri] / cc.loc[thu] - 1.0) if thu in cc.index else np.nan  # rendimento venerdi'
+        gap = oc.loc[mon] / cc.loc[fri] - 1.0
+        intr = cc.loc[mon] / oc.loc[mon] - 1.0                      # tradabile: open->close lunedi'
+        rows.append((mon, wk, fri_eq, gap, intr))
+    return pd.DataFrame(rows, columns=["mon", "wk", "fri_eq", "gap", "intr"]).dropna().set_index("mon")
+
+
+def main():
+    print("=" * 92)
+    print("  WEEKEND CRYPTO -> LUNEDI' AZIONARIO — validazione avversariale")
+    print("=" * 92)
+    for etf in ("QQQ", "SPY", "IWM"):
+        D = build(etf)
+        print(f"\n  ===== {etf}  (n={len(D)} lunedi', {D.index[0].date()}..{D.index[-1].date()}) =====")
+
+        # (A) INCREMENTALE vs venerdi' — regressione OLS standardizzata, t-stat su weekend_crypto
+        for tgt in ("gap", "intr"):
+            y = (D[tgt] - D[tgt].mean()) / D[tgt].std()
+            x1 = (D["wk"] - D["wk"].mean()) / D["wk"].std()
+            x2 = (D["fri_eq"] - D["fri_eq"].mean()) / D["fri_eq"].std()
+            X = np.column_stack([np.ones(len(D)), x1.values, x2.values])
+            beta, *_ = np.linalg.lstsq(X, y.values, rcond=None)
+            resid = y.values - X @ beta
+            se = np.sqrt(np.sum(resid**2) / (len(D) - 3) * np.diag(np.linalg.inv(X.T @ X)))
+            t_wk = beta[1] / se[1]; t_fri = beta[2] / se[2]
+            partial = float(pd.Series(resid).corr(x1))  # ~ contributo crypto al netto del resto
+            print(f"    [{tgt:4}] beta_weekendCrypto {beta[1]:+.3f} (t={t_wk:+.1f})  | "
+                  f"beta_fridayEq {beta[2]:+.3f} (t={t_fri:+.1f})  -> crypto {'INCREMENTALE' if abs(t_wk)>2 else 'non signif.'}")
+
+        # (B) TRADABILE: long lunedi' intraday se weekend crypto > 0, short se < 0 (net costi)
+        sig = np.sign(D["wk"].values)
+        gross = sig * D["intr"].values
+        net = gross - COST_RT
+        D2 = D.assign(net=net)
+        full = D2["net"]; oos = D2[D2.index >= OOS]["net"]; ins = D2[D2.index < OOS]["net"]
+        bh = D["intr"]  # baseline: sempre-long lunedi' intraday
+        hit = float((np.sign(gross) > 0).mean()) if False else float((sig == np.sign(D["intr"].values)).mean())
+        print(f"    TRADE (long/short Mon intraday su segno weekend-crypto, net {COST_RT*1e4:.0f}bps):")
+        print(f"      hit-rate segno {hit*100:.0f}% | Sharpe(sett.) FULL {_sh_weekly(full):.2f}  IS {_sh_weekly(ins):.2f}  OOS22+ {_sh_weekly(oos):.2f}")
+        print(f"      ritorno medio/lun {full.mean()*1e4:+.1f}bps (net) | baseline sempre-long {bh.mean()*1e4:+.1f}bps | "
+              f"ann.~{full.mean()*52*100:+.1f}%")
+        # long-flat (piu' realistico: long se crypto su, altrimenti cash)
+        lf = np.where(D["wk"].values > 0, D["intr"].values, 0.0) - np.where(D["wk"].values > 0, COST_RT, 0.0)
+        lfs = pd.Series(lf, index=D.index)
+        print(f"      variante LONG-FLAT (long se crypto su, else cash): Sharpe FULL {_sh_weekly(lfs):.2f}  "
+              f"OOS {_sh_weekly(lfs[lfs.index>=OOS]):.2f}  ann.~{lfs.mean()*52*100:+.1f}%")
+
+    print("\n  NB: ~52 lunedi'/anno -> Sharpe settimanale; OOS = 2022+. Multiple-testing: 3 ETF x 2 target.")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/eq_crypto_combo.py

@@ -0,0 +1,84 @@
+"""CROSS-MARKET COMBO — la sleeve equity-trend DIVERSIFICA il portafoglio crypto?
+
+(2) dopo EQ-GTAA01. La via che alza il Sharpe COMPLESSIVO senza cercare nuovo alpha: combinare due
+book scorrelati su mercati diversi. crypto = portafoglio attivo TP01+XS01+VRP01 (src.portfolio);
+equity = GTAA lf vt12% (la migliore sleeve equity, corr SPY 0.64, maxDD ~15%). Se la correlazione
+crypto<->equity e' bassa, il blend ha Sharpe > di ciascuno.
+
+ALLINEAMENTO ONESTO: crypto e' calendario-giornaliero (7gg), equity giorni di borsa. Compoundo i
+rendimenti crypto sul grid dei giorni di borsa (cattura i weekend) prima di combinare. Finestra =
+era crypto (TP01 dal 2019). Esecuzione split Deribit+IB (lo noto: e' un portafoglio cross-venue).
+"""
+import sys
+from pathlib import Path
+import numpy as np, pandas as pd
+
+ROOT = Path(__file__).resolve().parents[2]
+sys.path.insert(0, str(ROOT)); sys.path.insert(0, str(ROOT / "scripts" / "research"))
+from eq_sector_momentum import _sh, _cagr, _dd
+from eq_gtaa_trend import gtaa
+from eq_spy_trend import tsmom_exposure, backtest, _series
+from src.portfolio.sleeves import active_sleeves
+from src.portfolio.portfolio import StrategyPortfolio
+
+ANN = np.sqrt(252.0)
+
+
+def _ann_vol(r):
+    return float(np.std(np.asarray(r, float)) * ANN)
+
+
+def compound_to_grid(daily: pd.Series, grid: pd.DatetimeIndex) -> pd.Series:
+    """Compounda una serie di rendimenti (calendario) sul grid dato (giorni di borsa): per ogni data
+    del grid, somma-composta i rendimenti dal punto precedente."""
+    cum = (1.0 + daily).cumprod()
+    cum = cum.reindex(cum.index.union(grid)).ffill().reindex(grid)
+    return (cum / cum.shift(1) - 1.0).dropna()
+
+
+def main():
+    print("=" * 96)
+    print("  CROSS-MARKET COMBO — equity-trend (GTAA) x crypto (TP01+XS01+VRP01)")
+    print("=" * 96)
+    # crypto blend (rinormalizzato, date diverse)
+    crypto = StrategyPortfolio(active_sleeves()).combined_daily()
+    if crypto.index.tz is None:
+        crypto.index = crypto.index.tz_localize("UTC")
+    # equity sleeve (giorni di borsa)
+    eq = gtaa(target_vol=0.12).dropna()
+    # allinea: compounda crypto sui giorni di borsa dell'equity
+    grid = eq.index[eq.index >= crypto.index[0]]
+    cr = compound_to_grid(crypto, grid)
+    J = pd.concat({"crypto": cr, "equity": eq.reindex(cr.index)}, axis=1).dropna()
+    print(f"  finestra comune {J.index[0].date()}..{J.index[-1].date()}  ({len(J)} giorni di borsa)\n")
+
+    c, e = J["crypto"], J["equity"]
+    print("  --- STANDALONE (sulla finestra comune) ---")
+    for nm, r in (("crypto TP01+XS01+VRP01", c), ("equity GTAA vt12", e)):
+        print(f"    {nm:24} Sh {_sh(r):>5.2f}  CAGR {_cagr(r.values,r.index)*100:>5.1f}%  volAnn {_ann_vol(r)*100:>4.1f}%  maxDD {_dd(r.values)*100:>4.0f}%")
+    print(f"\n  --- CORRELAZIONE crypto <-> equity = {c.corr(e):+.3f}  (bassa = diversifica) ---")
+
+    print("\n  --- BLEND (capitale) ---")
+    print(f"    {'mix':18} {'Sharpe':>7} {'CAGR%':>6} {'volAnn%':>7} {'maxDD%':>6}")
+    for wc in (1.0, 0.75, 0.5, 0.25, 0.0):
+        b = wc * c + (1 - wc) * e
+        print(f"    crypto {int(wc*100):>3}/{int((1-wc)*100):<3} eq  {_sh(b):>7.2f} {_cagr(b.values,b.index)*100:>6.1f} {_ann_vol(b)*100:>7.1f} {_dd(b.values)*100:>6.0f}")
+
+    # risk-parity (peso inverso alla vol) — il blend "giusto" quando le vol differiscono
+    vc, ve = _ann_vol(c), _ann_vol(e)
+    wc_rp = (1/vc) / (1/vc + 1/ve)
+    b_rp = wc_rp * c + (1 - wc_rp) * e
+    print(f"\n  --- RISK-PARITY (inv-vol: crypto {wc_rp*100:.0f}% / eq {(1-wc_rp)*100:.0f}%) ---")
+    print(f"    Sharpe {_sh(b_rp):.2f}  CAGR {_cagr(b_rp.values,b_rp.index)*100:.1f}%  volAnn {_ann_vol(b_rp)*100:.1f}%  maxDD {_dd(b_rp.values)*100:.0f}%")
+
+    # verdetto: il blender batte il migliore dei due?
+    best_solo = max(_sh(c), _sh(e))
+    best_blend = max(_sh(0.5*c+0.5*e), _sh(b_rp))
+    print(f"\n  --- VERDETTO ---")
+    print(f"    miglior standalone Sharpe = {best_solo:.2f} | miglior blend Sharpe = {best_blend:.2f} "
+          f"-> {'DIVERSIFICA (blend > solo)' if best_blend > best_solo + 0.03 else 'nessun guadagno netto'}")
+    print(f"    (nota: portafoglio cross-venue Deribit+IB; finestra crypto corta ~{len(J)//252}y)")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/eq_gtaa_trend.py

@@ -0,0 +1,109 @@
+"""EQ-GTAA01 — Trend difensivo MULTI-ASSET (GTAA) sull'universo ETF.
+
+EQ-TREND01 ha mostrato che il trend long-flat su SPY taglia il DD (analogo TP01). La diversificazione
+delle SORGENTI di trend (azioni US/tech/small + bond + oro + high-yield) di solito migliora il
+rischio-aggiustato del trend mono-asset. Qui: ogni asset gestito col proprio trend long-flat
+(TSMOM multi-orizzonte), equal-weight tra gli asset DISPONIBILI (la quota "off" o assente -> cash).
+
+DATI: cache eqlib (ADJUSTED, nessun IB). Start diversi -> outer-join con peso rinormalizzato sugli
+asset esistenti (come gli sleeve crypto). Finestra lunga: SPY/QQQ/IWM da ~2000; TLT(2016)/GLD(2004)/
+HYG(2007) entrano dopo. Riporto anche la finestra 6-asset comune (2016+).
+
+GIUDIZIO: vs SPY buy&hold, vs EW statico (isola il valore del TIMING di trend), vs SPY-trend mono;
+Sharpe full/pre15/OOS + maxDD + plateau. Causale, netto fee.
+"""
+import sys
+from pathlib import Path
+import numpy as np, pandas as pd
+
+ROOT = Path(__file__).resolve().parents[2]
+sys.path.insert(0, str(ROOT / "scripts" / "research"))
+import eqlib
+from eqlib import load_eq
+from eq_sector_momentum import _sh, _cagr, _dd, EQ_HOLDOUT, spy_bh
+from eq_spy_trend import tsmom_exposure, backtest, _series
+
+ASSETS = ["SPY", "QQQ", "IWM", "TLT", "GLD", "HYG"]
+
+
+def gated_returns(sym, horizons=(21, 63, 126, 252), fee_side=0.0002, target_vol=None, lev_cap=1.0):
+    """Rendimenti netti daily di UN asset gestito col proprio trend long-flat (cash quando off)."""
+    px = _series(sym)
+    ex = tsmom_exposure(px, horizons=horizons, target_vol=target_vol, lev_cap=lev_cap)
+    return backtest(px, ex, fee_side=fee_side)
+
+
+def gtaa(assets=ASSETS, horizons=(21, 63, 126, 252), fee_side=0.0002, target_vol=None):
+    """Portafoglio GTAA: media (equal-weight) dei rendimenti trend-gated sugli asset disponibili
+    ogni giorno (outer-join). La quota di asset assenti/in-cash resta in cash."""
+    cols = {a: gated_returns(a, horizons, fee_side, target_vol) for a in assets}
+    R = pd.concat(cols, axis=1).sort_index()
+    return R.mean(axis=1, skipna=True)   # EW sugli asset esistenti quel giorno
+
+
+def ew_buyhold(assets=ASSETS):
+    cols = {a: _series(a).pct_change() for a in assets}
+    return pd.concat(cols, axis=1).sort_index().mean(axis=1, skipna=True)
+
+
+def _row(name, r, common=None, bench=None):
+    r = r.dropna() if common is None else r.reindex(common).fillna(0.0)
+    h = r[r.index >= EQ_HOLDOUT]; ii = r[r.index < EQ_HOLDOUT]
+    tim = float((r != 0).mean()) * 100
+    extra = ""
+    if bench is not None:
+        J = pd.concat({"r": r, "b": bench}, axis=1, join="inner").dropna()
+        extra = f"  corrSPY {J['r'].corr(J['b']):+.2f}"
+    print(f"  {name:26} CAGR {_cagr(r.values, r.index)*100:>5.1f}%  Sh {_sh(r):>5.2f} "
+          f"(pre15 {_sh(ii):>5.2f}|OOS {_sh(h):>5.2f})  maxDD {_dd(r.values)*100:>4.0f}%  inMkt {tim:>3.0f}%{extra}")
+
+
+def main():
+    print("=" * 100)
+    print("  EQ-GTAA01 — Trend difensivo MULTI-ASSET (GTAA)")
+    print("=" * 100)
+    spy = spy_bh()
+    g = gtaa()                          # outer-join, finestra lunga
+    gl = g.dropna()
+    print(f"  finestra lunga (outer-join) {gl.index[0].date()}..{gl.index[-1].date()} ({len(gl)}g)  OOS {EQ_HOLDOUT.date()}+\n")
+
+    print("  --- BASELINE & confronti (finestra lunga) ---")
+    cl = gl.index
+    _row("SPY buy&hold", spy.reindex(cl).fillna(0))
+    _row("EW statico (no trend)", ew_buyhold().reindex(cl).fillna(0), bench=spy.reindex(cl).fillna(0))
+    _row("SPY-trend mono (TREND01)", backtest(_series("SPY"), tsmom_exposure(_series("SPY"))).reindex(cl).fillna(0), bench=spy.reindex(cl).fillna(0))
+    print("\n  --- GTAA (multi-asset trend) ---")
+    _row("GTAA lf", gl, bench=spy.reindex(cl).fillna(0))
+    _row("GTAA lf vt12%", gtaa(target_vol=0.12).reindex(cl).fillna(0), bench=spy.reindex(cl).fillna(0))
+
+    # finestra 6-asset comune (tutti gli ETF esistono): 2016+
+    tlt0 = _series("TLT").index[0]
+    c6 = gl.index[gl.index >= tlt0]
+    print(f"\n  --- finestra 6-asset comune ({c6[0].date()}+) ---")
+    _row("SPY buy&hold (6a win)", spy.reindex(c6).fillna(0))
+    _row("GTAA lf (6a win)", g.reindex(c6).fillna(0), bench=spy.reindex(c6).fillna(0))
+    _row("GTAA lf vt12 (6a win)", gtaa(target_vol=0.12).reindex(c6).fillna(0), bench=spy.reindex(c6).fillna(0))
+
+    # MARGINALE vs SPY
+    print("\n  --- MARGINALE vs SPY (GTAA lf, finestra lunga) ---")
+    J = pd.concat({"spy": spy, "c": gl}, axis=1, join="inner").dropna(); JH = J[J.index >= EQ_HOLDOUT]
+    print(f"    corr full {J['spy'].corr(J['c']):+.3f} | OOS {JH['spy'].corr(JH['c']):+.3f}")
+    for wt in (0.5, 1.0):
+        bf = _sh((1-wt)*J['spy']+wt*J['c'])-_sh(J['spy']); bh = _sh((1-wt)*JH['spy']+wt*JH['c'])-_sh(JH['spy'])
+        lbl = "100% GTAA" if wt == 1.0 else "50/50 SPY/GTAA"
+        print(f"    {lbl:16}: uplift Sharpe FULL {bf:+.3f}  OOS {bh:+.3f}")
+    print("    DD nei bear (GTAA vs SPY):")
+    for lo, hi, lbl in [("2000-03-01","2002-12-31","dot-com"),("2007-10-01","2009-06-30","GFC"),
+                        ("2020-02-01","2020-04-30","COVID"),("2022-01-01","2022-12-31","2022")]:
+        seg=lambda s: _dd(s.reindex(cl).fillna(0)[(cl>=pd.Timestamp(lo,tz='UTC'))&(cl<=pd.Timestamp(hi,tz='UTC'))].values)*100
+        print(f"      {lbl:8} GTAA {seg(gl):.0f}%  | SPY {seg(spy):.0f}%")
+
+    print("\n  --- PLATEAU (Sharpe FULL/pre15/OOS, DD, CAGR) GTAA lf, finestra lunga ---")
+    print(f"    {'horizons':22} {'FULL':>6} {'pre15':>6} {'OOS':>6} {'DD%':>5} {'CAGR%':>6}")
+    for hz in [(63,126,252),(21,63,126,252),(126,252),(252,)]:
+        r = gtaa(horizons=hz).reindex(cl).fillna(0); h=r[r.index>=EQ_HOLDOUT]; ii=r[r.index<EQ_HOLDOUT]
+        print(f"    {'x'.join(map(str,hz)):22} {_sh(r):>6.2f} {_sh(ii):>6.2f} {_sh(h):>6.2f} {_dd(r.values)*100:>5.0f} {_cagr(r.values,r.index)*100:>6.1f}")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/eq_sector_momentum.py

@@ -0,0 +1,159 @@
+"""EQ-MOM01 — Momentum cross-sectional settoriale (SPDR), backtest onesto.
+
+Prima ricerca del fronte equity (branch research/equities-ib). L'edge "noioso e robusto" piu'
+plausibile in un mercato efficiente: ruotare nei settori a momentum forte. Domanda chiave (come nel
+crypto col soffitto TP01): NON "fa soldi?" (un long-only equity cavalca il toro) ma **batte/ADDS a
+SPY buy&hold?** — il baseline vero in equity. Anche vs equal-weight 9 settori (isola il timing del
+momentum dal tilt equal-weight).
+
+DATI: cache su disco eq_*.parquet (ADJUSTED div+split), via eqlib (nessun IB). 9 settori CLASSICI
+dal 1998 (27.5y) per il backtest lungo; 11 settori (2018+) come robustezza.
+
+COSTRUZIONE (causale): ogni REB giorni, momentum = blend di lookback [63,126,252]g con SKIP recente
+(12-1 classico), z-score cross-sectional mediato. long-only: full-invested nei top-k (confronto
+like-for-like con SPY). long-short: dollar-neutral top-k vs bottom-k. Posizione decisa a <= i-1,
+tenuta da i (W[:-1]*dret[1:]). Netto fee sul turnover. Opz. vol-target.
+
+GIUDIZIO: standalone (FULL / pre-2015 / hold-out 2015+ / per-anno, CAGR, Sharpe, maxDD) vs SPY e
+EW-settori; marginale vs SPY (corr, uplift blend full+hold, edge in-sample, persistenza multi-cut);
+plateau su lookback/k/reb/skip; sweep fee.
+"""
+import sys
+from pathlib import Path
+import numpy as np, pandas as pd
+
+ROOT = Path(__file__).resolve().parents[2]
+sys.path.insert(0, str(ROOT / "scripts" / "research"))
+import eqlib
+from eqlib import panel, load_eq, SECTORS_CLASSIC, SECTORS
+
+ANN = np.sqrt(252.0)
+EQ_HOLDOUT = pd.Timestamp("2015-01-01", tz="UTC")   # OOS lungo: ultimi ~11 anni (post-GFC, dove il momentum e' decaduto)
+
+
+def _sh(r):
+    r = np.asarray(pd.Series(r).dropna(), float)
+    return float(np.mean(r) / np.std(r) * ANN) if len(r) > 2 and np.std(r) > 0 else 0.0
+
+
+def _cagr(r, idx):
+    r = np.asarray(r, float); yrs = (idx[-1] - idx[0]).days / 365.25
+    return float(np.prod(1 + r) ** (1 / yrs) - 1) if yrs > 0 else 0.0
+
+
+def _dd(r):
+    eq = np.cumprod(1 + np.asarray(r, float)); pk = np.maximum.accumulate(eq)
+    return float(np.max((pk - eq) / pk)) if len(eq) else 0.0
+
+
+def momentum(universe=tuple(SECTORS_CLASSIC), lookbacks=(63, 126, 252), k=3, reb=21,
+             skip=21, mode="long", target_vol=None, fee_side=0.0002):
+    """Serie netta daily del book momentum settoriale. mode='long' (top-k full-invested) o
+    'ls' (dollar-neutral top-k vs bottom-k)."""
+    P = panel(universe, how="inner")
+    idx = P.index; px = P.values; n, A = px.shape
+    dret = np.vstack([np.zeros(A), px[1:] / px[:-1] - 1.0])
+    mlb = max(lookbacks) + skip
+    W = np.zeros((n, A)); w = np.zeros(A)
+    for i in range(n):
+        if i >= mlb and i % reb == 0:
+            score = np.zeros(A); cnt = 0
+            for Lb in lookbacks:
+                a, b = i - skip - Lb, i - skip
+                rL = px[b] / px[a] - 1.0
+                sd = rL.std()
+                if sd > 0:
+                    score += (rL - rL.mean()) / sd; cnt += 1
+            if cnt:
+                score /= cnt
+                order = np.argsort(score)
+                w = np.zeros(A)
+                if mode == "long":
+                    w[order[-k:]] = 1.0 / k                       # full-invested nei top-k
+                else:
+                    w[order[-k:]] = 0.5 / k; w[order[:k]] = -0.5 / k
+        W[i] = w
+    gross = np.zeros(n); gross[1:] = np.sum(W[:-1] * dret[1:], axis=1)
+    turn = np.zeros(n); turn[0] = np.abs(W[0]).sum(); turn[1:] = np.abs(np.diff(W, axis=0)).sum(axis=1)
+    net = gross - turn * fee_side
+    s = pd.Series(net, index=idx)
+    if target_vol:
+        rv = s.rolling(63, min_periods=20).std().shift(1) * ANN
+        scale = np.clip(np.nan_to_num(target_vol / rv.replace(0, np.nan).values, nan=0.0), 0, 3.0)
+        s = pd.Series(s.values * scale, index=idx)
+    return s
+
+
+def spy_bh():
+    d = load_eq("SPY")["close"].astype(float)
+    return pd.Series(d.values[1:] / d.values[:-1] - 1.0, index=d.index[1:])
+
+
+def ew_sectors_bh(universe=tuple(SECTORS_CLASSIC)):
+    P = panel(universe, how="inner"); dret = P.pct_change().dropna()
+    return dret.mean(axis=1)
+
+
+def _line(name, r, idx=None, bench=None):
+    idx = idx if idx is not None else r.index
+    r = r.reindex(idx).fillna(0.0) if hasattr(r, "reindex") else pd.Series(r, index=idx)
+    h = r[r.index >= EQ_HOLDOUT]; isamp = r[r.index < EQ_HOLDOUT]
+    extra = ""
+    if bench is not None:
+        J = pd.concat({"r": r, "b": bench}, axis=1, join="inner").dropna()
+        extra = f"  corr_SPY {J['r'].corr(J['b']):+.2f}"
+    print(f"  {name:26} CAGR {_cagr(r.values, r.index)*100:>5.1f}%  Sh {_sh(r):>5.2f}  "
+          f"(pre15 {_sh(isamp):>5.2f} | OOS15+ {_sh(h):>5.2f})  maxDD {_dd(r.values)*100:>4.0f}%{extra}")
+
+
+def main():
+    print("=" * 100)
+    print("  EQ-MOM01 — Momentum cross-sectional settoriale (9 SPDR classici, 1998+)")
+    print("=" * 100)
+    spy = spy_bh(); ew = ew_sectors_bh()
+    base = momentum()                                  # long, lb[63,126,252], k=3, reb21, skip21
+    common = base.index
+    print(f"  periodo {common[0].date()}..{common[-1].date()}  ({len(common)}g)  hold-out OOS = {EQ_HOLDOUT.date()}+\n")
+
+    print("  --- BASELINE da battere ---")
+    _line("SPY buy&hold", spy, common)
+    _line("EW 9 settori buy&hold", ew, common, bench=spy)
+    print("\n  --- EQ-MOM01 ---")
+    _line("MOM long top-3", base, common, bench=spy)
+    _line("MOM long top-3 vt15%", momentum(target_vol=0.15), common, bench=spy)
+    _line("MOM long-short top-3", momentum(mode="ls"), common, bench=spy)
+
+    # MARGINALE vs SPY (il test che conta in equity)
+    print("\n  --- MARGINALE vs SPY buy&hold (aggiunge al baseline?) ---")
+    J = pd.concat({"spy": spy, "c": base}, axis=1, join="inner").dropna()
+    JH = J[J.index >= EQ_HOLDOUT]
+    print(f"    corr(MOM,SPY) full {J['spy'].corr(J['c']):+.3f}  | OOS {JH['spy'].corr(JH['c']):+.3f}")
+    for wt in (0.25, 0.5):
+        bf = _sh((1-wt)*J["spy"]+wt*J["c"]) - _sh(J["spy"])
+        bh = _sh((1-wt)*JH["spy"]+wt*JH["c"]) - _sh(JH["spy"])
+        print(f"    blend {int((1-wt)*100)}/{int(wt*100)} SPY/MOM: uplift Sharpe FULL {bf:+.3f}  OOS {bh:+.3f}")
+    # per-decade (multi-cut onesto)
+    print("    Sharpe MOM per blocco: ", {f"{y}s": round(_sh(base[(base.index.year>=y)&(base.index.year<y+5)]),2)
+          for y in (1999,2004,2009,2014,2019,2024)})
+
+    # PLATEAU
+    print("\n  --- PLATEAU (Sharpe FULL / pre15 / OOS15+) long-only ---")
+    print(f"    {'cfg':24} {'FULL':>6} {'pre15':>6} {'OOS':>6} {'CAGR%':>6} {'DD%':>5}")
+    for lbs in [(126,), (63,126,252), (252,)]:
+        for k in (2, 3, 4):
+            for reb in (21,):
+                s = momentum(lookbacks=lbs, k=k, reb=reb)
+                tag = f"lb{'-'.join(map(str,lbs))} k{k}"
+                h=s[s.index>=EQ_HOLDOUT]; ii=s[s.index<EQ_HOLDOUT]
+                print(f"    {tag:24} {_sh(s):>6.2f} {_sh(ii):>6.2f} {_sh(h):>6.2f} {_cagr(s.values,s.index)*100:>6.1f} {_dd(s.values)*100:>5.0f}")
+
+    # sweep fee + robustezza 11 settori (2018+)
+    print("\n  --- ROBUSTEZZA ---")
+    for fee in (0.0, 0.0002, 0.0005, 0.001):
+        s = momentum(fee_side=fee); print(f"    fee {fee*100:.2f}%/lato: Sh FULL {_sh(s):.2f}  OOS {_sh(s[s.index>=EQ_HOLDOUT]):.2f}")
+    s11 = momentum(universe=tuple(SECTORS)); spy11 = spy.reindex(s11.index)
+    print(f"    11 settori (2018+): MOM Sh {_sh(s11):.2f} vs SPY {_sh(spy11):.2f}  (periodo {s11.index[0].date()}+)")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/eq_spy_trend.py

@@ -0,0 +1,146 @@
+"""EQ-TREND01 — Trend DIFENSIVO time-series su SPY (analogo equity di TP01).
+
+Il momentum cross-sectional settoriale e' morto (EQ-MOM01). Ma nel crypto l'unica cosa che ha retto
+NON era un alpha relative-value: era TP01, un trend DIFENSIVO che taglia il drawdown restando vicino
+al ritorno. L'equity ha lo stesso buco: SPY buy&hold fa Sharpe ~0.51 ma con maxDD 55% (due bear -50%:
+2000-02 e 2008-09). Domanda: un trend long-flat su SPY ALZA il Sharpe e DIMEZZA il DD restando
+investito nei tori? (NON cerchiamo di battere il CAGR — cerchiamo il taglio del rischio, come TP01.)
+
+DATI: cache su disco eq_spy/eq_tlt (ADJUSTED), via eqlib (nessun IB).
+COSTRUZIONE (causale, stile TP01): TSMOM multi-orizzonte [21,63,126,252]g (1/3/6/12 mesi); target =
+frazione di orizzonti in trend-up (0..1, allocazione graduale). Vol-target opz. Posizione decisa a
+<=i-1, tenuta da i. Netto fee sul turnover. Varianti: long-flat (cash in risk-off), long-bonds
+(TLT in risk-off, solo dal 2016), e SMA-200 binario (Faber) come riferimento classico.
+
+GIUDIZIO: vs SPY buy&hold (CAGR/Sharpe full-pre15-OOS15+/maxDD/time-in-market), marginale vs SPY,
+DD nei due bear storici, plateau (orizzonti/vol-target/cap leva), sweep fee.
+"""
+import sys
+from pathlib import Path
+import numpy as np, pandas as pd
+
+ROOT = Path(__file__).resolve().parents[2]
+sys.path.insert(0, str(ROOT / "scripts" / "research"))
+import eqlib
+from eqlib import load_eq
+from eq_sector_momentum import _sh, _cagr, _dd, EQ_HOLDOUT, spy_bh
+
+ANN = np.sqrt(252.0)
+
+
+def _series(sym):
+    d = load_eq(sym)["close"].astype(float)
+    return pd.Series(d.values, index=d.index)
+
+
+def tsmom_exposure(close: pd.Series, horizons=(21, 63, 126, 252), target_vol=None,
+                   lev_cap=1.0) -> pd.Series:
+    """Esposizione SPY in [0, lev_cap]: frazione di orizzonti in trend-up, opz. vol-targeted (causale)."""
+    px = close.values; n = len(px); tgt = np.zeros(n)
+    mh = max(horizons)
+    for i in range(mh, n):
+        tgt[i] = np.mean([1.0 if px[i] > px[i - H] else 0.0 for H in horizons])
+    s = pd.Series(tgt, index=close.index)
+    if target_vol:
+        ret = close.pct_change()
+        rv = ret.rolling(63, min_periods=20).std().shift(1) * ANN
+        scale = np.clip(np.nan_to_num(target_vol / rv.replace(0, np.nan).values, nan=0.0), 0, lev_cap / 0.0 if False else 10.0)
+        s = (s * scale).clip(0, lev_cap)
+    else:
+        s = s.clip(0, lev_cap)
+    return s
+
+
+def sma_timing(close: pd.Series, win=200) -> pd.Series:
+    """Faber: long se close > SMA(win), altrimenti flat. Binario {0,1}."""
+    sma = close.rolling(win, min_periods=win // 2).mean()
+    return (close > sma).astype(float)
+
+
+def backtest(close: pd.Series, exposure: pd.Series, risk_off: pd.Series | None = None,
+             fee_side=0.0002) -> pd.Series:
+    """Rendimenti netti: held = exposure ritardata di 1 (causale). La quota non in SPY (1-held, se
+    exposure<=1) va in risk_off (es. TLT) o cash (0). Fee sul turnover di SPY."""
+    ret = close.pct_change().fillna(0.0).values
+    exp = exposure.reindex(close.index).fillna(0.0).values
+    held = np.zeros(len(exp)); held[1:] = exp[:-1]
+    net = held * ret
+    if risk_off is not None:
+        ro = risk_off.reindex(close.index).pct_change().fillna(0.0).values
+        cash_w = np.clip(1.0 - held, 0.0, 1.0)        # quota fuori da SPY -> bonds
+        net = net + cash_w * ro
+    net = net - fee_side * np.abs(np.diff(held, prepend=0.0))
+    return pd.Series(net, index=close.index)
+
+
+def _row(name, r, common, bench=None):
+    r = r.reindex(common).fillna(0.0)
+    h = r[r.index >= EQ_HOLDOUT]; ii = r[r.index < EQ_HOLDOUT]
+    tim = float((r != 0).mean()) * 100
+    extra = ""
+    if bench is not None:
+        J = pd.concat({"r": r, "b": bench.reindex(common).fillna(0.0)}, axis=1).dropna()
+        extra = f"  corr {J['r'].corr(J['b']):+.2f}"
+    print(f"  {name:24} CAGR {_cagr(r.values, r.index)*100:>5.1f}%  Sh {_sh(r):>5.2f} "
+          f"(pre15 {_sh(ii):>5.2f}|OOS {_sh(h):>5.2f})  maxDD {_dd(r.values)*100:>4.0f}%  inMkt {tim:>3.0f}%{extra}")
+
+
+def _bear_dd(r, common, lo, hi, label):
+    seg = r.reindex(common).fillna(0.0)
+    seg = seg[(seg.index >= pd.Timestamp(lo, tz="UTC")) & (seg.index <= pd.Timestamp(hi, tz="UTC"))]
+    return f"{label}: {_dd(seg.values)*100:.0f}%"
+
+
+def main():
+    print("=" * 100)
+    print("  EQ-TREND01 — Trend DIFENSIVO time-series su SPY (analogo di TP01)")
+    print("=" * 100)
+    spy_px = _series("SPY"); spy = spy_bh()
+    common = spy_px.index[spy_px.index >= spy_px.index[252]]   # warmup 1y
+    print(f"  periodo {common[0].date()}..{common[-1].date()}  ({len(common)}g)  OOS = {EQ_HOLDOUT.date()}+\n")
+
+    print("  --- BASELINE ---")
+    _row("SPY buy&hold", spy, common)
+    print("\n  --- TREND long-flat (cash in risk-off) ---")
+    _row("SMA-200 (Faber)", backtest(spy_px, sma_timing(spy_px)), common, bench=spy)
+    _row("TSMOM lf cap1.0", backtest(spy_px, tsmom_exposure(spy_px)), common, bench=spy)
+    _row("TSMOM lf vt15 cap1.0", backtest(spy_px, tsmom_exposure(spy_px, target_vol=0.15, lev_cap=1.0)), common, bench=spy)
+    _row("TSMOM lf vt15 cap1.5", backtest(spy_px, tsmom_exposure(spy_px, target_vol=0.15, lev_cap=1.5)), common, bench=spy)
+
+    print("\n  --- TREND long-BONDS (TLT in risk-off, solo dove TLT esiste: 2016+) ---")
+    tlt = _series("TLT")
+    cb = spy_px.index[(spy_px.index >= tlt.index[0])]
+    _row("SPY b&h (2016+)", spy.reindex(cb), cb)
+    _row("TSMOM lf+TLT (2016+)", backtest(spy_px, tsmom_exposure(spy_px), risk_off=tlt), cb, bench=spy)
+
+    # MARGINALE vs SPY + DD nei bear
+    base = backtest(spy_px, tsmom_exposure(spy_px))
+    print("\n  --- MARGINALE vs SPY (TSMOM lf cap1.0) ---")
+    J = pd.concat({"spy": spy, "c": base}, axis=1, join="inner").dropna(); JH = J[J.index >= EQ_HOLDOUT]
+    print(f"    corr full {J['spy'].corr(J['c']):+.3f} | OOS {JH['spy'].corr(JH['c']):+.3f}")
+    for wt in (0.5, 1.0):
+        bf = _sh((1-wt)*J['spy']+wt*J['c']) - _sh(J['spy']); bh = _sh((1-wt)*JH['spy']+wt*JH['c']) - _sh(JH['spy'])
+        lbl = "100% TREND" if wt == 1.0 else f"{int((1-wt)*100)}/{int(wt*100)} SPY/TREND"
+        print(f"    {lbl:16}: uplift Sharpe FULL {bf:+.3f}  OOS {bh:+.3f}")
+    print("    DD nei bear storici (TSMOM vs SPY):")
+    for lo, hi, lbl in [("2000-03-01","2002-12-31","dot-com"), ("2007-10-01","2009-06-30","GFC"),
+                        ("2020-02-01","2020-04-30","COVID"), ("2022-01-01","2022-12-31","2022")]:
+        print(f"      {lbl:8}  TSMOM {_bear_dd(base,common,lo,hi,'')}  | SPY {_bear_dd(spy,common,lo,hi,'')}")
+
+    # PLATEAU
+    print("\n  --- PLATEAU (Sharpe FULL/pre15/OOS, maxDD, CAGR) long-flat cap1.0 ---")
+    print(f"    {'horizons':22} {'FULL':>6} {'pre15':>6} {'OOS':>6} {'DD%':>5} {'CAGR%':>6}")
+    for hz in [(63,126,252),(21,63,126,252),(126,252),(50,200),(200,)]:
+        ex = sma_timing(spy_px, 200) if hz == (200,) else tsmom_exposure(spy_px, horizons=hz)
+        r = backtest(spy_px, ex); h=r[r.index>=EQ_HOLDOUT]; ii=r[r.index<EQ_HOLDOUT]
+        tag = "SMA-200" if hz==(200,) else "x".join(map(str,hz))
+        print(f"    {tag:22} {_sh(r):>6.2f} {_sh(ii):>6.2f} {_sh(h):>6.2f} {_dd(r.values)*100:>5.0f} {_cagr(r.values,r.index)*100:>6.1f}")
+
+    print("\n  --- SWEEP FEE (TSMOM lf cap1.0) ---")
+    for fee in (0.0, 0.0002, 0.0005, 0.001):
+        r = backtest(spy_px, tsmom_exposure(spy_px), fee_side=fee)
+        print(f"    fee {fee*100:.2f}%/lato: Sh {_sh(r):.2f}  maxDD {_dd(r.values)*100:.0f}%")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/eq_tp01_gtaa_combo.py

@@ -0,0 +1,93 @@
+"""COMBO DEPLOYABLE — TP01 (Deribit) + GTAA (IB): le DUE gambe realmente eseguibili a basso capitale.
+
+Il combo crypto-pieno (TP01+XS01+VRP01)+GTAA diversificava (Sharpe 1.6->1.8), ma XS01/VRP01 sono
+STAT-MODE (non eseguibili a $600). Qui il combo ONESTO/deployable: solo le gambe eseguibili —
+  * TP01: TSMOM difensivo BTC/ETH, long-flat, gia' ARMATO live su Deribit;
+  * GTAA: trend difensivo multi-asset su ETF, eseguibile su IB (frazioni, switch mensile).
+Domanda: due trend difensivi su mercati diversi, scorrelati, danno un blend migliore di ciascuno?
+
+ALLINEAMENTO: TP01 e' calendario-giornaliero (crypto 7gg), GTAA giorni di borsa -> compoundo TP01 sul
+grid dei giorni di borsa (cattura i weekend). Finestra = era TP01 (BTC/ETH dal 2019).
+Confronto anche col combo crypto-pieno per quantificare il COSTO della deployability.
+"""
+import sys
+from pathlib import Path
+import numpy as np, pandas as pd
+
+ROOT = Path(__file__).resolve().parents[2]
+sys.path.insert(0, str(ROOT)); sys.path.insert(0, str(ROOT / "scripts" / "research"))
+from eq_sector_momentum import _sh, _cagr, _dd, EQ_HOLDOUT
+from eq_gtaa_trend import gtaa
+from eq_crypto_combo import compound_to_grid, _ann_vol
+from src.portfolio.sleeves import _tp01_returns, active_sleeves
+from src.portfolio.portfolio import StrategyPortfolio
+
+ANN = np.sqrt(252.0)
+
+
+def _stat(nm, r):
+    print(f"    {nm:26} Sh {_sh(r):>5.2f}  CAGR {_cagr(r.values,r.index)*100:>5.1f}%  "
+          f"volAnn {_ann_vol(r)*100:>4.1f}%  maxDD {_dd(r.values)*100:>4.0f}%")
+
+
+def main():
+    print("=" * 96)
+    print("  COMBO DEPLOYABLE — TP01 (Deribit) + GTAA (IB)")
+    print("=" * 96)
+    tp01 = _tp01_returns()
+    if tp01.index.tz is None:
+        tp01.index = tp01.index.tz_localize("UTC")
+    eq = gtaa(target_vol=0.12).dropna()
+    grid = eq.index[eq.index >= tp01.index[0]]
+    tp = compound_to_grid(tp01, grid)
+    J = pd.concat({"tp01": tp, "gtaa": eq.reindex(tp.index)}, axis=1).dropna()
+    print(f"  finestra comune {J.index[0].date()}..{J.index[-1].date()}  ({len(J)} giorni di borsa, ~{len(J)//252}y)\n")
+    t, g = J["tp01"], J["gtaa"]
+
+    print("  --- STANDALONE (finestra comune) ---")
+    _stat("TP01 (crypto, Deribit)", t)
+    _stat("GTAA vt12 (equity, IB)", g)
+    c = t.corr(g)
+    print(f"\n  --- CORRELAZIONE TP01 <-> GTAA = {c:+.3f} ---")
+
+    print("\n  --- BLEND (capitale) ---")
+    print(f"    {'mix':20} {'Sharpe':>7} {'CAGR%':>6} {'volAnn%':>7} {'maxDD%':>6}")
+    best = (-9, None)
+    for wt in (1.0, 0.75, 0.6, 0.5, 0.4, 0.25, 0.0):
+        b = wt * t + (1 - wt) * g
+        if _sh(b) > best[0]:
+            best = (_sh(b), wt)
+        print(f"    TP01 {int(wt*100):>3}/{int((1-wt)*100):<3} GTAA  {_sh(b):>7.2f} {_cagr(b.values,b.index)*100:>6.1f} {_ann_vol(b)*100:>7.1f} {_dd(b.values)*100:>6.0f}")
+
+    vt, vg = _ann_vol(t), _ann_vol(g)
+    wt_rp = (1/vt) / (1/vt + 1/vg)
+    b_rp = wt_rp * t + (1 - wt_rp) * g
+    print(f"\n  --- RISK-PARITY (inv-vol: TP01 {wt_rp*100:.0f}% / GTAA {(1-wt_rp)*100:.0f}%) ---")
+    _stat("risk-parity blend", b_rp)
+
+    # OOS (post-2023, per dare un taglio fuori dai primi anni TP01)
+    cut = pd.Timestamp("2023-01-01", tz="UTC")
+    print(f"\n  --- robustezza: blend 50/50 per taglio ---")
+    b50 = 0.5 * t + 0.5 * g
+    for lbl, sub in (("full", b50), (">=2023", b50[b50.index >= cut])):
+        print(f"    {lbl:10} Sh {_sh(sub):.2f}  maxDD {_dd(sub.values)*100:.0f}%")
+    print("    Sharpe 50/50 per anno:", {y: round(_sh(b50[b50.index.year == y]), 2) for y in sorted(set(b50.index.year))})
+
+    # VERDETTO + costo della deployability
+    best_solo = max(_sh(t), _sh(g))
+    print(f"\n  --- VERDETTO ---")
+    print(f"    miglior standalone {best_solo:.2f} | blend 50/50 {_sh(b50):.2f} | risk-parity {_sh(b_rp):.2f} | "
+          f"miglior cap-mix {best[0]:.2f} (TP01 {int(best[1]*100)}%)")
+    print(f"    -> {'DIVERSIFICA (blend > solo)' if max(_sh(b50),_sh(b_rp)) > best_solo + 0.03 else 'nessun guadagno netto'}")
+    # costo vs combo crypto-pieno
+    full = StrategyPortfolio(active_sleeves()).combined_daily()
+    if full.index.tz is None: full.index = full.index.tz_localize("UTC")
+    fc = compound_to_grid(full, J.index)
+    Jf = pd.concat({"f": fc, "g": g.reindex(fc.index)}, axis=1).dropna()
+    print(f"    costo deployability: blend 50/50 con crypto-PIENO Sh {_sh(0.5*Jf['f']+0.5*Jf['g']):.2f} "
+          f"vs deployable {_sh(b50):.2f}  (la differenza = cio' che lasciano XS01/VRP01 STAT-MODE)")
+    print(f"    (cross-venue Deribit+IB; entrambe switch mensile/basso turnover; frazionabili a $0.5-2k)")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/eqlib.py

@@ -0,0 +1,65 @@
+"""EQLIB — harness di ricerca EQUITY/ETF (branch research/equities-ib).
+
+Legge la CACHE su disco (data/raw/eq_*.parquet, ADJUSTED_LAST, scritta una volta da
+fetch_ib_equities.py) — MAI da IB. `lru_cache` -> ogni parquet si legge una sola volta per processo.
+"Memorizza i dati per non rileggerli ogni volta": la persistenza e' il parquet su disco + questa cache.
+
+Espone: universi (SECTORS/BROAD), load_eq(sym), panel(universe) allineato, e riusa lo scorer
+indurito di altlib (_sh, _dd_ret, _to_daily, marginal_vs_tp01) per giudicare i candidati con la
+stessa disciplina del lato crypto.
+"""
+from functools import lru_cache
+from pathlib import Path
+import numpy as np, pandas as pd
+
+ROOT = Path(__file__).resolve().parents[2]
+RAW = ROOT / "data" / "raw"
+
+# 11 SPDR settoriali. I 9 "classici" (sotto SECTORS_CLASSIC) partono 1998; XLRE 2015, XLC 2018.
+SECTORS = ["XLK", "XLF", "XLE", "XLV", "XLI", "XLP", "XLY", "XLU", "XLB", "XLRE", "XLC"]
+SECTORS_CLASSIC = ["XLK", "XLF", "XLE", "XLV", "XLI", "XLP", "XLY", "XLU", "XLB"]   # storia lunga (1998+)
+BROAD = ["SPY", "QQQ", "IWM", "TLT", "GLD", "HYG"]
+
+
+@lru_cache(maxsize=64)
+def load_eq(sym: str) -> pd.DataFrame:
+    """OHLCV aggiustato (dividendi+split) per `sym`, indicizzato datetime UTC. Cache su disco -> RAM."""
+    p = RAW / f"eq_{sym.lower()}_1d.parquet"
+    if not p.exists():
+        raise FileNotFoundError(f"{p} assente — gira: uv run --with ib_async python scripts/research/fetch_ib_equities.py")
+    d = pd.read_parquet(p).copy()
+    d.index = pd.to_datetime(d["timestamp"], unit="ms", utc=True)
+    return d[["open", "high", "low", "close", "volume"]]
+
+
+@lru_cache(maxsize=16)
+def _close_panel(universe: tuple) -> pd.DataFrame:
+    cols = {s: load_eq(s)["close"].astype(float) for s in universe}
+    return pd.concat(cols, axis=1).sort_index()
+
+
+def panel(universe=tuple(SECTORS), how: str = "inner") -> pd.DataFrame:
+    """Prezzi close aggiustati [date x asset]. how='inner' = date comuni a TUTTI (start = ETF piu' giovane);
+    'outer' = unione (NaN dove un ETF non esiste ancora)."""
+    P = _close_panel(tuple(universe))
+    return P.dropna(how="any") if how == "inner" else P
+
+
+def describe(universe=None):
+    universe = universe or (SECTORS + BROAD)
+    print(f"  {'sym':6} {'barre':>6} {'da':>11} {'a':>11} {'anni':>5}")
+    for s in universe:
+        try:
+            d = load_eq(s)
+            print(f"  {s:6} {len(d):>6} {str(d.index[0].date()):>11} {str(d.index[-1].date()):>11} "
+                  f"{(d.index[-1]-d.index[0]).days/365.25:>5.1f}")
+        except FileNotFoundError:
+            print(f"  {s:6}  (assente)")
+    Pc = panel(SECTORS_CLASSIC); Pa = panel(SECTORS)
+    print(f"\n  panel 9 settori CLASSICI: {Pc.shape[1]}x{len(Pc)}  start comune {Pc.index[0].date()}")
+    print(f"  panel 11 settori        : {Pa.shape[1]}x{len(Pa)}  start comune {Pa.index[0].date()}")
+
+
+if __name__ == "__main__":
+    print("=" * 70); print("  EQLIB — cache equity su disco (nessun IB)"); print("=" * 70)
+    describe()

scripts/research/eu_overnight_deepdive.py

@@ -0,0 +1,86 @@
+"""DEEP-DIVE: crypto[sessione USA] -> indice EUROPEO[overnight successivo]. Reale o artefatto?
+
+Lo sweep ha trovato un forte segnale: crypto[T-8h->T] (T~00:00 UTC, fine sessione USA) predice il
+future europeo (ESTX50/DAX)[T->T+6h] con t_crypto incrementale ~8, Sharpe 2.5, 3/3 anni. Verifiche:
+  (1) ROBUSTEZZA su T (ora entrata): un effetto vero non e' a coltello su una sola ora.
+  (2) vs SEMPRE-LONG: l'overnight ha un drift? il crypto AGGIUNGE oltre il long incondizionato?
+  (3) GAP 1-2h tra fine-segnale e inizio-cattura: se sopravvive, niente contaminazione di bordo.
+  (4) vs FUTURE-OWN: il crypto batte il momentum proprio del future (gia' nel t incrementale).
+Dati: cache (fut europei + crypto 1h). sqrt(252), net 2bps.
+"""
+import sys
+from pathlib import Path
+import numpy as np, pandas as pd
+
+ROOT = Path(__file__).resolve().parents[2]; RAW = ROOT / "data" / "raw"
+sys.path.insert(0, str(ROOT)); sys.path.insert(0, str(ROOT / "scripts" / "research"))
+from crypto_lead_harness import crypto_hourly, at
+COST = 0.0002
+
+
+def fut_hourly(sym):
+    d = pd.read_parquet(RAW / f"fut_{sym.lower()}_1h.parquet")
+    return pd.Series(d["close"].astype(float).values, index=pd.to_datetime(d["timestamp"], unit="ms", utc=True)).sort_index()
+
+
+def _sh(r):
+    r = np.asarray(r, float); r = r[np.isfinite(r)]
+    return float(np.mean(r) / np.std(r) * np.sqrt(252)) if len(r) > 5 and np.std(r) > 0 else 0.0
+
+
+def series(fut, bc, T, S=8, H=6, gap=0):
+    days = pd.date_range(fut.index[0].normalize(), fut.index[-1].normalize(), freq="D", tz="UTC")
+    rows = []
+    for D in days:
+        te = D + pd.Timedelta(hours=T)
+        f0, f1 = at(fut, te - pd.Timedelta(hours=S)), at(fut, te)
+        fcs, fce = at(fut, te + pd.Timedelta(hours=gap)), at(fut, te + pd.Timedelta(hours=gap + H))
+        c0, c1 = at(bc, te - pd.Timedelta(hours=S)), at(bc, te)
+        if not all(np.isfinite(v) and v > 0 for v in (f0, f1, fcs, fce, c0, c1)):
+            continue
+        rows.append((D, c1/c0 - 1, f1/f0 - 1, fce/fcs - 1))
+    Dd = pd.DataFrame(rows, columns=["d", "csig", "fctrl", "cap"]).set_index("d")
+    return Dd[Dd["cap"].abs() > 1e-9]
+
+
+def stats(Dd):
+    x, ctrl, y = Dd["csig"].values, Dd["fctrl"].values, Dd["cap"].values
+    def z(a):
+        s = a.std(); return (a - a.mean()) / s if s > 0 else a * 0
+    X = np.column_stack([np.ones(len(y)), z(x), z(ctrl)])
+    beta, *_ = np.linalg.lstsq(X, z(y), rcond=None)
+    resid = z(y) - X @ beta; dof = max(len(y) - 3, 1)
+    se = np.sqrt(np.sum(resid**2)/dof * np.diag(np.linalg.inv(X.T@X)))
+    t_c = float(beta[1]/se[1]) if se[1] > 0 else 0.0
+    crypto = np.sign(x)*y - COST
+    futown = np.sign(ctrl)*y - COST
+    always = y - COST
+    yrs = Dd.index.year.values
+    py = {int(v): round(_sh(crypto[yrs==v]),2) for v in sorted(set(yrs)) if (yrs==v).sum()>=30}
+    return dict(n=len(Dd), t_crypto=round(t_c,2), sh_crypto=round(_sh(crypto),2),
+                sh_futown=round(_sh(futown),2), sh_always=round(_sh(always),2),
+                ann_crypto=round(float(np.nanmean(crypto)*252*100),1), per_year=py)
+
+
+def main():
+    print("="*96); print("  DEEP-DIVE crypto -> indice EUROPEO overnight (reale o artefatto?)"); print("="*96)
+    bcs = {l: crypto_hourly(l) for l in ("BTC","ETH")}
+    for sym in ("ESTX50","DAX"):
+        fut = fut_hourly(sym)
+        print(f"\n  ===== {sym} =====")
+        print(f"  (1) ROBUSTEZZA su T (lead=BTC, S=8h H=6h gap=0): t_crypto | Sh crypto vs futown vs always-long")
+        for T in (20,21,22,23,0,1,2,3,4):
+            s = stats(series(fut, bcs["BTC"], T))
+            print(f"      T={T:>2}h: n={s['n']} t {s['t_crypto']:+.1f} | crypto {s['sh_crypto']:+.2f}  futown {s['sh_futown']:+.2f}  always {s['sh_always']:+.2f}  ann {s['ann_crypto']:+.1f}% {s['per_year']}")
+        print(f"  (3) GAP test (T=0h, cattura inizia +1h e +2h dopo il segnale):")
+        for g in (0,1,2):
+            s = stats(series(fut, bcs["BTC"], 0, gap=g))
+            print(f"      gap={g}h: t {s['t_crypto']:+.1f} | crypto {s['sh_crypto']:+.2f} always {s['sh_always']:+.2f}")
+        print(f"  (lead ETH, T=0h): ", end="")
+        s = stats(series(fut, bcs["ETH"], 0)); print(f"t {s['t_crypto']:+.1f} crypto {s['sh_crypto']:+.2f} always {s['sh_always']:+.2f} {s['per_year']}")
+    print("\n  LETTURA: se 'crypto' >> 'always' E >> 'futown' E regge su T e col gap -> lead REALE (crypto")
+    print("           anticipa il catch-up europeo). Se crypto ~ always -> e' solo overnight-drift europeo.")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/fetch_hl_funding.py

@@ -0,0 +1,137 @@
+"""FETCH + CERTIFY funding rate Hyperliquid (API pubblica, tokenless) — per ricerca CARRY cross-sectional.
+
+CONTESTO (2026-06-22, onda "nuova ricerca mirata"). Le due grandi ondate (sweep 104-ipotesi +
+ortho relative-value) hanno esaurito gli angoli DIREZIONALI e RELATIVE-VALUE sul *prezzo* BTC/ETH.
+L'unico meccanismo con una fonte di ritorno DIVERSA non ancora testato su dati certi e' il CARRY da
+funding (incassare il cashflow perp, delta-neutral). Scan di fattibilita':
+  * funding price-clock sul feed Deribit certificato -> gia' testato (agent_03 intraday) = FAIL.
+  * funding carry su Deribit (dove eseguiamo) -> ccxt fetch_funding_rate_history = 0 righe (bloccato).
+  * funding carry su Hyperliquid -> API pubblica /info {"type":"fundingHistory"} = DISPONIBILE,
+    cadenza ORARIA, tokenless, serie native dal 2023-05-12. HL e' gia' l'universo certificato di XS01.
+
+DISCIPLINA (lezione v2.0.0): nessuna fiducia nel dato finche' non e' certificato. Qui certifichiamo:
+  (1) cadenza ~1h coerente, (2) gap interni, (3) copertura (giorni nativi reali per coin),
+  (4) plausibilita' magnitudine (|funding orario| tipico < ~0.06%/h = cap HL; outlier flaggati),
+  (5) il funding e' un CASHFLOW, non un prezzo -> niente cross-venue OHLC; il sanity check e'
+      che il funding medio sia ~positivo e piccolo (premio long-pays-short tipico dei perp crypto).
+
+Universo = i 19 major di XS01 (quelli che la strategia live userebbe). Output:
+  data/raw/hlfund_<sym>_1h.parquet  (namespace dedicato 'hlfund', NON tocca hl_<sym>_1d di XS01).
+"""
+import sys, time, datetime as dt
+from pathlib import Path
+import numpy as np, pandas as pd, requests
+
+ROOT = Path(__file__).resolve().parents[2]
+RAW = ROOT / "data" / "raw"
+RAW.mkdir(parents=True, exist_ok=True)
+
+# 19 major di XS01 (CLAUDE.md)
+UNIVERSE = ["BTC","ETH","SOL","BNB","XRP","DOGE","AVAX","LINK","LTC","ADA",
+            "ARB","OP","SUI","APT","INJ","TIA","SEI","NEAR","AAVE"]
+
+HL_INFO = "https://api.hyperliquid.xyz/info"
+START = int(dt.datetime(2023, 1, 1, tzinfo=dt.timezone.utc).timestamp() * 1000)
+HOUR_MS = 3600 * 1000
+
+
+def _post(payload, max_retry=6):
+    """POST con backoff esponenziale su 429/5xx (l'API pubblica HL throttla)."""
+    delay = 1.0
+    for attempt in range(max_retry):
+        r = requests.post(HL_INFO, json=payload, timeout=30)
+        if r.status_code == 429 or r.status_code >= 500:
+            time.sleep(delay)
+            delay = min(delay * 2, 20)  # 1,2,4,8,16,20
+            continue
+        r.raise_for_status()
+        return r.json()
+    r.raise_for_status()
+    return r.json()
+
+
+def fetch_funding(coin: str) -> pd.DataFrame:
+    """Pagina fundingHistory (max 500/req, orario) avanzando startTime fino a oggi."""
+    rows, start = [], START
+    seen = set()
+    while True:
+        d = _post({"type": "fundingHistory", "coin": coin, "startTime": start})
+        if not d:
+            break
+        new = [x for x in d if x["time"] not in seen]
+        for x in new:
+            seen.add(x["time"])
+            rows.append((x["time"], float(x["fundingRate"]), float(x.get("premium", "nan"))))
+        last = d[-1]["time"]
+        if len(d) < 500:  # ultima pagina
+            break
+        nxt = last + 1
+        if nxt <= start:  # niente progresso -> stop
+            break
+        start = nxt
+        time.sleep(0.35)  # gentile con l'API pubblica
+    if not rows:
+        return pd.DataFrame(columns=["ts", "funding", "premium"]).set_index("ts")
+    df = pd.DataFrame(rows, columns=["ts", "funding", "premium"]).drop_duplicates("ts").sort_values("ts")
+    df["ts"] = pd.to_datetime(df["ts"], unit="ms", utc=True)
+    return df.set_index("ts")
+
+
+def certify(coin: str, df: pd.DataFrame) -> dict:
+    if df.empty:
+        return {"coin": coin, "n": 0, "status": "VUOTO"}
+    idx = df.index
+    span_days = (idx[-1] - idx[0]).total_seconds() / 86400
+    # cadenza: differenze in ore
+    deltas_h = np.diff(idx.view("int64")) / 1e9 / 3600
+    median_dt = float(np.median(deltas_h)) if len(deltas_h) else float("nan")
+    gaps = int((deltas_h > 1.5).sum())  # buchi > 1.5h
+    expected = int(round(span_days * 24)) + 1
+    coverage = len(df) / expected if expected else float("nan")
+    f = df["funding"].values
+    # statistiche funding (orario)
+    ann = float(np.nanmean(f)) * 24 * 365  # funding annualizzato (carry teorico per chi paga)
+    cap_hits = int((np.abs(f) > 0.0006).sum())  # cap HL ~0.06%/h (4%/8h clamp); fuori = sospetto
+    status = "OK"
+    if coverage < 0.97 or gaps > 50:
+        status = "GAP"
+    if span_days < 365:
+        status = "corto<365g"
+    return {"coin": coin, "n": len(df), "primo": idx[0].date(), "ultimo": idx[-1].date(),
+            "giorni": round(span_days), "cad_h": round(median_dt, 3), "gap>1.5h": gaps,
+            "cover%": round(coverage * 100, 1), "fund_med_bps": round(float(np.nanmedian(f)) * 1e4, 4),
+            "fund_ann%": round(ann * 100, 1), "cap_hit": cap_hits, "status": status}
+
+
+def main():
+    print("=" * 100)
+    print("  FETCH + CERTIFY funding Hyperliquid (orario, tokenless) — 19 major XS01 -> data/raw/hlfund_*")
+    print("=" * 100)
+    rep = []
+    for sym in UNIVERSE:
+        try:
+            df = fetch_funding(sym)
+        except Exception as e:
+            print(f"  {sym:5} ERR {repr(e)[:80]}")
+            rep.append({"coin": sym, "n": 0, "status": "ERR"})
+            continue
+        c = certify(sym, df)
+        rep.append(c)
+        if c.get("n", 0) > 0:
+            out = RAW / f"hlfund_{sym.lower()}_1h.parquet"
+            df.to_parquet(out)
+        print(f"  {sym:5} n={c.get('n',0):>6} {str(c.get('primo','')):>10}->{str(c.get('ultimo','')):>10} "
+              f"cad={c.get('cad_h','?')}h gap={c.get('gap>1.5h','?')} cov={c.get('cover%','?')}% "
+              f"med={c.get('fund_med_bps','?')}bps ann={c.get('fund_ann%','?')}% cap_hit={c.get('cap_hit','?')} "
+              f"[{c['status']}]")
+    ok = [r["coin"] for r in rep if r.get("status") in ("OK",)]
+    short = [r["coin"] for r in rep if r.get("status") == "corto<365g"]
+    print("-" * 100)
+    print(f"  CERTIFICATI OK ({len(ok)}): {ok}")
+    if short:
+        print(f"  CORTI <365g ({len(short)}): {short}")
+    print(f"  Scritti in data/raw/hlfund_<sym>_1h.parquet (funding ORARIO, serie nativa).")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/fetch_ib_equities.py

@@ -0,0 +1,134 @@
+"""FETCH + CERTIFY universo azioni/ETF da IB (ADJUSTED_LAST) -> data/raw/eq_<sym>_1d.parquet.
+
+Apre il fronte EQUITY (branch research/equities-ib). Disciplina v2.0.0: PRIMA il dato certificato,
+POI la strategia. IB dà storia daily aggiustata per dividendi+split (ADJUSTED_LAST), profonda
+(SPY dal 1996), sul conto paper. Namespace dedicato 'eq_' (NON tocca i parquet crypto).
+
+UNIVERSO (prima ricerca = momentum cross-sectional settoriale, l'edge robusto plausibile in equity):
+  * 11 SPDR settoriali (XLK..XLC);  * broad/macro SPY QQQ IWM TLT GLD HYG.
+NB: i 9 settori "classici" partono 1998; XLRE 2015, XLC 2018 -> lo start COMUNE a 11 e' 2018.
+Per backtest lunghi usare i 9 classici (1998+) o accettare lo start 2018 per gli 11.
+
+CERTIFICAZIONE (gemello equity di certify_feed.py):
+  (1) integrità: barre, range, date monotone, duplicati, flat bars (close invariato);
+  (2) gap: run di giorni-lavorativi mancanti > 5 (festivi normali, buchi lunghi = sospetti);
+  (3) sanità ritorni: max |daily ret| (un >50% non-evento = errore di adjustment);
+  (4) sanità adjustment: primo close aggiustato << ultimo (i dividendi abbassano lo storico).
+PREREQUISITO: gateway IB paper su 127.0.0.1:4002 (docker compose up -d ib-gateway).
+
+  uv run --with ib_async python scripts/research/fetch_ib_equities.py
+"""
+import sys, time
+from pathlib import Path
+import numpy as np, pandas as pd
+
+ROOT = Path(__file__).resolve().parents[2]
+RAW = ROOT / "data" / "raw"
+RAW.mkdir(parents=True, exist_ok=True)
+
+SECTORS = ["XLK", "XLF", "XLE", "XLV", "XLI", "XLP", "XLY", "XLU", "XLB", "XLRE", "XLC"]
+BROAD = ["SPY", "QQQ", "IWM", "TLT", "GLD", "HYG"]
+# espansione "diversi mercati" (intl / bond / credito / commodity / settori extra) per il lead-lag crypto
+BROAD2 = ["DIA", "EFA", "EEM", "FXI", "EWJ", "AGG", "LQD", "IEF", "USO", "SLV", "DBC", "VNQ"]
+UNIVERSE = SECTORS + BROAD + BROAD2
+
+
+def certify(sym: str, df: pd.DataFrame) -> dict:
+    if df.empty:
+        return {"sym": sym, "n": 0, "status": "VUOTO"}
+    idx = df.index
+    dup = int(idx.duplicated().sum())
+    mono = bool(idx.is_monotonic_increasing)
+    c = df["close"].values.astype(float)
+    ret = np.diff(c) / c[:-1]
+    flat = int((ret == 0).sum())
+    maxret = float(np.max(np.abs(ret))) if len(ret) else 0.0
+    # gap: giorni lavorativi attesi vs presenti, run lunghi mancanti
+    bdays = pd.bdate_range(idx[0], idx[-1])
+    missing = len(bdays) - len(idx.intersection(bdays))
+    gaps = bdays.difference(idx)
+    longgap = 0
+    if len(gaps):
+        g = pd.Series(1, index=gaps).resample("1D").sum().fillna(0)
+        # conta run consecutivi di bday mancanti
+        s = (gaps.to_series().diff().dt.days.fillna(1) > 3).cumsum()
+        longgap = int((gaps.to_series().groupby(s).size() > 5).sum())
+    span_y = (idx[-1] - idx[0]).days / 365.25
+    adj_ratio = round(float(c[0] / c[-1]), 3)  # primo/ultimo: <1 atteso (storico abbassato dai div)
+    status = "OK"
+    if dup or not mono:
+        status = "INTEGRITA'"
+    elif maxret > 0.5:
+        status = "SPIKE?"
+    elif longgap > 0:
+        status = "GAP-LUNGO"
+    elif span_y < 1:
+        status = "corto<1y"
+    return {"sym": sym, "n": len(df), "primo": idx[0].date(), "ultimo": idx[-1].date(),
+            "anni": round(span_y, 1), "dup": dup, "mono": mono, "flat": flat,
+            "maxret%": round(maxret * 100, 1), "miss_bd": missing, "gap_lunghi": longgap,
+            "adj_first/last": adj_ratio, "status": status}
+
+
+def main():
+    try:
+        from ib_async import IB, Stock
+    except Exception:
+        print("ib_async assente. Esegui con: uv run --with ib_async python scripts/research/fetch_ib_equities.py")
+        sys.exit(2)
+    ib = IB()
+    try:
+        ib.connect("127.0.0.1", 4002, clientId=90, timeout=15)
+    except Exception as e:
+        print(f"[CONNESSIONE FALLITA] 127.0.0.1:4002 -> {repr(e)[:120]}\n  Avvia: docker compose up -d ib-gateway")
+        sys.exit(1)
+
+    print("=" * 104)
+    print(f"  FETCH + CERTIFY azioni/ETF (ADJUSTED_LAST) -> data/raw/eq_* | acct {ib.managedAccounts()}")
+    print("=" * 104)
+    rep, ok = [], []
+    force = "--force" in sys.argv[1:]
+    universe = UNIVERSE
+    if "--only" in sys.argv[1:]:                      # refresh mirato (es. solo i 6 ETF GTAA per il cron)
+        universe = sys.argv[sys.argv.index("--only") + 1].upper().split(",")
+        force = True                                  # --only implica refresh dei simboli indicati
+    for sym in universe:
+        out_path = RAW / f"eq_{sym.lower()}_1d.parquet"
+        if out_path.exists() and not force:
+            print(f"  {sym:5} GIA' SU DISCO -> skip (usa --force per riscaricare)")
+            ok.append(sym)
+            continue
+        con = Stock(sym, "SMART", "USD")
+        try:
+            bars = ib.reqHistoricalData(con, endDateTime="", durationStr="30 Y", barSizeSetting="1 day",
+                                        whatToShow="ADJUSTED_LAST", useRTH=True, formatDate=1, timeout=60)
+        except Exception as e:
+            print(f"  {sym:5} ERR {repr(e)[:70]}"); rep.append({"sym": sym, "status": "ERR"}); time.sleep(1.2); continue
+        if not bars:
+            print(f"  {sym:5} 0 barre (subscription?)"); rep.append({"sym": sym, "n": 0, "status": "VUOTO"}); time.sleep(1.2); continue
+        df = pd.DataFrame([(pd.Timestamp(str(b.date)), b.open, b.high, b.low, b.close, b.volume) for b in bars],
+                          columns=["ts", "open", "high", "low", "close", "volume"]).set_index("ts").sort_index()
+        c = certify(sym, df)
+        rep.append(c)
+        if c.get("n", 0) > 0:
+            out = df.copy()
+            # ms epoch (come i parquet crypto), robusto alla risoluzione datetime64 (s/us/ns)
+            out["timestamp"] = out.index.astype("datetime64[ms]").astype("int64")
+            out.reset_index(drop=True).to_parquet(RAW / f"eq_{sym.lower()}_1d.parquet")
+            if c["status"] == "OK":
+                ok.append(sym)
+        print(f"  {sym:5} n={c.get('n',0):>5} {str(c.get('primo','')):>10}->{str(c.get('ultimo',''))} "
+              f"{c.get('anni','?')}y flat={c.get('flat','?')} maxret={c.get('maxret%','?')}% "
+              f"miss_bd={c.get('miss_bd','?')} gapL={c.get('gap_lunghi','?')} adj={c.get('adj_first/last','?')} [{c['status']}]")
+        time.sleep(1.2)  # pacing IB
+
+    print("-" * 104)
+    print(f"  CERTIFICATI OK ({len(ok)}/{len(UNIVERSE)}): {ok}")
+    sec_ok = [s for s in SECTORS if s in ok]
+    print(f"  settori OK: {len(sec_ok)}/11 {sec_ok}")
+    print(f"  -> scritti in data/raw/eq_<sym>_1d.parquet (ADJUSTED_LAST, namespace dedicato).")
+    ib.disconnect()
+
+
+if __name__ == "__main__":
+    main()

scripts/research/fetch_ib_futures.py

@@ -0,0 +1,64 @@
+"""FETCH futures indice ORARI (ES/NQ/RTY) da IB -> data/raw/fut_<sym>_1h.parquet (UTC).
+
+Per il test onesto dell'idea "monitor Deribit / trade IB": serve il path INTRADAY del future indice
+(che si trada di notte) per misurare finestre overnight NON sovrapposte col segnale crypto.
+ContFuture orario, in chunk da 1 anno (IB limita le durate intraday). Convertito in UTC.
+Resumable (salta i parquet gia' scritti). Per RETURNS/lead-lag il back-adjust del ContFuture e' ok
+(i ritorni infra-contratto sono preservati; i gap di roll ~4/anno sono trascurabili).
+
+  uv run --with ib_async python scripts/research/fetch_ib_futures.py
+"""
+import sys, time
+from pathlib import Path
+import numpy as np, pandas as pd
+
+ROOT = Path(__file__).resolve().parents[2]
+RAW = ROOT / "data" / "raw"
+# sym -> exchange (indici US + esteri + commodity + bond per la ricerca cross-mercato oltre SP500)
+SYMS = {"ES": "CME", "NQ": "CME", "RTY": "CME",
+        "GC": "COMEX", "CL": "NYMEX", "HG": "COMEX", "ZN": "CBOT",
+        "ESTX50": "EUREX", "DAX": "EUREX", "NKD": "CME"}
+N_CHUNKS = 5   # (non usato: ContFuture non accetta endDateTime -> chiamata singola)
+
+
+def main():
+    from ib_async import IB, ContFuture
+    ib = IB()
+    try:
+        ib.connect("127.0.0.1", 4002, clientId=144, timeout=15)
+    except Exception as e:
+        print(f"[CONNESSIONE FALLITA] {repr(e)[:100]}"); sys.exit(1)
+    print(f"  acct {ib.managedAccounts()} | fetch futures orari -> data/raw/fut_*")
+    for sym, exc in SYMS.items():
+        out = RAW / f"fut_{sym.lower()}_1h.parquet"
+        if out.exists():
+            print(f"  {sym}: gia' su disco -> skip"); continue
+        cf = ContFuture(sym, exchange=exc)
+        try:
+            ib.qualifyContracts(cf)
+        except Exception as e:
+            print(f"  {sym}: qualify ERR {repr(e)[:60]}"); continue
+        # ContFuture NON accetta endDateTime (Error 10339) -> chiamata singola, durata massima (~3y orari)
+        try:
+            b = ib.reqHistoricalData(cf, endDateTime="", durationStr="4 Y", barSizeSetting="1 hour",
+                                     whatToShow="TRADES", useRTH=False, formatDate=1, timeout=150)
+        except Exception as e:
+            print(f"  {sym}: ERR {repr(e)[:60]}"); continue
+        if not b:
+            print(f"  {sym}: VUOTO"); continue
+        D = pd.DataFrame([(pd.Timestamp(x.date), x.open, x.high, x.low, x.close, x.volume) for x in b],
+                         columns=["ts", "open", "high", "low", "close", "volume"]).drop_duplicates("ts").sort_values("ts").reset_index(drop=True)
+        # -> UTC ms (robusto alla risoluzione us/ns: naive-UTC -> datetime64[ms] -> int64)
+        ts = pd.to_datetime(D["ts"], utc=True).dt.tz_convert("UTC").dt.tz_localize(None)
+        D["timestamp"] = ts.values.astype("datetime64[ms]").astype("int64")
+        D = D.drop(columns=["ts"])
+        D.to_parquet(out)
+        u = pd.to_datetime(D["timestamp"], unit="ms", utc=True)
+        print(f"  {sym}: SCRITTO {len(D)} barre  {u.iloc[0]} .. {u.iloc[-1]}  -> {out.name}")
+        time.sleep(1.5)
+    ib.disconnect()
+    print("  done.")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/funding_carry_hl.py

@@ -0,0 +1,170 @@
+"""FC01 — Funding-CARRY cross-sectional su Hyperliquid (backtest onesto, STAT-MODE).
+
+IPOTESI. Il funding dei perp e' un CASHFLOW (long pagano short quando f>0). Un book
+dollar-neutral che VENDE i perp ad alto funding e COMPRA quelli a basso funding INCASSA
+il premio di funding (carry). E' una fonte di ritorno DIVERSA dal trend (TP01) e — forse —
+dal momentum cross-sectional (XS01). Domanda chiave: e' un edge reale o solo XS01 travestito?
+(gli asset ad alto funding sono spesso quelli pompati = forti = quelli che XS01 COMPRA; qui li
+SHORTIAMO -> potenziale ANTI-correlazione con XS01, oppure il carry domina).
+
+DATI (certificati). funding orario reale HL (scripts/research/fetch_hl_funding.py, dal 2023-05) +
+prezzi HL 1d (data/raw/hl_*_1d.parquet, gli stessi di XS01). Universo = i 19 major di XS01.
+
+COSTRUZIONE (causale, come XS01). Ogni H giorni: segnale = media causale del funding giornaliero
+realizzato sugli ultimi L giorni (solo dati <= i-1). Rank cross-section; SHORT i k ad alto funding,
+LONG i k a basso (dollar-neutral). Ritorno del perp per un LONG = price_ret - funding (chi e' long
+paga il funding); quindi r_book[i] = sum_a w[i-1,a] * (price_ret[i,a] - funding_realizzato[i,a]),
+meno fee sul turnover (0.05%/lato, come XS01), poi vol-target 20%.
+
+GIUDIZIO (metodologia indurita). Standalone (FULL/in-sample/hold-out, DD, anni) + correlazione a
+TP01/XS01/VRP01 + marginal_vs_tp01 (gate: edge in-sample>=0.5, persistenza multi-cut, hedge-vs-alpha,
+noise-null) + UPLIFT vs XS01 (la domanda di overlap) + plateau su L/k/direzione.
+
+CAVEAT ONESTO PRE-RISULTATO: e' market-neutral a 10 gambe -> NON eseguibile a $600 (STAT-MODE come
+XS01/VRP01). Il deliverable e' "esiste un carry reale e ORTOGONALE a XS01?", non un deploy.
+"""
+import sys
+from pathlib import Path
+import numpy as np, pandas as pd
+
+ROOT = Path(__file__).resolve().parents[2]
+sys.path.insert(0, str(ROOT))
+sys.path.insert(0, str(ROOT / "scripts" / "research" / "alt"))
+
+from src.portfolio.sleeves import XS_UNIVERSE, _xsec_returns, _vrp_combo_returns, _HL_DIR
+import altlib as L
+from altlib import _sh, _dd_ret, _to_daily, _uplift_series, HOLDOUT, marginal_vs_tp01
+
+FEE_SIDE = 0.0005   # 0.10% RT / 2, come XS01
+
+
+def _load_panel():
+    """Ritorna (PX, FUND): due DataFrame daily allineati [date x asset] con prezzo e funding
+    giornaliero realizzato (somma oraria). Solo asset con ENTRAMBI i dati."""
+    px, fund = {}, {}
+    for sym in XS_UNIVERSE:
+        pp = _HL_DIR / f"hl_{sym.lower()}_1d.parquet"
+        fp = _HL_DIR / f"hlfund_{sym.lower()}_1h.parquet"
+        if not (pp.exists() and fp.exists()):
+            continue
+        dp = pd.read_parquet(pp)
+        px[sym] = pd.Series(dp["close"].values.astype(float),
+                            index=pd.to_datetime(dp["timestamp"], unit="ms", utc=True)).resample("1D").last()
+        df = pd.read_parquet(fp)  # index ts (orario, utc), col funding
+        fund[sym] = df["funding"].resample("1D").sum()
+    PX = pd.concat(px, axis=1).sort_index()
+    FUND = pd.concat(fund, axis=1).sort_index().reindex(PX.index)
+    # tieni solo le date con prezzi per tutti + funding noto (0 dove manca un giorno isolato)
+    common = PX.dropna(how="any").index.intersection(FUND.index)
+    return PX.loc[common], FUND.loc[common].fillna(0.0)
+
+
+def carry_returns(L_lb=7, H=10, k=5, direction="carry", target_vol=0.20,
+                  fee_side=FEE_SIDE) -> pd.Series:
+    """Serie daily netta del book funding-carry cross-sectional. direction='carry' shorta l'alto
+    funding (incassa il premio); 'anti' lo compra."""
+    PX, FUND = _load_panel()
+    idx = PX.index
+    px = PX.values; fnd = FUND.values
+    n, A = px.shape
+    dret = np.vstack([np.zeros(A), px[1:] / px[:-1] - 1.0])
+    # segnale = media causale del funding giornaliero sugli ultimi L giorni (shiftato di 1)
+    sig = pd.DataFrame(fnd, index=idx).rolling(L_lb, min_periods=max(3, L_lb // 2)).mean().shift(1).values
+    W = np.zeros((n, A)); w = np.zeros(A)
+    for i in range(n):
+        if i >= L_lb + 1 and i % H == 0 and np.isfinite(sig[i]).sum() >= 2 * k:
+            s = sig[i].copy()
+            order = np.argsort(np.where(np.isfinite(s), s, np.nan))
+            order = order[np.isfinite(s[order])]
+            if len(order) >= 2 * k:
+                w = np.zeros(A)
+                lo, hi = order[:k], order[-k:]   # lo=basso funding, hi=alto funding
+                if direction == "carry":
+                    w[hi] = -0.5 / k; w[lo] = +0.5 / k     # SHORT alto funding (incassa), LONG basso
+                else:
+                    w[hi] = +0.5 / k; w[lo] = -0.5 / k
+        W[i] = w
+    # ritorno del perp per un LONG = price_ret - funding realizzato
+    perp_ret = dret - fnd
+    gross = np.zeros(n); gross[1:] = np.sum(W[:-1] * perp_ret[1:], axis=1)
+    turn = np.zeros(n); turn[0] = np.abs(W[0]).sum(); turn[1:] = np.abs(np.diff(W, axis=0)).sum(axis=1)
+    net = gross - turn * fee_side
+    s = pd.Series(net, index=idx)
+    rv = s.rolling(30, min_periods=15).std().shift(1) * np.sqrt(365.25)
+    scale = np.clip(np.nan_to_num(target_vol / rv.replace(0, np.nan).values, nan=0.0), 0, 3.0)
+    return pd.Series(s.values * scale, index=idx)
+
+
+def _stats(s: pd.Series) -> dict:
+    s = _to_daily(s)
+    h = s[s.index >= HOLDOUT]; isamp = s[s.index < HOLDOUT]
+    yrs = {y: round(_sh(s[s.index.year == y]), 2) for y in sorted(set(s.index.year))}
+    return dict(n=len(s), full=round(_sh(s), 2), insample=round(_sh(isamp), 2) if len(isamp) > 30 else None,
+                hold=round(_sh(h), 2) if len(h) > 30 else None, dd=round(_dd_ret(s), 3),
+                ann_ret=round(float(s.mean() * 365.25), 3), yearly=yrs)
+
+
+def main():
+    print("=" * 96)
+    print("  FC01 — FUNDING-CARRY cross-sectional su Hyperliquid (STAT-MODE)")
+    print("=" * 96)
+    PX, FUND = _load_panel()
+    print(f"  panel: {PX.shape[1]} asset x {len(PX)} giorni  [{PX.index[0].date()} .. {PX.index[-1].date()}]")
+    print(f"  asset: {list(PX.columns)}")
+    # funding medio annualizzato per asset (dispersione = materia prima del carry)
+    fann = (FUND.mean() * 365.25 * 100).round(1).sort_values(ascending=False)
+    print(f"  funding annualizzato% (carry teorico long-pays): "
+          f"max {fann.index[0]} {fann.iloc[0]:+.1f}  min {fann.index[-1]} {fann.iloc[-1]:+.1f}  "
+          f"mediana {fann.median():+.1f}  spread {fann.iloc[0]-fann.iloc[-1]:.1f}")
+
+    base = carry_returns()
+    print("\n  --- STANDALONE (config base L=7 H=10 k=5, direction=carry) ---")
+    st = _stats(base)
+    print(f"    FULL Sh {st['full']}  in-sample {st['insample']}  HOLD {st['hold']}  DD {st['dd']*100:.1f}%  "
+          f"ann.ret {st['ann_ret']*100:+.1f}%  ({st['n']}g)")
+    print(f"    per anno: {st['yearly']}")
+
+    # correlazioni vs gli sleeve attivi
+    print("\n  --- CORRELAZIONE vs sleeve attivi (daily, overlap comune) ---")
+    refs = {"TP01": L.tp01_baseline_daily(), "XS01": _to_daily(_xsec_returns()),
+            "VRP01": _to_daily(_vrp_combo_returns())}
+    bd = _to_daily(base)
+    for nm, r in refs.items():
+        J = pd.concat({"c": bd, "r": r}, axis=1, join="inner").dropna()
+        c = round(float(J["c"].corr(J["r"])), 3) if len(J) > 30 else None
+        print(f"    corr(FC01, {nm}) = {c}   (overlap {len(J)}g)")
+
+    # marginal vs TP01 (verdetto indurito completo)
+    print("\n  --- MARGINAL vs TP01 (scorer indurito) ---")
+    m = marginal_vs_tp01(bd)
+    for key in ("marginal_verdict", "corr_full", "cand_full_sharpe", "cand_insample_sharpe",
+                "has_insample_edge", "multicut_persistent", "is_hedge", "null_pctl_insample"):
+        print(f"    {key:22} = {m.get(key)}")
+    print(f"    blend w25: {m.get('blends', {}).get('w25')}")
+    print(f"    multicut_uplift: {m.get('multicut_uplift')}")
+
+    # UPLIFT vs XS01 (la domanda di overlap): XS01 da solo vs blend(XS01, FC01)
+    print("\n  --- UPLIFT vs XS01 (aggiunge a XS01 o e' ridondante?) ---")
+    xs = refs["XS01"]
+    J = pd.concat({"xs": xs, "fc": bd}, axis=1, join="inner").dropna()
+    JH = J[J.index >= HOLDOUT]
+    for w in (0.25, 0.5):
+        bf = _sh((1 - w) * J["xs"] + w * J["fc"]) - _sh(J["xs"])
+        bh = (_sh((1 - w) * JH["xs"] + w * JH["fc"]) - _sh(JH["xs"])) if len(JH) > 30 else None
+        print(f"    w={w}: uplift FULL {bf:+.3f}  HOLD {bh:+.3f}" if bh is not None
+              else f"    w={w}: uplift FULL {bf:+.3f}")
+    print(f"    XS01 standalone: FULL {_sh(J['xs']):.2f}  | FC01 standalone su overlap: {_sh(J['fc']):.2f}")
+
+    # PLATEAU su L, k, direzione
+    print("\n  --- PLATEAU (FULL / in-sample / HOLD Sharpe) ---")
+    print(f"    {'cfg':22} {'FULL':>6} {'in-s':>6} {'HOLD':>6} {'DD%':>6}")
+    for direction in ("carry", "anti"):
+        for Llb in (3, 7, 14, 30):
+            for k in (3, 5):
+                s = _stats(carry_returns(L_lb=Llb, k=k, direction=direction))
+                tag = f"{direction} L={Llb} k={k}"
+                print(f"    {tag:22} {str(s['full']):>6} {str(s['insample']):>6} {str(s['hold']):>6} {s['dd']*100:>5.1f}")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/fut_leadlag_generic.py

@@ -0,0 +1,108 @@
+"""LEAD-LAG GENERICO non-sovrapposto: crypto[T-S -> T] predice future[T -> T+H]? (ogni mercato/fuso).
+
+Market-agnostic. Per ogni giorno, entrata a ora T (UTC): segnale = crypto nella finestra [T-S, T]
+(finisce all'entrata), cattura = future nella finestra SUCCESSIVA [T, T+H] (non sovrapposta).
+Controllo = moto PROPRIO del future [T-S, T] -> isola se il crypto AGGIUNGE (anticipa) oltre il
+momentum del future. Sweep su T (copre apertura Europa ~07h, USA ~13h, Asia ~00h) e H.
+
+TRADE: sign(crypto[T-S,T]) * future[T,T+H] - costo. Sharpe sqrt(252), per-anno, OOS 2024+.
+Dati: data/raw/fut_*_1h.parquet (UTC) + crypto 1h. Solo cache, nessun IB.
+"""
+import sys, glob
+from pathlib import Path
+import numpy as np, pandas as pd
+
+ROOT = Path(__file__).resolve().parents[2]
+RAW = ROOT / "data" / "raw"
+sys.path.insert(0, str(ROOT)); sys.path.insert(0, str(ROOT / "scripts" / "research"))
+from crypto_lead_harness import crypto_hourly, at
+
+OOS = pd.Timestamp("2024-01-01", tz="UTC"); COST = 0.0002
+
+
+def fut_hourly(sym):
+    d = pd.read_parquet(RAW / f"fut_{sym.lower()}_1h.parquet")
+    return pd.Series(d["close"].astype(float).values, index=pd.to_datetime(d["timestamp"], unit="ms", utc=True)).sort_index()
+
+
+def _sh(r):
+    r = np.asarray(r, float); r = r[np.isfinite(r)]
+    return float(np.mean(r) / np.std(r) * np.sqrt(252)) if len(r) > 5 and np.std(r) > 0 else 0.0
+
+
+def run(fut, bc, T, S, H):
+    days = pd.date_range(fut.index[0].normalize(), fut.index[-1].normalize(), freq="D", tz="UTC")
+    rows = []
+    for D in days:
+        te = D + pd.Timedelta(hours=T)
+        ts = te - pd.Timedelta(hours=S)
+        tx = te + pd.Timedelta(hours=H)
+        f0, f1 = at(fut, ts), at(fut, te); f2 = at(fut, tx)
+        c0, c1 = at(bc, ts), at(bc, te)
+        if not all(np.isfinite(v) and v > 0 for v in (f0, f1, f2, c0, c1)):
+            continue
+        rows.append((D, c1/c0 - 1, f1/f0 - 1, f2/f1 - 1))
+    if len(rows) < 120:
+        return None
+    Dd = pd.DataFrame(rows, columns=["d", "csig", "fctrl", "cap"]).set_index("d")
+    # filtra giorni in cui la cattura e' identicamente 0 (mercato chiuso, prezzo stale)
+    Dd = Dd[Dd["cap"].abs() > 1e-9]
+    if len(Dd) < 120:
+        return None
+    x, ctrl, y = Dd["csig"].values, Dd["fctrl"].values, Dd["cap"].values
+
+    def z(a):
+        s = a.std(); return (a - a.mean()) / s if s > 0 else a * 0
+    X = np.column_stack([np.ones(len(y)), z(x), z(ctrl)])
+    beta, *_ = np.linalg.lstsq(X, z(y), rcond=None)
+    resid = z(y) - X @ beta; dof = max(len(y) - 3, 1)
+    se = np.sqrt(np.sum(resid**2) / dof * np.diag(np.linalg.inv(X.T @ X)))
+    t_c = float(beta[1] / se[1]) if se[1] > 0 else 0.0
+    C = np.sign(x) * y - COST
+    m = Dd.index >= OOS
+    yrs = Dd.index.year.values
+    py = {int(v): round(_sh(C[yrs == v]), 2) for v in sorted(set(yrs)) if (yrs == v).sum() >= 30}
+    pos = sum(1 for v in py.values() if v > 0)
+    return dict(T=T, H=H, n=len(Dd), t_crypto=round(t_c, 2), sh_full=round(_sh(C), 2),
+                sh_oos=round(_sh(C[m]), 2), ann=round(float(np.nanmean(C) * 252 * 100), 1),
+                years_pos=pos, years_tot=len(py), per_year=py)
+
+
+def main():
+    syms = sorted(p.name[4:-11].upper() for p in RAW.glob("fut_*_1h.parquet"))
+    print("=" * 98)
+    print(f"  LEAD-LAG GENERICO non-sovrapposto — futures: {syms}")
+    print("=" * 98)
+    print("  crypto[T-8h -> T] -> future[T -> T+6h], controllo=moto proprio future. net 2bps, OOS2024+\n")
+    bcs = {l: crypto_hourly(l) for l in ("BTC", "ETH")}
+    winners = []
+    for sym in syms:
+        fut = fut_hourly(sym)
+        best = None
+        for lead in ("BTC", "ETH"):
+            for T in (0, 4, 8, 12, 16, 20):
+                r = run(fut, bcs[lead], T, 8, 6)
+                if not r:
+                    continue
+                r["sym"] = sym; r["lead"] = lead
+                if best is None or r["sh_oos"] > best["sh_oos"]:
+                    best = r
+                # raccogli i forti (crypto significativo E robusto)
+                if r["t_crypto"] >= 2.5 and r["sh_oos"] > 0.5 and r["years_pos"] == r["years_tot"]:
+                    winners.append(r)
+        if best:
+            print(f"  {sym:7} miglior: {best['lead']}->T{best['T']}h: t_crypto {best['t_crypto']:+.1f} "
+                  f"Sh full {best['sh_full']:+.2f} OOS {best['sh_oos']:+.2f} ann {best['ann']:+.1f}% "
+                  f"{best['years_pos']}/{best['years_tot']}y {best['per_year']}")
+    print("\n  --- CANDIDATI FORTI (t_crypto>=2.5, OOS>0.5, tutti gli anni positivi) ---")
+    if not winners:
+        print("    NESSUNO. -> nessuna anticipazione crypto->future robusta oltre il rumore/beta.")
+    else:
+        winners.sort(key=lambda r: r["sh_oos"], reverse=True)
+        for w in winners[:10]:
+            print(f"    {w['sym']:7} {w['lead']}->T{w['T']}h H{w['H']}: t_crypto {w['t_crypto']} "
+                  f"Sh OOS {w['sh_oos']} full {w['sh_full']} ann {w['ann']}% {w['years_pos']}/{w['years_tot']}y {w['per_year']}")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/fut_overnight_leadlag.py

@@ -0,0 +1,116 @@
+"""TEST ONESTO finestre NON sovrapposte: crypto[notte presto] -> future indice[notte tardi].
+
+L'idea "monitor Deribit / trade IB" e' valida SOLO se il crypto anticipa il moto SUCCESSIVO del future
+(non lo stesso intervallo = look-ahead). Qui:
+  entrata a T (ora UTC nella notte, equity cash chiuso): osservato crypto fino a T.
+  segnale  = crypto[P 21:00 -> T]            (info nota a T)
+  controllo= future[P 21:00 -> T]            (il moto PROPRIO del future fino a T)
+  cattura  = future[T -> D 13:00]            (cio' che catturi entrando a T, fino a ~apertura cash)
+TEST: il segnale crypto predice la cattura INCREMENTALMENTE al controllo? (crypto vede prima del future?)
+TRADE: long/short future su sign(crypto[P21:00->T]), cattura future[T->open], net ~2bps. Sharpe/anno/OOS.
+Confronto col baseline sign(future[P21:00->T]) (momentum proprio del future) per isolare il valore crypto.
+
+Dati: data/raw/fut_<sym>_1h.parquet (UTC) + crypto 1h. sqrt(252).
+"""
+import sys, argparse, json
+from pathlib import Path
+import numpy as np, pandas as pd
+
+ROOT = Path(__file__).resolve().parents[2]
+RAW = ROOT / "data" / "raw"
+sys.path.insert(0, str(ROOT)); sys.path.insert(0, str(ROOT / "scripts" / "research"))
+from crypto_lead_harness import crypto_hourly, at
+
+OOS = pd.Timestamp("2024-01-01", tz="UTC"); COST = 0.0002
+CLOSE_H = 21; OPEN_H = 13
+
+
+def fut_hourly(sym):
+    d = pd.read_parquet(RAW / f"fut_{sym.lower()}_1h.parquet")
+    return pd.Series(d["close"].astype(float).values, index=pd.to_datetime(d["timestamp"], unit="ms", utc=True)).sort_index()
+
+
+def _sh(r):
+    r = np.asarray(r, float); r = r[np.isfinite(r)]
+    return float(np.mean(r) / np.std(r) * np.sqrt(252)) if len(r) > 5 and np.std(r) > 0 else 0.0
+
+
+def run(sym="ES", lead="BTC", h_entry=6):
+    fut = fut_hourly(sym); bc = crypto_hourly(lead)
+    days = pd.date_range(fut.index[0].normalize(), fut.index[-1].normalize(), freq="B", tz="UTC")
+    rows = []
+    for k in range(1, len(days)):
+        D = days[k]; P = days[k-1]
+        t_pc = P + pd.Timedelta(hours=CLOSE_H)        # P close 21:00
+        t_e = D + pd.Timedelta(hours=h_entry)          # entrata T
+        t_o = D + pd.Timedelta(hours=OPEN_H)           # ~apertura cash 13:00
+        fp, fe, fo = at(fut, t_pc), at(fut, t_e), at(fut, t_o)
+        cp, ce = at(bc, t_pc), at(bc, t_e)
+        if not all(np.isfinite(v) and v > 0 for v in (fp, fe, fo, cp, ce)):
+            continue
+        crypto_sig = ce / cp - 1.0          # crypto P21:00 -> T
+        fut_ctrl = fe / fp - 1.0            # future P21:00 -> T (controllo)
+        capture = fo / fe - 1.0             # future T -> open (cio' che catturi)
+        rows.append((D, crypto_sig, fut_ctrl, capture))
+    if len(rows) < 100:
+        return {"sym": sym, "lead": lead, "h_entry": h_entry, "err": f"n={len(rows)}"}
+    Dd = pd.DataFrame(rows, columns=["d", "csig", "fctrl", "cap"]).set_index("d")
+    x, ctrl, y = Dd["csig"].values, Dd["fctrl"].values, Dd["cap"].values
+
+    def z(a):
+        s = a.std(); return (a - a.mean()) / s if s > 0 else a * 0
+    # incrementale: cap ~ crypto + future_own_move
+    X = np.column_stack([np.ones(len(y)), z(x), z(ctrl)])
+    beta, *_ = np.linalg.lstsq(X, z(y), rcond=None)
+    resid = z(y) - X @ beta; dof = max(len(y) - 3, 1)
+    se = np.sqrt(np.sum(resid**2) / dof * np.diag(np.linalg.inv(X.T @ X)))
+    t_crypto = float(beta[1] / se[1]) if se[1] > 0 else 0.0
+    t_futown = float(beta[2] / se[2]) if se[2] > 0 else 0.0
+    # trade: crypto-signal vs future-own-signal
+    C_crypto = np.sign(x) * y - COST
+    C_futown = np.sign(ctrl) * y - COST
+    m = Dd.index >= OOS
+    yrs = Dd.index.year.values
+    py = {int(yv): round(_sh(C_crypto[yrs == yv]), 2) for yv in sorted(set(yrs)) if (yrs == yv).sum() >= 30}
+    return {"sym": sym, "lead": lead, "h_entry": h_entry, "n": len(Dd),
+            "corr_crypto_cap": round(float(np.corrcoef(x, y)[0, 1]), 3),
+            "t_crypto_incr": round(t_crypto, 2), "t_futown_incr": round(t_futown, 2),
+            "sh_crypto_full": round(_sh(C_crypto), 2), "sh_crypto_oos": round(_sh(C_crypto[m]), 2),
+            "sh_futown_full": round(_sh(C_futown), 2),
+            "ann_crypto_pct": round(float(np.nanmean(C_crypto) * 252 * 100), 1),
+            "per_year": py}
+
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument("--syms", default="ES,NQ,RTY")
+    ap.add_argument("--leads", default="BTC,ETH")
+    ap.add_argument("--entries", default="0,3,6,9")
+    args = ap.parse_args()
+    print("=" * 96)
+    print("  FUTURE OVERNIGHT — crypto[P21:00->T] predice future[T->open]? (finestre NON sovrapposte)")
+    print("=" * 96)
+    print("  cattura = future[T->open]; controllo = future[P21:00->T]; t_crypto_incr = crypto AL NETTO del future. net 2bps\n")
+    best = []
+    for sym in args.syms.split(","):
+        for lead in args.leads.split(","):
+            for h in [int(x) for x in args.entries.split(",")]:
+                r = run(sym, lead, h)
+                if "err" in r:
+                    print(f"  {sym}<-{lead} T={h}h: {r['err']}"); continue
+                print(f"  {sym}<-{lead} T={h:>2}h UTC: n={r['n']} corr {r['corr_crypto_cap']:+.3f} | "
+                      f"t_crypto(incr) {r['t_crypto_incr']:+.1f} t_futOwn {r['t_futown_incr']:+.1f} | "
+                      f"trade crypto Sh {r['sh_crypto_full']:+.2f} (OOS {r['sh_crypto_oos']:+.2f}) ann {r['ann_crypto_pct']:+.1f}% | "
+                      f"futOwn Sh {r['sh_futown_full']:+.2f}")
+                best.append(r)
+    best.sort(key=lambda r: r.get("sh_crypto_oos", -9), reverse=True)
+    print("\n  --- migliori per Sharpe OOS (trade su segnale crypto) ---")
+    for r in best[:5]:
+        print(f"    {r['sym']}<-{r['lead']} T={r['h_entry']}h: OOS {r['sh_crypto_oos']} full {r['sh_crypto_full']} "
+              f"t_crypto {r['t_crypto_incr']} per-anno {r['per_year']}")
+    print("\n  NB: t_crypto_incr alto E sh_crypto > sh_futOwn => il crypto anticipa il future (idea valida).")
+    print("      t_crypto ~0 o sh_crypto ~ futOwn => e' solo momentum del future, il crypto non aggiunge.")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/ib_equities_probe.py

@@ -0,0 +1,95 @@
+"""IB EQUITIES/ETF DATA PROBE — certifica cosa il paper IB dà per la ricerca su azioni/ETF.
+
+Gemello equity di certify_feed.py: PRIMA il dato (cosa c'è, quanto indietro, aggiustato per
+dividendi/split?, cosa costa), POI la strategia. Disciplina v2.0.0.
+
+Universo candidato per la prima ricerca equity (cross-sectional momentum / trend, l'edge "noioso e
+robusto" più plausibile in un mercato efficiente):
+  * 11 SPDR settoriali (XLK..XLC) — universo canonico del momentum cross-section settoriale;
+  * ETF broad / macro (SPY QQQ IWM TLT GLD HYG) — per trend e risk-on/off;
+  * 2 azioni (AAPL MSFT) per tarare profondità/qualità.
+
+Per ogni simbolo: profondità storica daily con whatToShow=ADJUSTED_LAST (split+dividendi, OBBLIGATORIO
+per un backtest equity onesto) e TRADES (raw), + flag se scatta errore di subscription market-data.
+
+  uv run --with ib_async python scripts/research/ib_equities_probe.py
+"""
+import argparse, sys
+
+SECTORS = ["XLK", "XLF", "XLE", "XLV", "XLI", "XLP", "XLY", "XLU", "XLB", "XLRE", "XLC"]
+BROAD = ["SPY", "QQQ", "IWM", "TLT", "GLD", "HYG"]
+STOCKS = ["AAPL", "MSFT"]
+
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument("--host", default="127.0.0.1")
+    ap.add_argument("--port", type=int, default=4002)
+    ap.add_argument("--client-id", type=int, default=88)
+    ap.add_argument("--years", default="20 Y", help="durata storica da richiedere")
+    args = ap.parse_args()
+
+    try:
+        from ib_async import IB, Stock
+    except Exception:
+        print("ib_async non importabile. Esegui con:  uv run --with ib_async python ...")
+        sys.exit(2)
+
+    ib = IB()
+    try:
+        ib.connect(args.host, args.port, clientId=args.client_id, timeout=15)
+    except Exception as e:
+        print(f"[CONNESSIONE FALLITA] {args.host}:{args.port} -> {repr(e)[:140]}")
+        sys.exit(1)
+
+    print("=" * 96)
+    print(f"  IB EQUITIES/ETF PROBE — {args.host}:{args.port} | acct {ib.managedAccounts()} | depth req {args.years}")
+    print("=" * 96)
+
+    universe = [("SECTOR", s) for s in SECTORS] + [("BROAD", s) for s in BROAD] + [("STOCK", s) for s in STOCKS]
+    print(f"  {'sym':6} {'tipo':7} {'ADJUSTED_LAST':>26} {'TRADES':>22}  note")
+    rows = []
+    for cat, sym in universe:
+        con = Stock(sym, "SMART", "USD")
+        try:
+            cds = ib.reqContractDetails(con)
+            if not cds:
+                print(f"  {sym:6} {cat:7} {'-- no contract --':>26}")
+                continue
+        except Exception as e:
+            print(f"  {sym:6} {cat:7} ERR resolve {repr(e)[:40]}")
+            continue
+
+        def hist(what):
+            try:
+                b = ib.reqHistoricalData(con, endDateTime="", durationStr=args.years,
+                                         barSizeSetting="1 day", whatToShow=what,
+                                         useRTH=True, formatDate=1, timeout=45)
+                if not b:
+                    return "0 barre", None
+                return f"{len(b)}b {b[0].date}..{b[-1].date}", b
+            except Exception as e:
+                return f"ERR {repr(e)[:30]}", None
+
+        adj_s, adj_b = hist("ADJUSTED_LAST")
+        trd_s, _ = hist("TRADES")
+        note = ""
+        if "ERR" in adj_s or "0 barre" in adj_s:
+            note = "subscription? prova delayed"
+        print(f"  {sym:6} {cat:7} {adj_s:>26} {trd_s:>22}  {note}")
+        if adj_b:
+            rows.append((sym, len(adj_b), str(adj_b[0].date), str(adj_b[-1].date)))
+
+    print("-" * 96)
+    if rows:
+        depth = min(r[1] for r in rows); start = max(r[2] for r in rows)
+        print(f"  CERTIFICABILI (ADJUSTED_LAST): {len(rows)}/{len(universe)} | profondità comune ~{depth}b | start comune {start}")
+        print(f"  -> per un backtest cross-sectional servono date allineate: lo start comune e' il limite.")
+    else:
+        print("  NESSUN simbolo ha reso storia ADJUSTED — probabile mancanza market-data subscription.")
+        print("  Ripiego: whatToShow='TRADES' (raw, non adj) o dati 'delayed' / fonte esterna certificabile.")
+    ib.disconnect()
+
+
+if __name__ == "__main__":
+    main()

scripts/research/ib_probe.py

@@ -0,0 +1,108 @@
+"""IB DATA PROBE — enumera cosa un conto paper Interactive Brokers espone (dati storici).
+
+NON e' una strategia: e' lo scan di FATTIBILITA' DATI, gemello di certify_feed.py per il mondo IB.
+Disciplina v2.0.0: prima il dato (cosa c'e', quanto indietro, che qualita', cosa costa), poi la
+strategia. "Solo dati, decido dopo".
+
+PREREQUISITO: una sessione IB Gateway/TWS PAPER loggata e raggiungibile (default 127.0.0.1:4002).
+  Tipico: avvii IB Gateway (Paper) sul tuo PC con API abilitata su 4002, poi reverse-tunnel
+  SSH verso questo server:  ssh -R 4002:localhost:4002 utente@server
+ESECUZIONE (senza sporcare le dipendenze del progetto):
+  uv run --with ib_async python scripts/research/ib_probe.py
+  uv run --with ib_async python scripts/research/ib_probe.py --port 7497   # TWS paper
+
+Cosa fa, in ordine, e si ferma con diagnosi chiara al primo errore:
+  (1) connette e stampa server version + account paper;
+  (2) risolve un set di contratti rilevanti per QUESTO progetto:
+        - CME crypto: BTC (5 BTC), MBT (micro 0.1 BTC), ETH, MET (micro);  -> per il BASIS/carry
+        - spot crypto Paxos (se abilitato): BTC, ETH;
+        - 2 azioni/ETF di riferimento (SPY, AAPL) per tarare durate/qualita';
+  (3) per ogni contratto risolto: chiede un piccolo storico (durata breve, 1 day bars) e riporta
+      n barre, range, e se scatta un errore di SUBSCRIPTION mancante (codice 354/10089/10090...);
+  (4) sintesi: cosa e' scaricabile GRATIS su paper vs cosa richiede market-data a pagamento.
+"""
+import argparse, sys
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument("--host", default="127.0.0.1")
+    ap.add_argument("--port", type=int, default=7497, help="7497=TWS paper (default), 4002=GW paper, 4001/7496=live")
+    ap.add_argument("--client-id", type=int, default=77)
+    args = ap.parse_args()
+
+    try:
+        from ib_async import IB, Future, Stock, Crypto, util
+    except Exception:
+        print("ib_async non importabile. Esegui con:  uv run --with ib_async python scripts/research/ib_probe.py")
+        sys.exit(2)
+
+    ib = IB()
+    try:
+        ib.connect(args.host, args.port, clientId=args.client_id, timeout=15)
+    except Exception as e:
+        print(f"[CONNESSIONE FALLITA] {args.host}:{args.port} -> {repr(e)[:160]}")
+        print("  Verifica: IB Gateway/TWS Paper acceso, API abilitata, porta giusta, tunnel attivo.")
+        sys.exit(1)
+
+    print("=" * 90)
+    print(f"  IB DATA PROBE — connesso {args.host}:{args.port} | serverVersion={ib.client.serverVersion()}")
+    try:
+        accts = ib.managedAccounts()
+        print(f"  account: {accts}  (paper se inizia per 'D' tipicamente)")
+    except Exception as e:
+        print(f"  account: ? ({repr(e)[:60]})")
+    print("=" * 90)
+
+    # (2) contratti rilevanti
+    candidates = []
+    # CME crypto futures: lasciamo che IB scelga il front-month (no expiry -> reqContractDetails)
+    candidates += [("CME BTC fut", Future("BTC", exchange="CME")),
+                   ("CME MBT micro", Future("MBT", exchange="CME")),
+                   ("CME ETH fut", Future("ETH", exchange="CME")),
+                   ("CME MET micro", Future("MET", exchange="CME"))]
+    # spot crypto Paxos (puo' non essere abilitato su paper)
+    candidates += [("Paxos BTC", Crypto("BTC", exchange="PAXOS", currency="USD")),
+                   ("Paxos ETH", Crypto("ETH", exchange="PAXOS", currency="USD"))]
+    # riferimenti equity
+    candidates += [("SPY ETF", Stock("SPY", "SMART", "USD")),
+                   ("AAPL", Stock("AAPL", "SMART", "USD"))]
+
+    resolved = []
+    print("\n  (A) RISOLUZIONE CONTRATTI")
+    for label, c in candidates:
+        try:
+            cds = ib.reqContractDetails(c)
+            if not cds:
+                print(f"    {label:16} -> NESSUN match")
+                continue
+            # per i futures prendi la scadenza piu' vicina disponibile
+            cd = sorted(cds, key=lambda d: getattr(d.contract, "lastTradeDateOrContractMonth", "") or "")[0]
+            con = cd.contract
+            extra = f" exp={con.lastTradeDateOrContractMonth}" if getattr(con, "lastTradeDateOrContractMonth", "") else ""
+            print(f"    {label:16} -> OK  {con.localSymbol or con.symbol} {con.exchange}{extra}  (n match={len(cds)})")
+            resolved.append((label, con))
+        except Exception as e:
+            print(f"    {label:16} -> ERR {repr(e)[:70]}")
+
+    # (3) prova storico breve
+    print("\n  (B) STORICO DI PROVA (durata 10 D, barre 1 day)")
+    for label, con in resolved:
+        try:
+            bars = ib.reqHistoricalData(con, endDateTime="", durationStr="10 D",
+                                        barSizeSetting="1 day", whatToShow="TRADES",
+                                        useRTH=False, formatDate=1, timeout=30)
+            if not bars:
+                print(f"    {label:16} -> 0 barre (forse serve subscription o whatToShow diverso)")
+            else:
+                print(f"    {label:16} -> {len(bars)} barre  {bars[0].date} .. {bars[-1].date}  close={bars[-1].close}")
+        except Exception as e:
+            print(f"    {label:16} -> ERR {repr(e)[:90]}")
+
+    print("\n  (C) NOTE")
+    print("    - errori 354/10089/10090/10168 = market-data subscription mancante (paper la eredita dal live).")
+    print("    - per il BASIS/carry servono i MULTIPLI futures (front+next) -> poi reqContractDetails senza filtro expiry.")
+    ib.disconnect()
+
+
+if __name__ == "__main__":
+    main()

scripts/research/intraday/fill_haircut.py

@@ -0,0 +1,153 @@
+"""fill_haircut — quanto il fill REALE a basso capitale erode il lead PREVDAY (e TP01)?
+
+Lo scettico d'esecuzione (diario 2026-06-21) ha segnalato che il vol-target di PREVDAY fa ~8500
+ribilanciamenti/anno, di cui 97-98% < $1 di nozionale a $600: a quel capitale NON puoi piazzare
+quegli ordini (min_order $5), quindi il libro MODELED (ribilanciamento continuo, frictionless) è
+una finzione. Il forward-monitor traccia MODELED-$2000 vs REAL-$600 per misurare il gap nei mesi
+a venire — qui lo stimiamo SUBITO su tutto lo storico, replicando la STESSA logica di paper_prevday.
+
+Due libri, identici tranne il fill:
+  MODELED : ribilancia ad ogni barra alla posizione-bersaglio (fee proporzionale su ogni |Δ|).
+  REAL-$C : capitale C, salta i ribilanciamenti con nozionale |Δpos|*leg_cap < min_order ($5)
+            (posizione resta "stale" -> tracking error, ma niente fee sui trade infinitesimi).
+
+Sweep capitale {600, 2000, 20000} per mostrare a quanto l'haircut svanisce. Poi la domanda-soldi:
+il blend 80%TP01+20%PREVDAY conserva l'uplift hold-out (+0.56 modellato) usando PREVDAY-REAL-$600?
+
+    uv run python scripts/research/intraday/fill_haircut.py
+"""
+import sys
+from pathlib import Path
+
+import numpy as np
+import pandas as pd
+
+ROOT = Path(__file__).resolve().parents[3]
+sys.path.insert(0, str(ROOT))
+from src.backtest.harness import load                       # noqa: E402
+from src.strategies.prevday_breakout import target as pv_target  # noqa: E402
+from src.portfolio.portfolio import to_daily                # noqa: E402
+from src.portfolio.sleeves import _tp01_returns             # noqa: E402
+
+HOLD = pd.Timestamp("2025-01-01", tz="UTC")
+FEE_SIDE = 0.0005          # 0.05%/side = 0.10% RT
+MIN_ORDER = 5.0
+WEIGHT = 0.5
+ASSETS = ["BTC", "ETH"]
+
+
+def _sh(x):
+    x = x.dropna()
+    return float(x.mean() / x.std() * np.sqrt(365.25)) if len(x) > 2 and x.std() > 0 else 0.0
+
+
+def _dd(x):
+    eq = (1 + x.fillna(0)).cumprod()
+    return float(((eq - eq.cummax()) / eq.cummax()).min())
+
+
+def simulate(targets: dict, rets: dict, idx_dt, capital):
+    """Bar-by-bar 50/50 book. capital=None -> MODELED (continuous). Returns (daily_ret, stats)."""
+    n = len(idx_dt)
+    held = {a: 0.0 for a in ASSETS}
+    net = np.zeros(n)
+    exec_ct = {a: 0 for a in ASSETS}
+    skip_ct = {a: 0 for a in ASSETS}
+    fee_tot = 0.0
+    for i in range(n):
+        step = 0.0
+        for a in ASSETS:
+            tgt = float(targets[a][i]); r = float(rets[a][i]); h = held[a]
+            if capital is None:                      # MODELED: always rebalance
+                new_h = tgt; traded = abs(tgt - h)
+                exec_ct[a] += 1 if traded > 1e-9 else 0
+            else:                                    # REAL-$C: skip sub-min_order
+                leg_cap = capital * WEIGHT
+                if abs(tgt - h) * leg_cap >= MIN_ORDER:
+                    new_h = tgt; traded = abs(tgt - h); exec_ct[a] += 1
+                else:
+                    new_h = h; traded = 0.0; skip_ct[a] += 1
+            fee = FEE_SIDE * traded
+            fee_tot += WEIGHT * fee
+            step += WEIGHT * (h * r - fee)           # earn on position HELD into bar, pay fee on rebalance
+            held[a] = new_h
+        net[i] = step
+    s = pd.Series(net, index=idx_dt)
+    daily = s.groupby(s.index.floor("1D")).sum()
+    yrs = (idx_dt[-1] - idx_dt[0]).days / 365.25
+    ex = sum(exec_ct.values()); sk = sum(skip_ct.values())
+    stats = dict(execs_per_yr=ex / yrs, skip_frac=sk / (ex + sk) if (ex + sk) else 0.0,
+                 fee_drag_per_yr=fee_tot / yrs)
+    return daily, stats
+
+
+def build_targets():
+    targets, rets, ts_sets = {}, {}, {}
+    dts = {}
+    for a in ASSETS:
+        df = load(a, "1h").reset_index(drop=True)
+        c = df["close"].values.astype(float)
+        r = np.zeros(len(c)); r[1:] = c[1:] / c[:-1] - 1.0
+        targets[a] = pv_target(df); rets[a] = r
+        ts = df["timestamp"].values.astype("int64")
+        ts_sets[a] = ts
+        dts[a] = pd.to_datetime(df["datetime"], utc=True).values
+    common = sorted(set(ts_sets["BTC"]).intersection(ts_sets["ETH"]))
+    pos = {a: {int(t): i for i, t in enumerate(ts_sets[a])} for a in ASSETS}
+    T, R = {a: [] for a in ASSETS}, {a: [] for a in ASSETS}
+    dt_out = []
+    for t in common:
+        i_btc = pos["BTC"][int(t)]
+        dt_out.append(dts["BTC"][i_btc])
+        for a in ASSETS:
+            i = pos[a][int(t)]
+            T[a].append(targets[a][i]); R[a].append(rets[a][i])
+    idx = pd.to_datetime(dt_out, utc=True)
+    return {a: np.array(T[a]) for a in ASSETS}, {a: np.array(R[a]) for a in ASSETS}, idx
+
+
+def row(label, daily):
+    J = daily.dropna(); JH = J[J.index >= HOLD]
+    yrs = (J.index[-1] - J.index[0]).days / 365.25
+    cagr = (1 + J).prod() ** (1 / yrs) - 1
+    return (f"  {label:<18s} FULL Sh {_sh(J):+5.2f}  HOLD Sh {_sh(JH):+5.2f}  "
+            f"CAGR {cagr*100:+5.1f}%  DD {_dd(J)*100:4.0f}%")
+
+
+def main():
+    print("=" * 92)
+    print("  FILL-HAIRCUT — PREVDAY: libro MODELED (continuo) vs REAL-$C (skip < $5 min-order)")
+    print("=" * 92)
+    T, R, idx = build_targets()
+    print(f"  path 1h: {len(idx)} barre  {idx[0].date()} -> {idx[-1].date()}\n")
+
+    books = {}
+    for cap, lab in [(None, "MODELED ($∞)"), (20000, "REAL-$20k"), (2000, "REAL-$2000"), (600, "REAL-$600")]:
+        daily, st = simulate(T, R, idx, cap)
+        books[lab] = daily
+        print(row(lab, daily) +
+              f"   | rebal/yr {st['execs_per_yr']:6.0f}  skip {st['skip_frac']*100:4.1f}%  "
+              f"fee-drag/yr {st['fee_drag_per_yr']*100:4.2f}%")
+
+    mod = books["MODELED ($∞)"]; real = books["REAL-$600"]
+    hc_full = _sh(mod.dropna()) - _sh(real.dropna())
+    JHm = mod[mod.index >= HOLD]; JHr = real[real.index >= HOLD]
+    hc_hold = _sh(JHm) - _sh(JHr)
+    print(f"\n  >> HAIRCUT $600 (MODELED - REAL): FULL Sharpe {hc_full:+.2f}  |  HOLD-OUT Sharpe {hc_hold:+.2f}")
+
+    # money question: does the blend uplift survive at REAL-$600?
+    print("\n" + "-" * 92)
+    print("  BLEND 80%TP01 + 20%PREVDAY — sopravvive l'uplift hold-out col fill reale?")
+    tp = to_daily(_tp01_returns())
+    for lab, pv in [("MODELED", mod), ("REAL-$600", real)]:
+        J = pd.concat({"TP": tp, "PV": pv}, axis=1).dropna(); JH = J[J.index >= HOLD]
+        for w in (0.20, 0.30):
+            b = (1 - w) * J["TP"] + w * J["PV"]; bh = (1 - w) * JH["TP"] + w * JH["PV"]
+            print(f"  PV={lab:<9s} w={w:.0%}  FULL {_sh(b):+.2f} (upl {_sh(b)-_sh(J['TP']):+.2f})  "
+                  f"HOLD {_sh(bh):+.2f} (upl {_sh(bh)-_sh(JH['TP']):+.2f})")
+    print(f"  [TP01 solo: FULL {_sh(tp.dropna()):+.2f}  HOLD {_sh(tp[tp.index>=HOLD]):+.2f}]")
+    print("=" * 92)
+
+
+if __name__ == "__main__":
+    main()

scripts/research/intraday/prevday_bootstrap.py

@@ -0,0 +1,142 @@
+"""prevday_bootstrap — l'edge di PREVDAY è coda-fortuna o persistente? (blocker #2/#3)
+
+CHIARIMENTO: il "top-5 giorni = 76-83% del PnL" del diario intraday era sulle GAMBE REVERT del combo
+a 5 segnali (vol_event/volume_spike/gap_fill), poi SCARTATE. Il sopravvissuto in forward-monitor è
+PREVDAY (breakout-continuation). Qui testiamo la concentrazione e la robustezza di PREVDAY STESSO —
+e in particolare della sua GAMBA SHORT, che (prevday_turnover) è l'intero valore di portafoglio.
+
+Due test:
+  A) CONCENTRAZIONE — quota del PnL nei top-K giorni (riproduce la metrica del diario su PREVDAY,
+     full / short-only / long-only, vs TP01 come riferimento: PREVDAY è PIÙ concentrato di ciò che
+     già deployamo?). + giorni per arrivare al 50% del guadagno cumulato.
+  B) CIRCULAR BLOCK BOOTSTRAP (blocchi da 20g, preserva autocorrelazione/regime) — distribuzione di:
+     standalone Sharpe (full + hold-out) e dell'UPLIFT hold-out del blend 80%TP01+20%PREVDAY (la
+     metrica-soldi). %>0 e 5° percentile = quanto l'edge dipende da quali blocchi sono capitati.
+
+    uv run python scripts/research/intraday/prevday_bootstrap.py
+"""
+import sys
+from pathlib import Path
+
+import numpy as np
+import pandas as pd
+
+ROOT = Path(__file__).resolve().parents[3]
+sys.path.insert(0, str(ROOT))
+from src.backtest.harness import load                       # noqa: E402
+from src.strategies import prevday_breakout as pb           # noqa: E402
+from src.portfolio.portfolio import to_daily                # noqa: E402
+from src.portfolio.sleeves import _tp01_returns             # noqa: E402
+
+HOLD = pd.Timestamp("2025-01-01", tz="UTC")
+FEE_SIDE = 0.0005
+WEIGHT = 0.5
+ASSETS = ["BTC", "ETH"]
+RNG = np.random.default_rng(12345)
+B = 3000
+BLOCK = 20
+
+
+def _sh(x):
+    x = np.asarray(x, float); x = x[np.isfinite(x)]
+    return float(x.mean() / x.std() * np.sqrt(365.25)) if len(x) > 2 and x.std() > 0 else 0.0
+
+
+def _leg_daily(dfs, leg):
+    """Ritorni daily 50/50 di PREVDAY restringendo la direzione: 'full'|'short'|'long'."""
+    out = None
+    for a in ASSETS:
+        df = dfs[a]
+        c = df["close"].values.astype(float)
+        r = np.zeros(len(c)); r[1:] = c[1:] / c[:-1] - 1.0
+        d = pb._breakout_direction(df, pb.ANCHOR_DAYS, pb.BUFFER_K, True)
+        if leg == "short":
+            d = np.minimum(d, 0.0)
+        elif leg == "long":
+            d = np.maximum(d, 0.0)
+        tgt = np.nan_to_num(pb._vol_target(d, df, pb.TARGET_VOL, pb.VOL_WIN_DAYS, pb.LEV_CAP), nan=0.0)
+        held = np.zeros(len(tgt)); held[1:] = tgt[:-1]
+        net = held * r - FEE_SIDE * np.abs(np.diff(tgt, prepend=tgt[0]))
+        s = pd.Series(net, index=pd.to_datetime(df["datetime"], utc=True))
+        dd = s.groupby(s.index.floor("1D")).sum()
+        out = dd if out is None else out.add(dd, fill_value=0)
+    return WEIGHT * out
+
+
+def concentration(daily, label):
+    s = daily.dropna()
+    total = s.sum()
+    pos = s[s > 0].sum()
+    topk = {k: s.nlargest(k).sum() / total if total != 0 else float("nan") for k in (5, 10, 20)}
+    # giorni (in ordine decrescente) per arrivare al 50% del guadagno lordo positivo
+    cum = s.sort_values(ascending=False).cumsum()
+    d50 = int((cum < 0.5 * pos).sum()) + 1 if pos > 0 else -1
+    n = len(s)
+    print(f"  {label:<22s} n={n}  totRet {total*100:+6.0f}%  "
+          f"top5 {topk[5]*100:4.0f}%  top10 {topk[10]*100:4.0f}%  top20 {topk[20]*100:4.0f}%  "
+          f"giorni->50% gain: {d50} ({d50/n*100:.1f}% dei giorni)")
+
+
+def block_boot_joint(tp, pv, n_iter=B, block=BLOCK):
+    """Bootstrap a blocchi circolari della serie CONGIUNTA (tp,pv) allineata. Ritorna i campioni di
+    (sh_pv, uplift_blend_80_20)."""
+    J = pd.concat({"TP": tp, "PV": pv}, axis=1, sort=True).dropna()
+    a = J["TP"].values; b = J["PV"].values
+    n = len(a)
+    nblocks = int(np.ceil(n / block))
+    sh_pv, upl = [], []
+    base_tp = _sh(a)
+    for _ in range(n_iter):
+        starts = RNG.integers(0, n, size=nblocks)
+        idx = np.concatenate([(np.arange(s, s + block) % n) for s in starts])[:n]
+        ta, tb = a[idx], b[idx]
+        sh_pv.append(_sh(tb))
+        blend = 0.8 * ta + 0.2 * tb
+        upl.append(_sh(blend) - _sh(ta))
+    return np.array(sh_pv), np.array(upl), base_tp
+
+
+def report_boot(name, sh, upl):
+    def q(x, p):
+        return float(np.percentile(x, p))
+    print(f"  {name}")
+    print(f"    PREVDAY Sharpe   : mediana {np.median(sh):+.2f}  [5°,95°]=[{q(sh,5):+.2f},{q(sh,95):+.2f}]  %>0 {np.mean(sh>0)*100:.0f}%")
+    print(f"    blend 80/20 UPLIFT: mediana {np.median(upl):+.2f}  [5°,95°]=[{q(upl,5):+.2f},{q(upl,95):+.2f}]  "
+          f"%>0 {np.mean(upl>0)*100:.0f}%  %>+0.10 {np.mean(upl>0.10)*100:.0f}%")
+
+
+def main():
+    print("=" * 100)
+    print("  PREVDAY bootstrap — l'edge è coda-fortuna o persistente? (blocco da 20g, B=%d)" % B)
+    print("=" * 100)
+    dfs = {a: load(a, "1h").reset_index(drop=True) for a in ASSETS}
+    pv_full = _leg_daily(dfs, "full")
+    pv_short = _leg_daily(dfs, "short")
+    pv_long = _leg_daily(dfs, "long")
+    tp = to_daily(_tp01_returns())
+
+    print("\n[A] CONCENTRAZIONE del PnL nei top-K giorni (più alto = più coda-fortuna):")
+    concentration(pv_full, "PREVDAY full")
+    concentration(pv_short, "PREVDAY short-only")
+    concentration(pv_long, "PREVDAY long-only")
+    concentration(tp, "TP01 (riferimento)")
+
+    print(f"\n[B] CIRCULAR BLOCK BOOTSTRAP — FULL ({pv_full.dropna().index.min().date()}->{pv_full.dropna().index.max().date()}):")
+    sh, upl, base = block_boot_joint(tp, pv_full)
+    print(f"    [TP01 base full Sharpe {base:+.2f}; uplift osservato +0.28 a w20]")
+    report_boot("full sample:", sh, upl)
+
+    print(f"\n[B] HOLD-OUT (2025+):")
+    tph = tp[tp.index >= HOLD]; pvh = pv_full[pv_full.index >= HOLD]
+    shH, uplH, baseH = block_boot_joint(tph, pvh)
+    print(f"    [TP01 base hold Sharpe {baseH:+.2f}; uplift osservato +0.56 a w20]")
+    report_boot("hold-out:", shH, uplH)
+
+    print(f"\n[B] SHORT-ONLY hold-out (la gamba che è tutto il valore):")
+    shS, uplS, _ = block_boot_joint(tph, pv_short[pv_short.index >= HOLD])
+    report_boot("short-only hold-out:", shS, uplS)
+    print("=" * 100)
+
+
+if __name__ == "__main__":
+    main()

scripts/research/intraday/prevday_turnover.py

@@ -0,0 +1,134 @@
+"""prevday_turnover — la fee di PREVDAY viene dai FLIP, non dai micro-ribilanciamenti. Si può tagliare?
+
+Da fill_haircut.py: il libro REAL-$600 salta il 98.4% dei ribilanciamenti del vol-target e la
+fee-drag scende solo 2.49% -> 2.39%/anno. Quindi la fee (~2.4%/anno) e' dominata dai ~50 FLIP di
+direzione/anno, non dal churn sub-dollaro. Un deadband d'esecuzione e' inutile; la leva e' ridurre i
+flip a LIVELLO DI SEGNALE. Qui sweepiamo le leve che riducono i flip e misuriamo il trade-off
+turnover <-> edge:
+  * BUFFER_K (break piu' deciso = meno flip)              {0.30 base, 0.50, 0.75, 1.00}
+  * ANCHOR_DAYS (range multi-giorno = livelli piu' larghi){1 base, 2, 3, 5}
+  * MIN_HOLD (non flippare entro N ore dall'ultimo flip)  {0 base, 24, 72}
+e in piu' LONG-ONLY vs LONG-SHORT (isola la gamba short = l'hedge del blocker #1).
+
+Per ogni config: flip/anno, fee-drag/anno, FULL/HOLD Sharpe, corr a TP01, e l'uplift hold-out del
+blend 80%TP01+20%PV (la metrica che conta). Libro MODELED (l'haircut di fill e' +0.01, irrilevante).
+
+    uv run python scripts/research/intraday/prevday_turnover.py
+"""
+import sys
+from pathlib import Path
+
+import numpy as np
+import pandas as pd
+
+ROOT = Path(__file__).resolve().parents[3]
+sys.path.insert(0, str(ROOT))
+from src.backtest.harness import load                       # noqa: E402
+from src.strategies import prevday_breakout as pb           # noqa: E402
+from src.portfolio.portfolio import to_daily                # noqa: E402
+from src.portfolio.sleeves import _tp01_returns             # noqa: E402
+
+HOLD = pd.Timestamp("2025-01-01", tz="UTC")
+FEE_SIDE = 0.0005
+WEIGHT = 0.5
+ASSETS = ["BTC", "ETH"]
+
+
+def _sh(x):
+    x = x.dropna()
+    return float(x.mean() / x.std() * np.sqrt(365.25)) if len(x) > 2 and x.std() > 0 else 0.0
+
+
+def _dd(x):
+    eq = (1 + x.fillna(0)).cumprod()
+    return float(((eq - eq.cummax()) / eq.cummax()).min())
+
+
+def _min_hold(direction, min_hold_bars):
+    """Sopprime i flip entro min_hold_bars dall'ultimo cambio di segno (riduce il churn di segnale)."""
+    if min_hold_bars <= 0:
+        return direction
+    out = direction.copy()
+    last_flip = -10**9
+    cur = out[0]
+    for i in range(len(out)):
+        if np.sign(out[i]) != np.sign(cur):
+            if i - last_flip >= min_hold_bars:
+                cur = out[i]; last_flip = i
+            else:
+                out[i] = cur               # mantieni la posizione (flip soppresso)
+        else:
+            cur = out[i]
+    return out
+
+
+def build(dfs, anchor, k, min_hold_bars, allow_short):
+    """Ritorni daily 50/50 + flip/anno + fee-drag/anno per una config di segnale (libro modeled)."""
+    legs, idx_ref = [], None
+    flips = 0; fee_tot = 0.0; yrs = None
+    daily_sum = None
+    for a in ASSETS:
+        df = dfs[a]
+        c = df["close"].values.astype(float)
+        r = np.zeros(len(c)); r[1:] = c[1:] / c[:-1] - 1.0
+        direction = pb._breakout_direction(df, anchor, k, allow_short)
+        direction = _min_hold(direction, min_hold_bars)
+        tgt = pb._vol_target(direction, df, pb.TARGET_VOL, pb.VOL_WIN_DAYS, pb.LEV_CAP)
+        tgt = np.nan_to_num(tgt, nan=0.0)
+        held = np.zeros(len(tgt)); held[1:] = tgt[:-1]
+        turn = np.abs(np.diff(tgt, prepend=tgt[0]))
+        net = held * r - FEE_SIDE * turn
+        dt = pd.to_datetime(df["datetime"], utc=True)
+        s = pd.Series(net, index=dt)
+        d = s.groupby(s.index.floor("1D")).sum()
+        daily_sum = d if daily_sum is None else daily_sum.add(d, fill_value=0)
+        flips += int((np.sign(direction[1:]) != np.sign(direction[:-1])).sum())
+        fee_tot += float((FEE_SIDE * turn).sum())
+        yrs = (dt.iloc[-1] - dt.iloc[0]).days / 365.25
+    daily = WEIGHT * daily_sum
+    return daily, flips / 2 / yrs, WEIGHT * fee_tot / yrs    # flips mediati sulle 2 gambe
+
+
+def main():
+    print("=" * 104)
+    print("  PREVDAY turnover-reduction — la fee viene dai FLIP. Si taglia senza perdere l'edge?")
+    print("=" * 104)
+    dfs = {a: load(a, "1h").reset_index(drop=True) for a in ASSETS}
+    tp = to_daily(_tp01_returns())
+
+    def line(label, daily, flips, fee):
+        J = pd.concat({"TP": tp, "PV": daily}, axis=1, sort=True).dropna(); JH = J[J.index >= HOLD]
+        b = 0.8 * J["TP"] + 0.2 * J["PV"]; bh = 0.8 * JH["TP"] + 0.2 * JH["PV"]
+        upl_h = _sh(bh) - _sh(JH["TP"])
+        print(f"  {label:<26s} flip/yr {flips:5.0f}  fee {fee*100:4.2f}%  "
+              f"FULL {_sh(J['PV']):+5.2f}  HOLD {_sh(JH['PV']):+5.2f}  DD {_dd(J['PV'])*100:4.0f}%  "
+              f"corrTP {J['PV'].corr(J['TP']):+.2f}  blendHOLDupl {upl_h:+.2f}")
+
+    print(f"  [TP01 solo: FULL {_sh(tp.dropna()):+.2f}  HOLD {_sh(tp[tp.index>=HOLD]):+.2f}]\n")
+
+    print("  -- BASE (congelato: anchor=1, k=0.30, no min-hold, long-short) --")
+    d, f, fee = build(dfs, 1, 0.30, 0, True); line("BASE", d, f, fee)
+
+    print("\n  -- BUFFER_K piu' ampio (break piu' deciso) --")
+    for k in (0.50, 0.75, 1.00):
+        d, f, fee = build(dfs, 1, k, 0, True); line(f"k={k:.2f}", d, f, fee)
+
+    print("\n  -- ANCHOR_DAYS multi-giorno (range piu' largo) --")
+    for an in (2, 3, 5):
+        d, f, fee = build(dfs, an, 0.30, 0, True); line(f"anchor={an}", d, f, fee)
+
+    print("\n  -- MIN_HOLD (no flip entro N ore) --")
+    for mh in (24, 72):
+        d, f, fee = build(dfs, 1, 0.30, mh, True); line(f"min_hold={mh}h", d, f, fee)
+
+    print("\n  -- combo low-turnover (k=0.75 + anchor=2 + min_hold=24h) --")
+    d, f, fee = build(dfs, 2, 0.75, 24, True); line("combo-LT", d, f, fee)
+
+    print("\n  -- LONG-ONLY vs LONG-SHORT (isola la gamba short = hedge del blocker #1) --")
+    d, f, fee = build(dfs, 1, 0.30, 0, False); line("long-only (no short)", d, f, fee)
+    d, f, fee = build(dfs, 1, 0.30, 0, True);  line("long-short (BASE)", d, f, fee)
+    print("=" * 104)
+
+
+if __name__ == "__main__":
+    main()

scripts/research/skyhook/check_v1.py

@@ -0,0 +1,13 @@
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+import skyhooklib as sk
+from src.strategies.skyhook import SkyhookParams
+V1 = SkyhookParams(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35, vola_hi=95, vol_lo=0.0)
+rep = sk.study("SKH01-V1", V1)
+print(sk.fmt(rep))
+print("causality:", sk.causality(V1))
+print("\n--- marginal vs TP01 (does it ADD as a sleeve?) ---")
+import altlib as al
+print(al.fmt_marginal(dict(name="SKH01-V1", marginal=sk.marginal(V1),
+      abs_grade=rep["verdict"]["grade"], marginal_verdict=sk.marginal(V1).get("marginal_verdict"),
+      earns_slot=False)))

scripts/research/skyhook/grid.py

@@ -0,0 +1,26 @@
+"""Combined grid over the scout-winning levers -> rank by min-asset HOLD-OUT (gate minFull>=0.5)."""
+import sys, itertools
+from dataclasses import replace
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+import skyhooklib as sk
+from src.strategies.skyhook import SkyhookParams
+
+base = SkyhookParams()
+def quick(p):
+    rs = {a: sk.run_asset(a, p, sk.FEE_RT) for a in sk.CERTIFIED}
+    return (min(rs[a]["full"]["sharpe"] for a in rs),
+            min(rs[a]["holdout"]["sharpe"] for a in rs),
+            min(rs[a]["full"]["n_trades"] for a in rs),
+            round(sum(rs[a]["full"]["maxdd"] for a in rs)/2,3))
+
+rows=[]
+for ptn_n,(sl,tp),vol_lo,(vlo,vhi) in itertools.product(
+        (8,21,55), ((2.0,5.0),(2.5,6.0),(3.0,8.0)), (0.0,40.0,50.0), ((35.0,95.0),(25.0,95.0))):
+    p=replace(base, ptn_n=ptn_n, sl_atr=sl, tp_atr=tp, vol_lo=vol_lo, vola_lo=vlo, vola_hi=vhi)
+    mf,mh,mt,dd=quick(p)
+    rows.append((mh,mf,mt,dd,ptn_n,sl,tp,vol_lo,vlo,vhi))
+rows.sort(reverse=True)
+print(f"{'minH':>6s}{'minF':>6s}{'tr':>5s}{'dd':>5s}  ptn  sl   tp  vlo vola")
+for mh,mf,mt,dd,ptn_n,sl,tp,vol_lo,vlo,vhi in rows[:18]:
+    gate = "PASS" if (mf>=0.5 and mh>=0.2 and mt>=20) else ""
+    print(f"{mh:>+6.2f}{mf:>+6.2f}{mt:>5d}{dd*100:>4.0f}%  {ptn_n:>3d} {sl:>3.1f} {tp:>4.1f} {vol_lo:>4.0f} [{vlo:.0f},{vhi:.0f}] {gate}")

scripts/research/skyhook/runs/SKH2_ASYM_LS.py

@@ -0,0 +1,289 @@
+"""SKH2_ASYM_LS — long/short RISK ASYMMETRY family (Skyhook DD-cut wave).
+
+Hypothesis: shorts are essential (prior finding) but they carry the standalone draw-down —
+in crypto a short gets steamrolled by a vol-up move. Keep the verified V2-winner risk on the
+LONG side, but put TIGHTER risk on the SHORT side: a shorter time-stop (uscitashort) and/or a
+tighter SL (smaller sl_atr, or a fixed 'pct' SL), and a leaner TP so shorts take profit fast
+instead of bleeding into a reversal.
+
+SkyhookParams has uscitalong/uscitashort but a SINGLE sl_atr/tp_atr, so direction-asymmetric
+STOPS require CUSTOM entries. We reuse the engine's regime+pattern signal (htf_features +
+merge_htf_to_ltf) UNCHANGED — only the per-direction (sl, tp, max_bars) differ. This is causal:
+the only thing that depends on direction is the offset magnitude applied to close[i]; the SIGNAL
+(comp_long/comp_short) is computed exactly as the verified winner.
+
+Causality: proven by truncated-prefix recompute on the CUSTOM entries (same scheme as
+sk.causality): an entry emitted on a prefix must match the full-run entry at that index.
+"""
+from __future__ import annotations
+
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+
+import numpy as np
+import pandas as pd
+
+import skyhooklib as sk
+import altlib as al
+from src.backtest.harness import backtest_signals
+from src.strategies import skyhook as S
+from src.strategies.skyhook import SkyhookParams
+
+HOLDOUT = sk.HOLDOUT
+FEE = sk.FEE_RT
+
+# Verified V2 winner signal config (the regime/pattern gate we keep).
+WIN = dict(ptn_n=45, vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+# Winner's symmetric risk (used for longs, and as the symmetric reference):
+WIN_RISK = dict(sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16)
+
+
+# ---------------------------------------------------------------------------
+# Custom asymmetric entries. Longs keep `long_risk`; shorts use `short_risk`.
+# Risk dicts: {'mode':'atr'|'pct', 'sl':..., 'tp':..., 'mb':int}
+#   atr -> sl/tp are ATR multiples ; pct -> sl/tp are fractions of close.
+# ---------------------------------------------------------------------------
+def asym_entries(ltf, htf, base_p: SkyhookParams, long_risk: dict, short_risk: dict) -> list:
+    feat = S.htf_features(htf, base_p)
+    m = S.merge_htf_to_ltf(ltf, feat)
+    c = m["close"].values.astype(float)
+    a = S.atr(m, base_p.ltf_atr_win)
+    comp_long = np.nan_to_num(m["comp_long"].values).astype(bool)
+    comp_short = np.nan_to_num(m["comp_short"].values).astype(bool)
+    days = pd.to_datetime(m["datetime"]).dt.floor("D").values
+
+    entries = [None] * len(m)
+    count_today: dict = {}
+    for i in range(len(m)):
+        if not np.isfinite(a[i]) or a[i] <= 0:
+            continue
+        day = days[i]
+        if count_today.get(day, 0) >= base_p.max_per_day:
+            continue
+        if comp_long[i]:
+            direction, rk = 1, long_risk
+        elif comp_short[i]:
+            direction, rk = -1, short_risk
+        else:
+            continue
+        if rk["mode"] == "atr":
+            sl_off, tp_off = rk["sl"] * a[i], rk["tp"] * a[i]
+        else:
+            sl_off, tp_off = rk["sl"] * c[i], rk["tp"] * c[i]
+        if direction == 1:
+            sl, tp = c[i] - sl_off, c[i] + tp_off
+        else:
+            sl, tp = c[i] + sl_off, c[i] - tp_off
+        entries[i] = {"dir": direction, "tp": float(tp), "sl": float(sl), "max_bars": int(rk["mb"])}
+        count_today[day] = count_today.get(day, 0) + 1
+    return entries
+
+
+# ---------------------------------------------------------------------------
+# Causality on the CUSTOM entries: prefix recompute must match the full run.
+# ---------------------------------------------------------------------------
+def causality_struct(base_p, long_risk, short_risk, asset="BTC", tail=200) -> dict:
+    ltf, htf = sk.frames(asset)
+    full = asym_entries(ltf, htf, base_p, long_risk, short_risk)
+    n = len(ltf)
+    bad = 0
+    checked = 0
+    for frac in (0.80, 0.92):
+        cut = int(n * frac)
+        cut_ts = int(ltf["timestamp"].iloc[cut - 1])
+        htf_cut = htf[htf["timestamp"] <= cut_ts].reset_index(drop=True)
+        sub = asym_entries(ltf.iloc[:cut].reset_index(drop=True), htf_cut, base_p, long_risk, short_risk)
+        for i in range(max(0, cut - tail), cut):
+            checked += 1
+            x, y = full[i], sub[i]
+            if (x is None) != (y is None):
+                bad += 1
+            elif x is not None and (x["dir"] != y["dir"]
+                                    or abs(x["sl"] - y["sl"]) > 1e-6
+                                    or abs(x["tp"] - y["tp"]) > 1e-6
+                                    or x["max_bars"] != y["max_bars"]):
+                bad += 1
+    return dict(ok=bool(bad == 0), mismatches=int(bad), checked=int(checked))
+
+
+def _split(eq, idx, mask):
+    e = eq[mask]
+    if len(e) < 5:
+        return dict(sharpe=0.0, ret=0.0, maxdd=0.0)
+    r = np.diff(e) / e[:-1]
+    r = r[np.isfinite(r)]
+    ix = idx[mask]
+    dt = pd.Series(ix).diff().dt.total_seconds().median() or 86400
+    bpy = 86400 * 365.25 / dt
+    sh = float(np.mean(r) / np.std(r) * np.sqrt(bpy)) if len(r) and np.std(r) > 0 else 0.0
+    pk = np.maximum.accumulate(e)
+    dd = float(np.max((pk - e) / pk))
+    return dict(sharpe=round(sh, 3), ret=round(float(e[-1] / e[0] - 1), 4), maxdd=round(dd, 4))
+
+
+def study_asym(name, base_p, long_risk, short_risk):
+    per_asset = {}
+    fee_ok_all = True
+    for a in ("BTC", "ETH"):
+        ltf, htf = sk.frames(a)
+        ent = asym_entries(ltf, htf, base_p, long_risk, short_risk)
+        m = backtest_signals(ltf, ent, fee_rt=FEE, leverage=1.0, asset=a, tf="230m")
+        eq = m.equity
+        idx = pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True))
+        hmask = np.asarray(idx >= HOLDOUT)
+        full = dict(sharpe=round(m.sharpe, 3), ret=round(m.net_return, 4), maxdd=round(m.max_dd, 4),
+                    n_trades=int(m.n_trades), win_rate=round(m.win_rate, 1))
+        hold = _split(eq, idx, hmask)
+        sweep = {}
+        for f in (0.0, 0.001, 0.002, 0.003):
+            mf = backtest_signals(ltf, ent, fee_rt=f, leverage=1.0, asset=a, tf="230m")
+            sweep[f"{f*100:.2f}%"] = round(mf.sharpe, 3)
+        fee_ok_all = fee_ok_all and (sweep["0.30%"] > 0)
+        # short-only vs long-only DD diagnostic
+        per_asset[a] = dict(full=full, hold=hold, fee_sweep=sweep,
+                            yearly={int(y): round(v, 4) for y, v in m.yearly.items()})
+    mf = min(per_asset[a]["full"]["sharpe"] for a in per_asset)
+    mh = min(per_asset[a]["hold"]["sharpe"] for a in per_asset)
+    mt = min(per_asset[a]["full"]["n_trades"] for a in per_asset)
+    mdd = max(per_asset[a]["full"]["maxdd"] for a in per_asset)
+    grade = "PASS" if (mt >= 20 and mf >= 0.5 and mh >= 0.2 and fee_ok_all) else \
+            ("WEAK" if (mt >= 20 and mf >= 0.3 and mh >= 0.0) else "FAIL")
+    return dict(grade=grade, minFull=mf, minHold=mh, minTr=mt, maxDD=mdd, fee_ok=fee_ok_all,
+                per_asset=per_asset)
+
+
+def daily_5050(base_p, long_risk, short_risk):
+    series = {}
+    for a in ("BTC", "ETH"):
+        ltf, htf = sk.frames(a)
+        ent = asym_entries(ltf, htf, base_p, long_risk, short_risk)
+        m = backtest_signals(ltf, ent, fee_rt=FEE, leverage=1.0, asset=a, tf="230m")
+        s = pd.Series(m.equity, index=pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True)))
+        series[a] = s.resample("1D").last().ffill().pct_change().dropna()
+    J = pd.concat(series, axis=1, join="inner").fillna(0.0)
+    return 0.5 * J["BTC"] + 0.5 * J["ETH"]
+
+
+def marginal_asym(base_p, long_risk, short_risk):
+    return al.marginal_vs_tp01(daily_5050(base_p, long_risk, short_risk))
+
+
+def print_study(name, r):
+    print(f"\n=== {name} -> {r['grade']}  (minFull={r['minFull']:+.2f} minHold={r['minHold']:+.2f}"
+          f" minTr={r['minTr']} maxDD={r['maxDD']*100:.0f}% feeOK={r['fee_ok']})")
+    for a, pa in r["per_asset"].items():
+        yr = " ".join(f"{y}:{v*100:+.0f}%" for y, v in pa["yearly"].items())
+        print(f"  {a}: FULL Sh={pa['full']['sharpe']:+.2f} ret={pa['full']['ret']*100:+.0f}%"
+              f" DD={pa['full']['maxdd']*100:.0f}% n={pa['full']['n_trades']} wr={pa['full']['win_rate']:.0f}%"
+              f" | HOLD Sh={pa['hold']['sharpe']:+.2f} ret={pa['hold']['ret']*100:+.0f}% DD={pa['hold']['maxdd']*100:.0f}%")
+        print(f"      fee: " + " ".join(f"{k}={v:+.2f}" for k, v in pa["fee_sweep"].items()))
+        print(f"      yr: {yr}")
+
+
+if __name__ == "__main__":
+    base_p = SkyhookParams(**WIN, **WIN_RISK)  # signal + winner risk (used for shape only)
+
+    # ---- 0) REFERENCE: rebuild the verified symmetric winner via our custom path -------
+    long_winner = dict(mode="atr", sl=2.5, tp=7.0, mb=24)
+    sym_short = dict(mode="atr", sl=2.5, tp=7.0, mb=16)
+    rREF = study_asym("REF symmetric-winner (rebuilt)", base_p, long_winner, sym_short)
+    print_study("REF symmetric-winner (rebuilt)", rREF)
+
+    # Long side: ROUND 1 showed pct-SL shorts lift everything but ETH DD sticks ~30.5%.
+    # The standalone DD comes from BOTH directions, so we also tighten the LONG pct-SL a
+    # touch to bring the combined DD under 30 while keeping the winner's long TP behaviour.
+    # We test two long variants: the verified winner (atr) AND a pct long.
+    long_variants = [
+        ("Latr", dict(mode="atr", sl=2.5, tp=7.0, mb=24)),
+        ("Lpct", dict(mode="pct", sl=0.04, tp=0.10, mb=24)),
+    ]
+
+    # ---- 1) GRID over asymmetric SHORT risk: pct-SL family is the winner; push SL tighter
+    # to knock ETH DD under 30. Keep the strong tp=0.08 and a couple of mb / SL choices.
+    short_grid = []
+    for mb_s in (12, 14, 16):
+        for slp in (0.02, 0.025, 0.03):
+            for tpp in (0.06, 0.08):
+                short_grid.append((f"Spct_mb{mb_s}_sl{slp}_tp{tpp}",
+                                   dict(mode="pct", sl=slp, tp=tpp, mb=mb_s)))
+    # a few tight-ATR shorts for completeness
+    for mb_s in (12, 14):
+        for sl_s in (1.5, 2.0):
+            short_grid.append((f"Satr_mb{mb_s}_sl{sl_s}_tp5.0",
+                               dict(mode="atr", sl=sl_s, tp=5.0, mb=mb_s)))
+
+    candidates = []
+    for lname, lr in long_variants:
+        for sname, sr in short_grid:
+            candidates.append((f"{lname}|{sname}", lr, sr))
+
+    results = []
+    for name, lr, sr in candidates:
+        r = study_asym(name, base_p, lr, sr)
+        results.append((name, lr, sr, r))
+
+    # Rank: feasible (grade != FAIL, fee ok) by lowest DD, then highest minHold.
+    feas = [(n, lr, sr, r) for n, lr, sr, r in results if r["grade"] != "FAIL"]
+    feas.sort(key=lambda t: (t[3]["maxDD"], -t[3]["minHold"]))
+
+    print("\n\n##### GRID RANK (feasible, by lowest standalone maxDD) #####")
+    for n, lr, sr, r in feas[:16]:
+        print(f"  {n:28s} DD={r['maxDD']*100:4.0f}% minFull={r['minFull']:+.2f}"
+              f" minHold={r['minHold']:+.2f} minTr={r['minTr']} grade={r['grade']}")
+
+    # ---- 2) Detailed study + marginal on the top DD-cutters that keep hold-out ---------
+    # pick best candidates: DD<30 with decent hold-out
+    qualifying = [t for t in feas if t[3]["maxDD"] < 0.30 and t[3]["minHold"] >= 0.50]
+    qualifying.sort(key=lambda t: (t[3]["maxDD"], -t[3]["minHold"]))
+    probe = qualifying[:5] if qualifying else feas[:5]
+
+    print("\n\n##### DETAIL + MARGINAL on top probes #####")
+    best = None
+    for n, long_risk, sr, r in probe:
+        print_study(n, r)
+        caus = causality_struct(base_p, long_risk, sr, "BTC")
+        caus_e = causality_struct(base_p, long_risk, sr, "ETH")
+        mg = marginal_asym(base_p, long_risk, sr)
+        w25 = mg.get("blends", {}).get("w25", {})
+        w50 = mg.get("blends", {}).get("w50", {})
+        earns = (r["grade"] != "FAIL" and mg.get("marginal_verdict") == "ADDS"
+                 and mg.get("robust_oos") and not mg.get("is_hedge"))
+        beats = bool(earns and r["maxDD"] < 0.30
+                     and (w25.get("uplift_hold") or -9) >= 0.55
+                     and r["minHold"] >= 0.65)
+        print(f"  CAUSALITY BTC={caus} ETH={caus_e}")
+        print(f"  MARGINAL: verdict={mg.get('marginal_verdict')} corr_full={mg.get('corr_full')}"
+              f" insample_edge={mg.get('has_insample_edge')} cand_is_sh={mg.get('cand_insample_sharpe')}"
+              f" hedge={mg.get('is_hedge')} robust_oos={mg.get('robust_oos')}"
+              f" multicut={mg.get('multicut_persistent')} cleanYr={mg.get('clean_year_uplift')}")
+        print(f"  BLEND w25: uplift_hold={w25.get('uplift_hold')} uplift_full={w25.get('uplift_full')}"
+              f" | w50: full={w50.get('full')} hold={w50.get('hold')} dd={w50.get('dd')}")
+        print(f"  => earns_slot={earns}  beats_winner={beats}")
+        cand = dict(name=n, long_risk=long_risk, short_risk=sr, study=r, caus=caus, caus_e=caus_e,
+                    mg=mg, earns=earns, beats=beats)
+        # prefer beats; else lowest-DD earns; else lowest-DD feasible
+        if best is None:
+            best = cand
+        else:
+            key = lambda x: (x["beats"], x["earns"], -x["study"]["maxDD"], x["study"]["minHold"])
+            if key(cand) > key(best):
+                best = cand
+
+    print("\n\n##### FINAL BEST #####")
+    b = best
+    r = b["study"]
+    mg = b["mg"]
+    w25 = mg.get("blends", {}).get("w25", {})
+    w50 = mg.get("blends", {}).get("w50", {})
+    print(f"BEST CONFIG: signal={WIN} long_risk={b['long_risk']} short_risk={b['short_risk']}")
+    print(f"  minFull={r['minFull']:+.2f} minHold={r['minHold']:+.2f} maxDD={r['maxDD']*100:.0f}%"
+          f" minTrades={r['minTr']} fee@0.30%_ok={r['fee_ok']}")
+    print(f"  causality BTC={b['caus']} ETH={b['caus_e']}")
+    print(f"  marginal: verdict={mg.get('marginal_verdict')} corr_full={mg.get('corr_full')}"
+          f" insample_edge={mg.get('has_insample_edge')} hedge={mg.get('is_hedge')}"
+          f" robust_oos={mg.get('robust_oos')} multicut={mg.get('multicut_persistent')}"
+          f" cleanYr={mg.get('clean_year_uplift')}")
+    print(f"  blend w25 uplift_hold={w25.get('uplift_hold')} uplift_full={w25.get('uplift_full')}"
+          f" | w50 full={w50.get('full')} hold={w50.get('hold')} dd={w50.get('dd')}")
+    print(f"  earns_slot={b['earns']}  beats_winner={b['beats']}")
+    print(f"  BTC_DD={r['per_asset']['BTC']['full']['maxdd']} ETH_DD={r['per_asset']['ETH']['full']['maxdd']}")

scripts/research/skyhook/runs/SKH2_CHANDE_WIN.py

@@ -0,0 +1,225 @@
+"""SKH2_CHANDE_WIN — DD-reduction wave: re-tune indicator WINDOWS (Chande/ATR) for DD.
+
+Family task: smoother indicators -> more stable regime -> potentially lower standalone maxDD.
+We hold the VERIFIED V2 winner's pattern/exits/bands FIXED and sweep ONLY the windows:
+    n_vola, n_volume in {7,13,21,34}
+    atr_win           in {10,14,21}
+    ltf_atr_win       in {10,14,21}
+
+Everything is expressible via SkyhookParams -> the SHARED honest harness sk.study() applies
+the exact leak-free FULL+HOLDOUT+fee-sweep+per-year machinery, and sk.causality / sk.marginal
+give the same comparable numbers as every other agent.
+
+WINNER (baseline to beat):
+    SkyhookParams(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16,
+                  vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+    minFull +0.83 minHold +0.81 ; standalone DD BTC 34% / ETH 31% (>30% = the problem).
+
+GOAL: max_dd < 0.30 while keeping minHold >= ~0.70 and earns_slot True, blend w25 uplift_hold >= 0.55.
+"""
+from __future__ import annotations
+
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+
+import itertools
+
+import skyhooklib as sk
+from src.strategies.skyhook import SkyhookParams
+
+# ---- fixed winner spine (pattern / exits / bands) --------------------------
+FIXED = dict(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16,
+             vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+
+
+def mk(n_vola, n_volume, atr_win, ltf_atr_win):
+    return SkyhookParams(n_vola=n_vola, n_volume=n_volume, atr_win=atr_win,
+                         ltf_atr_win=ltf_atr_win, **FIXED)
+
+
+def cheap_eval(p):
+    """Fast standalone screen: FULL+HOLD on BTC&ETH only (no fee-sweep/marginal)."""
+    rb = sk.run_asset("BTC", p)
+    re = sk.run_asset("ETH", p)
+    min_full = min(rb["full"]["sharpe"], re["full"]["sharpe"])
+    min_hold = min(rb["holdout"]["sharpe"], re["holdout"]["sharpe"])
+    max_dd = max(rb["full"]["maxdd"], re["full"]["maxdd"])
+    min_tr = min(rb["full"]["n_trades"], re["full"]["n_trades"])
+    return dict(min_full=min_full, min_hold=min_hold, max_dd=max_dd, min_tr=min_tr,
+                btc_dd=rb["full"]["maxdd"], eth_dd=re["full"]["maxdd"],
+                btc_hold=rb["holdout"]["sharpe"], eth_hold=re["holdout"]["sharpe"])
+
+
+def earns_slot(rep, marg):
+    return (rep["verdict"]["grade"] != "FAIL"
+            and marg.get("marginal_verdict") == "ADDS"
+            and bool(marg.get("robust_oos"))
+            and not bool(marg.get("is_hedge")))
+
+
+def beats_winner(rep, marg, ev):
+    es = earns_slot(rep, marg)
+    w25 = marg.get("blends", {}).get("w25", {}).get("uplift_hold")
+    mh = rep["verdict"]["min_asset_holdout_sharpe"]
+    return bool(es and ev["max_dd"] < 0.30 and (w25 is not None and w25 >= 0.55) and mh >= 0.65)
+
+
+# ---- WINNER reference (so DD comparison is apples-to-apples in THIS harness) ----
+def winner_params():
+    return SkyhookParams(**FIXED)
+
+
+if __name__ == "__main__":
+    print("########## STAGE 1: cheap window screen (FULL+HOLD+DD, BTC&ETH) ##########")
+
+    # winner reference in this harness
+    wev = cheap_eval(winner_params())
+    print(f"[WINNER ref] minFull={wev['min_full']:+.2f} minHold={wev['min_hold']:+.2f} "
+          f"maxDD={wev['max_dd']*100:.0f}% (BTC {wev['btc_dd']*100:.0f}% / ETH {wev['eth_dd']*100:.0f}%) "
+          f"minTr={wev['min_tr']}")
+
+    n_vola_grid = [7, 13, 21, 34]
+    n_volume_grid = [7, 13, 21, 34]
+    atr_grid = [10, 14, 21]
+    ltf_grid = [10, 14, 21]
+
+    rows = []
+    for nva, nvo, aw, law in itertools.product(n_vola_grid, n_volume_grid, atr_grid, ltf_grid):
+        p = mk(nva, nvo, aw, law)
+        ev = cheap_eval(p)
+        rows.append((nva, nvo, aw, law, ev))
+
+    # Sort by lowest DD among those that keep some hold-out edge & enough trades
+    def keyf(r):
+        ev = r[4]
+        return ev["max_dd"]
+
+    viable = [r for r in rows if r[4]["min_tr"] >= 20]
+    viable.sort(key=keyf)
+
+    print(f"\n--- {len(rows)} configs screened. Top 15 by LOWEST standalone maxDD "
+          f"(min_tr>=20) ---")
+    print(f"{'nva':>4}{'nvo':>4}{'aw':>4}{'law':>5}  {'maxDD':>7} {'btcDD':>6} {'ethDD':>6} "
+          f"{'minFull':>8} {'minHold':>8} {'minTr':>6}")
+    for nva, nvo, aw, law, ev in viable[:15]:
+        print(f"{nva:>4}{nvo:>4}{aw:>4}{law:>5}  {ev['max_dd']*100:>6.1f}% "
+              f"{ev['btc_dd']*100:>5.1f}% {ev['eth_dd']*100:>5.1f}% "
+              f"{ev['min_full']:>+8.2f} {ev['min_hold']:>+8.2f} {ev['min_tr']:>6}")
+
+    # STAGE-1 LEARNING (from broad probes): no window combo gets BOTH BTC&ETH sub-30%
+    # (BTC & ETH DD move in OPPOSITE directions vs n_vola/atr_win). The best DD-CUT that
+    # also keeps hold-out is the SLOWER-INDICATOR corner. Study the lowest-DD configs that
+    # still keep minHold>=0.50 (the real DD/hold tradeoff frontier), plus a couple extras
+    # found by the broad probe (n_vola=13, slower atr_win/ltf_atr_win).
+    extra = [(13, 13, 18, 18), (13, 13, 14, 18), (13, 13, 21, 18)]  # (nva,nvo,aw,law)
+    # candidates that cut DD while keeping hold-out
+    cands = [r for r in viable
+             if r[4]["min_hold"] >= 0.50 and r[4]["min_full"] >= 0.50]
+    cands.sort(key=lambda r: r[4]["max_dd"])  # lowest DD first
+    study_keys = set()
+    study_list = []
+    for r in cands[:5]:
+        k = (r[0], r[1], r[2], r[3])
+        if k not in study_keys:
+            study_keys.add(k); study_list.append(r)
+    # ensure the broad-probe extras are studied (they may not be on the coarse grid)
+    for nva, nvo, aw, law in extra:
+        k = (nva, nvo, aw, law)
+        if k not in study_keys:
+            ev = cheap_eval(mk(nva, nvo, aw, law))
+            study_keys.add(k); study_list.append((nva, nvo, aw, law, ev))
+    if not study_list:
+        study_list = viable[:4]
+
+    print(f"\n########## STAGE 2: FULL study + causality + marginal on "
+          f"{len(study_list)} candidate(s) ##########")
+
+    results = []
+    for nva, nvo, aw, law, ev in study_list:
+        p = mk(nva, nvo, aw, law)
+        name = f"CW_nva{nva}_nvo{nvo}_aw{aw}_law{law}"
+        rep = sk.study(name, p)
+        caus_b = sk.causality(p, "BTC")
+        caus_e = sk.causality(p, "ETH")
+        marg = sk.marginal(p)
+        caus_ok = bool(caus_b["ok"] and caus_e["ok"])
+        es = earns_slot(rep, marg)
+        bw = beats_winner(rep, marg, ev) and caus_ok
+        w25 = marg.get("blends", {}).get("w25", {}).get("uplift_hold")
+        w50 = marg.get("blends", {}).get("w50", {})
+        results.append(dict(name=name, p=p, rep=rep, marg=marg, ev=ev,
+                            caus_ok=caus_ok, es=es, bw=bw, w25=w25, w50=w50,
+                            cfg=dict(n_vola=nva, n_volume=nvo, atr_win=aw, ltf_atr_win=law)))
+        print("\n" + sk.fmt(rep))
+        print(f"  causality BTC={caus_b} ETH={caus_e}")
+        print(f"  marginal: verdict={marg.get('marginal_verdict')} corr_full={marg.get('corr_full')} "
+              f"has_insample_edge={marg.get('has_insample_edge')} is_hedge={marg.get('is_hedge')} "
+              f"robust_oos={marg.get('robust_oos')} multicut_persistent={marg.get('multicut_persistent')}")
+        print(f"            clean_year_uplift={marg.get('clean_year_uplift')} "
+              f"blend_w25_uplift_hold={w25} w50={w50}")
+        print(f"  >> earns_slot={es}  beats_winner={bw}  standaloneDD={ev['max_dd']*100:.1f}%")
+
+    # ---- pick best: prefer beats_winner; else lowest DD among earns_slot; else lowest DD ----
+    def rank(r):
+        # higher is better. Priority: (1) beats_winner, (2) earns_slot, then PORTFOLIO VALUE
+        # (blend w25 uplift_hold + min-asset hold-out) which the wave objectives (2)&(3) reward,
+        # with DD as a final tiebreak. NOTE: no window combo reaches max_dd<0.30 (DD wall is
+        # structural: BTC & ETH DD move in OPPOSITE directions vs the vola window) so we report
+        # the strongest earns_slot config rather than chasing an unreachable DD gate.
+        w25 = r["w25"] if r["w25"] is not None else -9
+        return (1 if r["bw"] else 0,
+                1 if r["es"] else 0,
+                round(w25, 3),
+                r["rep"]["verdict"]["min_asset_holdout_sharpe"],
+                -r["ev"]["max_dd"])
+
+    if results:
+        best = sorted(results, key=rank, reverse=True)[0]
+        b = best
+        rep = b["rep"]; marg = b["marg"]; ev = b["ev"]
+        v = rep["verdict"]
+        print("\n" + "=" * 78)
+        print("FINAL BEST CONFIG (CHANDE_WIN family)")
+        print("=" * 78)
+        print(f"  config            = {b['cfg']}  (+ fixed winner spine {FIXED})")
+        print(f"  name              = {b['name']}")
+        print(f"  minFull           = {v['min_asset_full_sharpe']:+.3f}")
+        print(f"  minHold           = {v['min_asset_holdout_sharpe']:+.3f}  "
+              f"(BTC {rep['per_asset']['BTC']['holdout']['sharpe']:+.2f} / "
+              f"ETH {rep['per_asset']['ETH']['holdout']['sharpe']:+.2f})")
+        print(f"  standalone max_dd = {ev['max_dd']:.4f}  "
+              f"(BTC {ev['btc_dd']:.4f} / ETH {ev['eth_dd']:.4f})")
+        print(f"  n_trades_min      = {v['min_trades']}")
+        print(f"  fee_survives 0.30%= {v['fee_survives']}")
+        print(f"  causality_ok      = {b['caus_ok']}")
+        print(f"  grade             = {v['grade']}")
+        print(f"  --- marginal vs TP01 ---")
+        print(f"  corr_full         = {marg.get('corr_full')}")
+        print(f"  marginal_verdict  = {marg.get('marginal_verdict')}")
+        print(f"  has_insample_edge = {marg.get('has_insample_edge')}")
+        print(f"  is_hedge          = {marg.get('is_hedge')}")
+        print(f"  robust_oos        = {marg.get('robust_oos')}")
+        print(f"  multicut_persist  = {marg.get('multicut_persistent')}")
+        print(f"  clean_year_uplift = {marg.get('clean_year_uplift')}")
+        print(f"  blend w25 uplift_hold = {b['w25']}")
+        print(f"  blend w50         = {b['w50']}")
+        print(f"  earns_slot        = {b['es']}")
+        print(f"  BEATS_WINNER      = {b['bw']}")
+        print("=" * 78)
+
+        # machine-readable line for the harness operator
+        import json
+        out = dict(
+            family="CHANDE_WIN", best_config=b["cfg"], fixed=FIXED, name=b["name"],
+            grade=v["grade"], min_full=v["min_asset_full_sharpe"],
+            min_hold=v["min_asset_holdout_sharpe"], max_dd=ev["max_dd"],
+            btc_dd=ev["btc_dd"], eth_dd=ev["eth_dd"], n_trades_min=v["min_trades"],
+            fee_survives=v["fee_survives"], causality_ok=b["caus_ok"],
+            corr_full=marg.get("corr_full"), marginal_verdict=marg.get("marginal_verdict"),
+            has_insample_edge=marg.get("has_insample_edge"), is_hedge=marg.get("is_hedge"),
+            robust_oos=marg.get("robust_oos"),
+            multicut_persistent=marg.get("multicut_persistent"),
+            clean_year_uplift=marg.get("clean_year_uplift"),
+            blend_w25_uplift_hold=b["w25"], earns_slot=b["es"], beats_winner=b["bw"],
+        )
+        print("RESULT_JSON " + json.dumps(out, default=str))

scripts/research/skyhook/runs/SKH2_DDKILL.py

@@ -0,0 +1,381 @@
+"""SKH2_DDKILL — CAUSAL drawdown kill-switch overlay on Skyhook entries.
+
+Family: drawdown kill-switch on entries [DDKILL].
+
+Idea
+----
+Walk the trade-by-trade REALIZED equity of the V2-winner Skyhook. Track the running peak.
+Once standalone DD from the running peak exceeds `dd_kill`, enter a "killed" state and SUPPRESS
+new entries until equity recovers within `recover` of the running peak (i.e. DD shrinks back
+below `recover`). This is sequential & causal: the kill decision for a new entry at bar i uses
+ONLY the equity realized by trades that closed at/before i (busy_until <= i).
+
+Implementation
+--------------
+1. Build base entries with the winner SkyhookParams.
+2. Run backtest_signals -> realized equity path (mark-at-trade-exit, forward-filled).
+3. From that equity path compute a per-bar boolean `killed[i]` (causal: peak/DD use eq up to i,
+   and eq[i] only changes at a trade-exit bar -> the state at the moment we'd open a new entry
+   at i reflects only past closed trades).
+4. Null entries where killed -> re-run. Equity changes (suppressed losers/winners during DD),
+   so ITERATE to a fixed point (state stabilizes, usually 2-5 iters).
+5. Evaluate FULL + HOLD-OUT + fee-sweep + per-year on BOTH assets, marginal-vs-TP01,
+   combined-curve max-DD, causality (truncated-prefix recompute of the FINAL entries).
+
+CAUSALITY of the overlay itself
+-------------------------------
+The base skyhook_entries are already causal (sk.causality). The kill mask at bar i is a function
+of equity[0..i], and equity[j] for j<=i only embeds trades whose exit_idx <= i. The mask never
+references a future bar. We additionally PROVE it by a truncated-prefix recompute: re-deriving the
+final (killed) entries on a data prefix must match the full-run final entries on the overlap.
+"""
+from __future__ import annotations
+
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+
+import numpy as np
+import pandas as pd
+
+import skyhooklib as sk
+from src.backtest.harness import backtest_signals
+from src.strategies.skyhook import SkyhookParams, skyhook_entries
+
+import altlib as al
+
+HOLDOUT = sk.HOLDOUT
+FEE = sk.FEE_RT
+FEE_SWEEP = (0.0, 0.001, 0.002, 0.003)
+
+# The verified V2 winner from the prior wave.
+WINNER = dict(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16,
+              vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+
+
+def winner_params() -> SkyhookParams:
+    return SkyhookParams(**WINNER)
+
+
+# ---------------------------------------------------------------------------
+# Causal DD kill-switch overlay
+# ---------------------------------------------------------------------------
+def killed_mask_from_equity(equity: np.ndarray, dd_kill: float, recover: float) -> np.ndarray:
+    """Per-bar boolean: True = entries SUPPRESSED at this bar.
+
+    State machine over the realized equity path (hysteresis):
+      - track running peak.
+      - if not killed and DD_from_peak > dd_kill  -> killed.
+      - if killed and DD_from_peak <= recover     -> un-killed (recovered).
+    Causal: peak[i] and dd[i] use equity[0..i] only.
+    """
+    n = len(equity)
+    killed = np.zeros(n, dtype=bool)
+    peak = equity[0]
+    state = False
+    for i in range(n):
+        if equity[i] > peak:
+            peak = equity[i]
+        dd = (peak - equity[i]) / peak if peak > 0 else 0.0
+        if state:
+            if dd <= recover:
+                state = False
+        else:
+            if dd > dd_kill:
+                state = True
+        killed[i] = state
+    return killed
+
+
+def apply_kill(entries: list, killed: np.ndarray) -> list:
+    out = list(entries)
+    for i in range(len(out)):
+        if i < len(killed) and killed[i]:
+            out[i] = None
+    return out
+
+
+def ddkill_entries_for_asset(asset: str, p: SkyhookParams, dd_kill: float, recover: float,
+                             fee_rt: float = FEE, max_iter: int = 8):
+    """Iterate the kill-switch to a fixed point. Returns (final_entries, ltf, n_iters)."""
+    ltf, htf = sk.frames(asset)
+    base = skyhook_entries(ltf, htf, p)
+    cur = base
+    prev_killcount = -1
+    iters = 0
+    for it in range(max_iter):
+        m = backtest_signals(ltf, cur, fee_rt=fee_rt, leverage=1.0, asset=asset, tf="230m")
+        killed = killed_mask_from_equity(m.equity, dd_kill, recover)
+        nxt = apply_kill(base, killed)   # always re-derive from BASE so a recovered state re-enables entries
+        iters = it + 1
+        kc = int(killed.sum())
+        # fixed point: same set of nulled entries as before
+        same = all((a is None) == (b is None) for a, b in zip(nxt, cur))
+        cur = nxt
+        if same and kc == prev_killcount:
+            break
+        prev_killcount = kc
+    return cur, ltf, iters
+
+
+def _split(eq: np.ndarray, idx: pd.DatetimeIndex, mask: np.ndarray) -> dict:
+    e = eq[mask]
+    if len(e) < 5:
+        return dict(sharpe=0.0, ret=0.0, maxdd=0.0, n=int(len(e)))
+    r = np.diff(e) / e[:-1]
+    r = r[np.isfinite(r)]
+    ix = idx[mask]
+    dt = pd.Series(ix).diff().dt.total_seconds().median() or 86400
+    bpy = 86400 * 365.25 / dt
+    sharpe = float(np.mean(r) / np.std(r) * np.sqrt(bpy)) if len(r) and np.std(r) > 0 else 0.0
+    pk = np.maximum.accumulate(e)
+    dd = float(np.max((pk - e) / pk))
+    return dict(sharpe=round(sharpe, 3), ret=round(float(e[-1] / e[0] - 1), 4),
+                maxdd=round(dd, 4), n=int(len(e)))
+
+
+def study_ddkill(name: str, p: SkyhookParams, dd_kill: float, recover: float):
+    per_asset = {}
+    fee_ok_all = True
+    entries_by_asset = {}
+    for a in ("BTC", "ETH"):
+        ent, ltf, iters = ddkill_entries_for_asset(a, p, dd_kill, recover)
+        entries_by_asset[a] = (ent, ltf)
+        m = backtest_signals(ltf, ent, fee_rt=FEE, leverage=1.0, asset=a, tf="230m")
+        eq = m.equity
+        idx = pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True))
+        hmask = np.asarray(idx >= HOLDOUT)
+        full = dict(sharpe=round(m.sharpe, 3), ret=round(m.net_return, 4), maxdd=round(m.max_dd, 4),
+                    n_trades=int(m.n_trades), win_rate=round(m.win_rate, 1))
+        hold = _split(eq, idx, hmask)
+        sweep = {}
+        for f in FEE_SWEEP:
+            mf = backtest_signals(ltf, ent, fee_rt=f, leverage=1.0, asset=a, tf="230m")
+            sweep[f"{f*100:.2f}%"] = round(mf.sharpe, 3)
+        fee_ok_all = fee_ok_all and (sweep["0.30%"] > 0)
+        per_asset[a] = dict(full=full, hold=hold,
+                            yearly={int(y): round(v, 4) for y, v in m.yearly.items()},
+                            fee_sweep=sweep, iters=iters)
+    mf = min(per_asset[a]["full"]["sharpe"] for a in per_asset)
+    mh = min(per_asset[a]["hold"]["sharpe"] for a in per_asset)
+    mt = min(per_asset[a]["full"]["n_trades"] for a in per_asset)
+    mdd = max(per_asset[a]["full"]["maxdd"] for a in per_asset)
+    grade = "PASS" if (mt >= 20 and mf >= 0.5 and mh >= 0.2 and fee_ok_all) else \
+            ("WEAK" if (mt >= 20 and mf >= 0.3 and mh >= 0.0) else "FAIL")
+    print(f"\n=== {name} (dd_kill={dd_kill:.0%} recover={recover:.0%}) -> {grade}  "
+          f"(minFull={mf:+.2f} minHold={mh:+.2f} minTr={mt} maxDD={mdd*100:.0f}% feeOK={fee_ok_all})")
+    for a in per_asset:
+        pa = per_asset[a]
+        yr = " ".join(f"{y}:{r*100:+.0f}%" for y, r in pa["yearly"].items())
+        print(f"  {a}: FULL Sh={pa['full']['sharpe']:+.2f} ret={pa['full']['ret']*100:+.0f}% "
+              f"DD={pa['full']['maxdd']*100:.0f}% n={pa['full']['n_trades']} wr={pa['full']['win_rate']:.0f}% "
+              f"| HOLD Sh={pa['hold']['sharpe']:+.2f} ret={pa['hold']['ret']*100:+.0f}% DD={pa['hold']['maxdd']*100:.0f}% "
+              f"[iters={pa['iters']}]")
+        print(f"      fee: " + " ".join(f"{k}={v:+.2f}" for k, v in pa["fee_sweep"].items()))
+        print(f"      yr: {yr}")
+    return dict(grade=grade, minFull=mf, minHold=mh, minTr=mt, maxDD=mdd, fee_ok=fee_ok_all,
+                per_asset=per_asset, entries_by_asset=entries_by_asset,
+                dd_kill=dd_kill, recover=recover)
+
+
+def marginal_ddkill(p: SkyhookParams, dd_kill: float, recover: float):
+    def daily(a):
+        ent, ltf, _ = ddkill_entries_for_asset(a, p, dd_kill, recover)
+        m = backtest_signals(ltf, ent, fee_rt=FEE, leverage=1.0, asset=a, tf="230m")
+        s = pd.Series(m.equity, index=pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True)))
+        return s.resample("1D").last().ffill().pct_change().dropna()
+    sb, se = daily("BTC"), daily("ETH")
+    J = pd.concat({"BTC": sb, "ETH": se}, axis=1, join="inner").fillna(0.0)
+    cand = 0.5 * J["BTC"] + 0.5 * J["ETH"]
+    return al.marginal_vs_tp01(cand)
+
+
+def combined_curve_maxdd(res: dict) -> float:
+    """Max-DD of the COMBINED 50/50 BTC+ETH bar-level equity (single standalone curve)."""
+    curves = []
+    for a in ("BTC", "ETH"):
+        ent, ltf = res["entries_by_asset"][a]
+        m = backtest_signals(ltf, ent, fee_rt=FEE, leverage=1.0, asset=a, tf="230m")
+        s = pd.Series(m.equity, index=pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True)))
+        curves.append(s.resample("1D").last().ffill().pct_change().fillna(0.0))
+    J = pd.concat({"BTC": curves[0], "ETH": curves[1]}, axis=1, join="inner").fillna(0.0)
+    r = 0.5 * J["BTC"] + 0.5 * J["ETH"]
+    eq = (1.0 + r).cumprod().values
+    pk = np.maximum.accumulate(eq)
+    return float(np.max((pk - eq) / pk))
+
+
+# ---------------------------------------------------------------------------
+# Causality of the FINAL (killed) entries via truncated-prefix recompute.
+# ---------------------------------------------------------------------------
+def causality_ddkill(p: SkyhookParams, dd_kill: float, recover: float, asset: str = "BTC",
+                     tail: int = 200) -> dict:
+    """Re-derive final killed entries on a data PREFIX; they must match the full-run final
+    entries on the overlap tail. Proves the kill mask uses no future bar."""
+    full_ent, ltf_full = (lambda r: r[:2])(ddkill_entries_for_asset(asset, p, dd_kill, recover))
+    n = len(ltf_full)
+    ltf, htf = sk.frames(asset)
+    bad = 0; checked = 0
+    for frac in (0.80, 0.92):
+        cut = int(n * frac)
+        cut_ts = int(ltf["timestamp"].iloc[cut - 1])
+        ltf_sub = ltf.iloc[:cut].reset_index(drop=True)
+        htf_sub = htf[htf["timestamp"] <= cut_ts].reset_index(drop=True)
+        # re-run the whole kill iteration on the prefix
+        base_sub = skyhook_entries(ltf_sub, htf_sub, p)
+        cur = base_sub
+        for _ in range(8):
+            m = backtest_signals(ltf_sub, cur, fee_rt=FEE, leverage=1.0, asset=asset, tf="230m")
+            killed = killed_mask_from_equity(m.equity, dd_kill, recover)
+            nxt = apply_kill(base_sub, killed)
+            same = all((x is None) == (y is None) for x, y in zip(nxt, cur))
+            cur = nxt
+            if same:
+                break
+        for i in range(max(0, cut - tail), cut):
+            checked += 1
+            aE, bE = full_ent[i], cur[i]
+            if (aE is None) != (bE is None):
+                bad += 1
+            elif aE is not None and (aE["dir"] != bE["dir"]
+                                     or abs(aE["sl"] - bE["sl"]) > 1e-6
+                                     or abs(aE["tp"] - bE["tp"]) > 1e-6):
+                bad += 1
+    return dict(ok=bool(bad == 0), mismatches=int(bad), checked=int(checked))
+
+
+def earns_slot_of(res: dict, mg: dict) -> bool:
+    return (res["grade"] != "FAIL"
+            and mg.get("marginal_verdict") == "ADDS"
+            and bool(mg.get("robust_oos"))
+            and not bool(mg.get("is_hedge")))
+
+
+def beats_winner(res: dict, mg: dict, max_dd: float, earns: bool) -> bool:
+    w25 = mg.get("blends", {}).get("w25", {}).get("uplift_hold")
+    return bool(earns and max_dd < 0.30
+                and (w25 is not None and w25 >= 0.55)
+                and res["minHold"] >= 0.65)
+
+
+if __name__ == "__main__":
+    p = winner_params()
+
+    print("########## BASELINE (winner, no kill) for reference ##########")
+    base_rep = sk.study("WINNER (no kill)", p)
+    print(sk.fmt(base_rep))
+    base_dd = max(base_rep["per_asset"][a]["full"]["maxdd"] for a in base_rep["per_asset"])
+    print(f"  winner standalone maxDD (per-asset max) = {base_dd*100:.0f}%")
+
+    # Grid of (dd_kill, recover). recover < dd_kill (hysteresis: re-enable once DD shrinks back).
+    grid = [
+        (0.15, 0.10),
+        (0.15, 0.12),
+        (0.18, 0.12),
+        (0.18, 0.15),
+        (0.20, 0.15),
+        (0.22, 0.16),
+        (0.25, 0.18),
+        (0.30, 0.22),
+    ]
+
+    results = []
+    for dd_kill, recover in grid:
+        res = study_ddkill(f"DDKILL", p, dd_kill, recover)
+        results.append(res)
+
+    print("\n\n########## MARGINAL + combined-DD + earns_slot ##########")
+    summary = []
+    for res in results:
+        mg = marginal_ddkill(p, res["dd_kill"], res["recover"])
+        cdd = combined_curve_maxdd(res)
+        per_asset_dd = res["maxDD"]
+        # standalone max_dd per the brief = max(full.maxdd over BTC & ETH) for the overlay too
+        max_dd = per_asset_dd
+        earns = earns_slot_of(res, mg)
+        bw = beats_winner(res, mg, max_dd, earns)
+        w25 = mg.get("blends", {}).get("w25", {}).get("uplift_hold")
+        w50 = mg.get("blends", {}).get("w50", {})
+        summary.append(dict(dd_kill=res["dd_kill"], recover=res["recover"], grade=res["grade"],
+                            minFull=res["minFull"], minHold=res["minHold"], minTr=res["minTr"],
+                            per_asset_dd=per_asset_dd, combined_dd=cdd, max_dd=max_dd,
+                            corr_full=mg.get("corr_full"), verdict=mg.get("marginal_verdict"),
+                            insample=mg.get("has_insample_edge"), hedge=mg.get("is_hedge"),
+                            robust=mg.get("robust_oos"), multicut=mg.get("multicut_persistent"),
+                            cleanYr=mg.get("clean_year_uplift"), w25=w25, w50=w50,
+                            fee_ok=res["fee_ok"], earns=earns, beats=bw, mg=mg, res=res))
+        print(f"[dd_kill={res['dd_kill']:.0%} recover={res['recover']:.0%}] grade={res['grade']} "
+              f"minFull={res['minFull']:+.2f} minHold={res['minHold']:+.2f} "
+              f"perAssetDD={per_asset_dd*100:.0f}% combinedDD={cdd*100:.0f}% "
+              f"| corr={mg.get('corr_full')} verdict={mg.get('marginal_verdict')} "
+              f"insample={mg.get('has_insample_edge')} hedge={mg.get('is_hedge')} "
+              f"robust={mg.get('robust_oos')} multicut={mg.get('multicut_persistent')} "
+              f"cleanYr={mg.get('clean_year_uplift')} w25uplift={w25} "
+              f"| earns={earns} BEATS={bw}")
+
+    # Pick best honestly: prefer beats_winner; then earns_slot AND a healthy hold-out floor
+    # (>=0.65) so we never pick a DD win that kills the hold-out; then lowest per-asset DD;
+    # then highest minHold.
+    def rank(s):
+        healthy = bool(s["earns"]) and (s["minHold"] or -9) >= 0.65
+        return (not s["beats"], not healthy, not s["earns"], s["max_dd"], -(s["minHold"] or -9))
+    summary.sort(key=rank)
+    best = summary[0]
+    res = best["res"]; mg = best["mg"]
+
+    # Causality of the FINAL killed entries on the best config, both assets.
+    cz_btc = causality_ddkill(p, best["dd_kill"], best["recover"], "BTC")
+    cz_eth = causality_ddkill(p, best["dd_kill"], best["recover"], "ETH")
+    cz_ok = cz_btc["ok"] and cz_eth["ok"]
+
+    print("\n\n################## BEST CONFIG ##################")
+    print(f"config: WINNER + DDKILL(dd_kill={best['dd_kill']:.0%}, recover={best['recover']:.0%})")
+    print(f"  minFull        = {best['minFull']:+.3f}")
+    print(f"  minHold        = {best['minHold']:+.3f}")
+    print(f"  per-asset maxDD= {best['per_asset_dd']*100:.1f}%  (max over BTC&ETH full.maxdd)")
+    print(f"  combined  maxDD= {best['combined_dd']*100:.1f}%  (50/50 daily curve)")
+    print(f"  n_trades_min   = {best['minTr']}")
+    print(f"  fee@0.30%      = {best['fee_ok']}")
+    print(f"  causality      = BTC {cz_btc} | ETH {cz_eth} -> ok={cz_ok}")
+    print(f"  --- MARGINAL vs TP01 ---")
+    print(f"  corr_full           = {mg.get('corr_full')}")
+    print(f"  corr_hold           = {mg.get('corr_hold')}")
+    print(f"  marginal_verdict    = {mg.get('marginal_verdict')}")
+    print(f"  has_insample_edge   = {mg.get('has_insample_edge')}")
+    print(f"  is_hedge            = {mg.get('is_hedge')}")
+    print(f"  robust_oos          = {mg.get('robust_oos')}")
+    print(f"  multicut_persistent = {mg.get('multicut_persistent')}")
+    print(f"  clean_year_uplift   = {mg.get('clean_year_uplift')}")
+    print(f"  jackknife_min_uplift= {mg.get('jackknife_min_uplift')}")
+    print(f"  cand_insample_sharpe= {mg.get('cand_insample_sharpe')}")
+    print(f"  blends.w25          = {mg.get('blends', {}).get('w25')}")
+    print(f"  blends.w50          = {mg.get('blends', {}).get('w50')}")
+    earns = best["earns"]
+    print(f"  earns_slot          = {earns}")
+    print(f"  BEATS_WINNER        = {best['beats']}")
+
+    # Emit a compact machine-readable line for the orchestrator.
+    import json
+    w25 = mg.get("blends", {}).get("w25", {}).get("uplift_hold")
+    out = dict(
+        family="ddkill_entries",
+        best_config=dict(base=WINNER, dd_kill=best["dd_kill"], recover=best["recover"]),
+        ran_ok=True, grade=res["grade"],
+        min_full_sharpe=round(float(best["minFull"]), 3),
+        min_hold_sharpe=round(float(best["minHold"]), 3),
+        max_dd=round(float(best["max_dd"]), 4),
+        combined_dd=round(float(best["combined_dd"]), 4),
+        n_trades_min=int(best["minTr"]),
+        fee_survives_030=bool(best["fee_ok"]),
+        causality_ok=bool(cz_ok),
+        marginal_verdict=mg.get("marginal_verdict"),
+        has_insample_edge=bool(mg.get("has_insample_edge")),
+        is_hedge=bool(mg.get("is_hedge")),
+        robust_oos=bool(mg.get("robust_oos")),
+        multicut_persistent=bool(mg.get("multicut_persistent")),
+        clean_year_uplift=mg.get("clean_year_uplift"),
+        corr_full=mg.get("corr_full"),
+        blend_w25_uplift_hold=w25,
+        earns_slot=bool(earns),
+        beats_winner=bool(best["beats"]),
+    )
+    print("\nRESULT_JSON " + json.dumps(out, default=str))

scripts/research/skyhook/runs/SKH2_DUALTF_PTN.py

@@ -0,0 +1,399 @@
+"""SKH2_DUALTF_PTN — LTF (230m) CONFIRMATION of the HTF (690m) breakout at entry.
+
+FAMILY: DUALTF_PTN. Hypothesis (DD-cut): the V2 winner enters on a fresh HTF Donchian
+breakout regardless of where the LTF exec-frame is. If we ALSO require the LTF to confirm
+the breakout at the entry bar (LTF close[i] above its own EMA(n) for longs / below for
+shorts, or LTF short-term momentum agrees), we avoid entering against a freshly-turned LTF.
+Fewer "fight the exec-frame" fills -> fewer immediate stop-outs -> lower standalone maxDD,
+ideally without gutting the hold-out edge.
+
+BASELINE (V2 winner): SkyhookParams(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24,
+uscitashort=16, vola_lo=35.0, vola_hi=95.0, vol_lo=0.0).
+  minFull +0.83, minHold +0.81, maxDD BTC 34% / ETH 31% (THE PROBLEM), marginal ADDS.
+
+WHAT THIS SCRIPT DOES (all leak-free):
+  * Reuse S.htf_features (V2 composer, Chande regime) -> comp_long/comp_short on HTF close.
+  * merge backward to LTF (S.merge_htf_to_ltf) -> causal HTF signal at each LTF bar.
+  * Compute LTF confirmation features at close[i]: EMA(n) of LTF close, and an LTF
+    momentum (close[i] vs close[i-mom]). All strictly causal (no shift into the future).
+  * AND the HTF composer with the LTF confirmation: long only if comp_long & ltf_up;
+    short only if comp_short & ltf_dn. (ltf_up/ltf_dn defined by chosen confirm mode.)
+  * Same entry/exit machinery as V2 (sl/tp ATR multiples, asymmetric max_bars, 1/day).
+
+CAUSALITY: every LTF feature uses ltf data with index <= i. EMA via ewm(adjust=False) is a
+pure causal recursion; momentum uses close[i] and close[i-mom]. We prove it with a
+truncated-prefix recompute (same protocol as sk.causality) on our custom entries.
+"""
+from __future__ import annotations
+
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+
+import numpy as np
+import pandas as pd
+
+import skyhooklib as sk
+from src.backtest.harness import backtest_signals
+from src.strategies import skyhook as S
+from src.strategies.skyhook import SkyhookParams
+
+HOLDOUT = sk.HOLDOUT
+FEE = sk.FEE_RT
+
+# V2 winner params (the baseline to beat). LTF confirmation rides on top of these.
+WINNER = dict(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16,
+              vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+
+
+def winner_params(**kw):
+    base = dict(WINNER)
+    base.update(kw)
+    return SkyhookParams(**base)
+
+
+# ---------------------------------------------------------------------------
+# Causal LTF confirmation features (computed on the 230m exec frame, at close[i]).
+# ---------------------------------------------------------------------------
+def ltf_confirm(ltf_close: np.ndarray, *, ema_n: int, mom_n: int) -> tuple[np.ndarray, np.ndarray]:
+    """Return (ltf_up, ltf_dn) boolean masks per LTF bar, strictly causal.
+    ltf_up := close[i] > EMA_n(close)[i]  AND  close[i] > close[i-mom_n]  (momentum agrees)
+    ltf_dn := close[i] < EMA_n(close)[i]  AND  close[i] < close[i-mom_n]
+    EMA via ewm(adjust=False): a causal recursion (uses only data <= i)."""
+    c = pd.Series(np.asarray(ltf_close, float))
+    ema = c.ewm(span=ema_n, adjust=False).mean().values
+    cc = c.values
+    mom_up = np.zeros(len(cc), dtype=bool)
+    mom_dn = np.zeros(len(cc), dtype=bool)
+    if mom_n > 0:
+        mom_up[mom_n:] = cc[mom_n:] > cc[:-mom_n]
+        mom_dn[mom_n:] = cc[mom_n:] < cc[:-mom_n]
+    else:
+        mom_up[:] = True
+        mom_dn[:] = True
+    up = (cc > ema) & mom_up
+    dn = (cc < ema) & mom_dn
+    return up, dn
+
+
+def ltf_confirm_modes(ltf_close: np.ndarray, *, ema_n: int, mom_n: int, mode: str,
+                      slope_n: int = 0):
+    """Causal LTF confirmation masks (ltf_up, ltf_dn). All features use data <= i.
+    Components:
+      ema_up  := close[i] > EMA_n(close)[i]
+      mom_up  := close[i] > close[i-mom_n]   (sustained move over mom_n bars)
+      slope_up:= EMA_n(close)[i] > EMA_n(close)[i-slope_n]  (LTF trend is rising)  if slope_n>0
+    Modes:
+      'ema'   -> ema_up
+      'mom'   -> mom_up
+      'both'  -> ema_up & mom_up
+      'or'    -> ema_up | mom_up
+      'slope' -> slope_up only (EMA itself rising/falling)
+      'ema_slope' -> ema_up & slope_up  (above a rising EMA = real LTF uptrend, strict)
+      'all'   -> ema_up & mom_up & slope_up (strictest)
+    """
+    c = pd.Series(np.asarray(ltf_close, float))
+    ema = c.ewm(span=ema_n, adjust=False).mean().values
+    cc = c.values
+    n = len(cc)
+    ema_up = cc > ema
+    ema_dn = cc < ema
+    mom_up = np.zeros(n, dtype=bool)
+    mom_dn = np.zeros(n, dtype=bool)
+    if mom_n > 0:
+        mom_up[mom_n:] = cc[mom_n:] > cc[:-mom_n]
+        mom_dn[mom_n:] = cc[mom_n:] < cc[:-mom_n]
+    else:
+        mom_up[:] = True
+        mom_dn[:] = True
+    slope_up = np.zeros(n, dtype=bool)
+    slope_dn = np.zeros(n, dtype=bool)
+    if slope_n > 0:
+        slope_up[slope_n:] = ema[slope_n:] > ema[:-slope_n]
+        slope_dn[slope_n:] = ema[slope_n:] < ema[:-slope_n]
+    else:
+        slope_up[:] = True
+        slope_dn[:] = True
+    if mode == "ema":
+        return ema_up, ema_dn
+    if mode == "mom":
+        return mom_up, mom_dn
+    if mode == "or":
+        return (ema_up | mom_up), (ema_dn | mom_dn)
+    if mode == "slope":
+        return slope_up, slope_dn
+    if mode == "ema_slope":
+        return (ema_up & slope_up), (ema_dn & slope_dn)
+    if mode == "all":
+        return (ema_up & mom_up & slope_up), (ema_dn & mom_dn & slope_dn)
+    # default 'both'
+    return (ema_up & mom_up), (ema_dn & mom_dn)
+
+
+def ltf_not_overextended(ltf_close: np.ndarray, ltf_atr: np.ndarray, *,
+                         ema_n: int, max_ext_atr: float):
+    """REJECT (return False) when the LTF is already overextended from its EMA at entry:
+    a long-breakout fired when close[i] is already > ema + max_ext_atr*ATR_LTF[i] is a LATE
+    fill (mean-reversion-prone, big-stop risk). Confirmation = NOT overextended.
+    ltf_up := (close - ema) <= max_ext_atr*ATR   (still room to run, not blown off)
+    ltf_dn := (ema - close) <= max_ext_atr*ATR
+    All causal: ema, ATR, close all at i."""
+    c = pd.Series(np.asarray(ltf_close, float))
+    ema = c.ewm(span=ema_n, adjust=False).mean().values
+    cc = c.values
+    a = np.asarray(ltf_atr, float)
+    a = np.where(np.isfinite(a) & (a > 0), a, np.nan)
+    ext = (cc - ema) / a
+    # long: not too far ABOVE ema ; short: not too far BELOW ema
+    up = np.where(np.isfinite(ext), ext <= max_ext_atr, False)
+    dn = np.where(np.isfinite(ext), (-ext) <= max_ext_atr, False)
+    return up, dn
+
+
+# ---------------------------------------------------------------------------
+# Custom entries: V2 HTF composer AND LTF confirmation.
+# ---------------------------------------------------------------------------
+def dualtf_entries(ltf: pd.DataFrame, htf: pd.DataFrame, p: SkyhookParams,
+                   *, ema_n: int, mom_n: int, mode: str, slope_n: int = 0,
+                   max_ext_atr: float = 0.0) -> list:
+    feat = S.htf_features(htf, p)              # V2 composer (Chande regime + Donchian)
+    m = S.merge_htf_to_ltf(ltf, feat)          # causal backward merge
+
+    c = m["close"].values.astype(float)
+    a = S.atr(m, p.ltf_atr_win)
+    comp_long = np.nan_to_num(m["comp_long"].values).astype(bool)
+    comp_short = np.nan_to_num(m["comp_short"].values).astype(bool)
+
+    if mode == "notext":
+        ltf_up, ltf_dn = ltf_not_overextended(c, a, ema_n=ema_n, max_ext_atr=max_ext_atr)
+    else:
+        ltf_up, ltf_dn = ltf_confirm_modes(c, ema_n=ema_n, mom_n=mom_n, mode=mode, slope_n=slope_n)
+    comp_long = comp_long & ltf_up
+    comp_short = comp_short & ltf_dn
+
+    days = pd.to_datetime(m["datetime"]).dt.floor("D").values
+    entries: list = [None] * len(m)
+    count_today: dict = {}
+    for i in range(len(m)):
+        if not np.isfinite(a[i]) or a[i] <= 0:
+            continue
+        day = days[i]
+        if count_today.get(day, 0) >= p.max_per_day:
+            continue
+        if comp_long[i]:
+            direction, mb = 1, p.uscitalong
+        elif comp_short[i]:
+            direction, mb = -1, p.uscitashort
+        else:
+            continue
+        if p.exit_mode == "atr":
+            sl_off, tp_off = p.sl_atr * a[i], p.tp_atr * a[i]
+        else:
+            sl_off, tp_off = p.sl_pct * c[i], p.tp_pct * c[i]
+        if direction == 1:
+            sl, tp = c[i] - sl_off, c[i] + tp_off
+        else:
+            sl, tp = c[i] + sl_off, c[i] - tp_off
+        entries[i] = {"dir": direction, "tp": float(tp), "sl": float(sl), "max_bars": int(mb)}
+        count_today[day] = count_today.get(day, 0) + 1
+    return entries
+
+
+# ---------------------------------------------------------------------------
+# Eval helpers (FULL + HOLD-OUT + fee sweep + per-year, both assets) — mirrors sk.study.
+# ---------------------------------------------------------------------------
+def _split(eq, idx, mask):
+    e = eq[mask]
+    if len(e) < 5:
+        return dict(sharpe=0.0, ret=0.0, maxdd=0.0, n=int(len(e)))
+    r = np.diff(e) / e[:-1]; r = r[np.isfinite(r)]
+    ix = idx[mask]
+    dt = pd.Series(ix).diff().dt.total_seconds().median() or 86400
+    bpy = 86400 * 365.25 / dt
+    sh = float(np.mean(r) / np.std(r) * np.sqrt(bpy)) if len(r) and np.std(r) > 0 else 0.0
+    pk = np.maximum.accumulate(e); dd = float(np.max((pk - e) / pk))
+    return dict(sharpe=round(sh, 3), ret=round(float(e[-1] / e[0] - 1), 4), maxdd=round(dd, 4), n=int(len(e)))
+
+
+def study_dualtf(name, p, confirm):
+    per_asset = {}
+    fee_ok_all = True
+    for a in ("BTC", "ETH"):
+        ltf, htf = sk.frames(a)
+        ent = dualtf_entries(ltf, htf, p, **confirm)
+        m = backtest_signals(ltf, ent, fee_rt=FEE, leverage=1.0, asset=a, tf="230m")
+        eq = m.equity
+        idx = pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True))
+        hmask = np.asarray(idx >= HOLDOUT)
+        full = dict(sharpe=round(m.sharpe, 3), ret=round(m.net_return, 4), maxdd=round(m.max_dd, 4),
+                    n_trades=int(m.n_trades), win_rate=round(m.win_rate, 1))
+        hold = _split(eq, idx, hmask)
+        sweep = {}
+        for f in (0.0, 0.001, 0.002, 0.003):
+            mf = backtest_signals(ltf, ent, fee_rt=f, leverage=1.0, asset=a, tf="230m")
+            sweep[f"{f*100:.2f}%"] = round(mf.sharpe, 3)
+        fee_ok_all = fee_ok_all and (sweep["0.30%"] > 0)
+        per_asset[a] = dict(full=full, hold=hold, fee_sweep=sweep,
+                            yearly={int(y): round(v, 4) for y, v in m.yearly.items()})
+    mf = min(per_asset[a]["full"]["sharpe"] for a in per_asset)
+    mh = min(per_asset[a]["hold"]["sharpe"] for a in per_asset)
+    mt = min(per_asset[a]["full"]["n_trades"] for a in per_asset)
+    mdd = max(per_asset[a]["full"]["maxdd"] for a in per_asset)
+    grade = "PASS" if (mt >= 20 and mf >= 0.5 and mh >= 0.2 and fee_ok_all) else \
+            ("WEAK" if (mt >= 20 and mf >= 0.3 and mh >= 0.0) else "FAIL")
+    print(f"\n=== {name} -> {grade}  (minFull={mf:+.2f} minHold={mh:+.2f} minTr={mt} maxDD={mdd*100:.0f}% feeOK={fee_ok_all})")
+    for a in per_asset:
+        pa = per_asset[a]
+        yr = " ".join(f"{y}:{r*100:+.0f}%" for y, r in pa["yearly"].items())
+        print(f"  {a}: FULL Sh={pa['full']['sharpe']:+.2f} ret={pa['full']['ret']*100:+.0f}% DD={pa['full']['maxdd']*100:.0f}%"
+              f" n={pa['full']['n_trades']} wr={pa['full']['win_rate']:.0f}% | HOLD Sh={pa['hold']['sharpe']:+.2f} ret={pa['hold']['ret']*100:+.0f}% DD={pa['hold']['maxdd']*100:.0f}%")
+        print(f"      fee: " + " ".join(f"{k}={v:+.2f}" for k, v in pa["fee_sweep"].items()))
+        print(f"      yr: {yr}")
+    return dict(grade=grade, minFull=mf, minHold=mh, minTr=mt, maxDD=mdd, fee_ok=fee_ok_all, per_asset=per_asset)
+
+
+def marginal_dualtf(p, confirm):
+    import altlib as al
+    def daily(a):
+        ltf, htf = sk.frames(a)
+        ent = dualtf_entries(ltf, htf, p, **confirm)
+        m = backtest_signals(ltf, ent, fee_rt=FEE, leverage=1.0, asset=a, tf="230m")
+        s = pd.Series(m.equity, index=pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True)))
+        return s.resample("1D").last().ffill().pct_change().dropna()
+    sb, se = daily("BTC"), daily("ETH")
+    J = pd.concat({"BTC": sb, "ETH": se}, axis=1, join="inner").fillna(0.0)
+    cand = 0.5 * J["BTC"] + 0.5 * J["ETH"]
+    return al.marginal_vs_tp01(cand)
+
+
+def check_causality(p, confirm, asset="BTC", tail=150):
+    ltf, htf = sk.frames(asset)
+    full = dualtf_entries(ltf, htf, p, **confirm)
+    n = len(ltf); bad = 0; checked = 0
+    for frac in (0.80, 0.92):
+        cut = int(n * frac)
+        cut_ts = int(ltf["timestamp"].iloc[cut - 1])
+        htf_cut = htf[htf["timestamp"] <= cut_ts].reset_index(drop=True)
+        sub = dualtf_entries(ltf.iloc[:cut].reset_index(drop=True), htf_cut, p, **confirm)
+        for i in range(max(0, cut - tail), cut):
+            checked += 1
+            a, b = full[i], sub[i]
+            if (a is None) != (b is None):
+                bad += 1
+            elif a is not None and (a["dir"] != b["dir"]
+                                    or abs(a["sl"] - b["sl"]) > 1e-6
+                                    or abs(a["tp"] - b["tp"]) > 1e-6):
+                bad += 1
+    return dict(ok=bool(bad == 0), mismatches=int(bad), checked=int(checked))
+
+
+if __name__ == "__main__":
+    print("########## SKH2_DUALTF_PTN: LTF confirmation of HTF breakout ##########")
+
+    # Reference: V2 winner WITHOUT LTF confirmation (mode 'none' via wide-open masks).
+    p = winner_params()
+
+    # --- Reference (no LTF confirm) using sk.run_asset directly ---
+    print("\n--- V2 WINNER reference (no LTF confirm) ---")
+    refF, refH, refDD, refTr = [], [], [], []
+    for a in ("BTC", "ETH"):
+        r = sk.run_asset(a, p, FEE)
+        refF.append(r["full"]["sharpe"]); refH.append(r["holdout"]["sharpe"])
+        refDD.append(r["full"]["maxdd"]); refTr.append(r["full"]["n_trades"])
+        print(f"  {a}: FULL Sh={r['full']['sharpe']:+.2f} DD={r['full']['maxdd']*100:.0f}% n={r['full']['n_trades']}"
+              f" | HOLD Sh={r['holdout']['sharpe']:+.2f}")
+    print(f"  REF minFull={min(refF):+.2f} minHold={min(refH):+.2f} maxDD={max(refDD)*100:.0f}% minTr={min(refTr)}")
+
+    # --- Sweep LTF confirmation configs ---
+    # ema_n / mom_n on 230m bars. ~6.26 bars/day. EMA 10~1.6d, 20~3.2d. mom small (1-6 bars).
+    # Directional confirms are near-redundant with a fresh breakout (barely filter).
+    # The real DD lever in this family: REJECT OVEREXTENDED LTF fills (late, blow-off,
+    # mean-reversion-prone, big-stop). max_ext_atr = max allowed (close-ema)/ATR_LTF at entry.
+    configs = {
+        # reference directional confirm (keeps ~all trades)
+        "mom_only_m3":      dict(ema_n=20, mom_n=3, mode="mom"),
+        # NOT-OVEREXTENDED gate: tighter max_ext -> fewer late fills -> aim lower DD
+        "notext_e20_x4":    dict(ema_n=20, mom_n=0, mode="notext", max_ext_atr=4.0),
+        "notext_e20_x3":    dict(ema_n=20, mom_n=0, mode="notext", max_ext_atr=3.0),
+        "notext_e20_x2":    dict(ema_n=20, mom_n=0, mode="notext", max_ext_atr=2.0),
+        "notext_e20_x1_5":  dict(ema_n=20, mom_n=0, mode="notext", max_ext_atr=1.5),
+        "notext_e30_x3":    dict(ema_n=30, mom_n=0, mode="notext", max_ext_atr=3.0),
+        "notext_e30_x2":    dict(ema_n=30, mom_n=0, mode="notext", max_ext_atr=2.0),
+        "notext_e10_x2":    dict(ema_n=10, mom_n=0, mode="notext", max_ext_atr=2.0),
+        "notext_e10_x1_5":  dict(ema_n=10, mom_n=0, mode="notext", max_ext_atr=1.5),
+        "notext_e30_x1_5":  dict(ema_n=30, mom_n=0, mode="notext", max_ext_atr=1.5),
+    }
+
+    results = {}
+    for tag, cfg in configs.items():
+        r = study_dualtf(f"DUALTF_{tag}", p, cfg)
+        results[tag] = (cfg, r)
+
+    # --- Pick best: priority (1) DD<30, (2) earns_slot, (3) minHold high ---
+    print("\n\n##### MARGINAL vs TP01 (configs with minTr>=20) #####")
+    scored = []
+    for tag, (cfg, r) in results.items():
+        if r["minTr"] < 20:
+            print(f"[{tag}] minTr={r['minTr']} <20 -> skip marginal")
+            continue
+        mg = marginal_dualtf(p, cfg)
+        verdict = mg.get("marginal_verdict")
+        robust = bool(mg.get("robust_oos"))
+        hedge = bool(mg.get("is_hedge"))
+        earns = (r["grade"] != "FAIL") and (verdict == "ADDS") and robust and (not hedge)
+        w25 = mg.get("blends", {}).get("w25", {}).get("uplift_hold")
+        beats = earns and (r["maxDD"] < 0.30) and (w25 is not None and w25 >= 0.55) and (r["minHold"] >= 0.65)
+        scored.append((tag, cfg, r, mg, earns, beats))
+        print(f"[{tag}] grade={r['grade']} minFull={r['minFull']:+.2f} minHold={r['minHold']:+.2f} maxDD={r['maxDD']*100:.0f}%"
+              f" | corr_full={mg.get('corr_full')} verdict={verdict} insample={mg.get('has_insample_edge')}"
+              f" hedge={hedge} robust={robust} w25uplift={w25} earns_slot={earns} BEATS={beats}")
+
+    if not scored:
+        print("\nNo config with enough trades.")
+        sys.exit(0)
+
+    # Rank: beats_winner first, then DD<30 & earns, then by minHold, then by lowest DD.
+    def rank_key(item):
+        tag, cfg, r, mg, earns, beats = item
+        w25 = mg.get("blends", {}).get("w25", {}).get("uplift_hold") or -9
+        return (beats, (r["maxDD"] < 0.30 and earns), earns, r["minHold"], -r["maxDD"])
+    scored.sort(key=rank_key, reverse=True)
+    best_tag, best_cfg, best_r, best_mg, best_earns, best_beats = scored[0]
+
+    # --- Causality on best ---
+    cb = check_causality(p, best_cfg, "BTC")
+    ce = check_causality(p, best_cfg, "ETH")
+    caus_ok = cb["ok"] and ce["ok"]
+
+    w25 = best_mg.get("blends", {}).get("w25", {}).get("uplift_hold")
+    w50 = best_mg.get("blends", {}).get("w50", {})
+    fee030_min = min(best_r["per_asset"][a]["fee_sweep"]["0.30%"] for a in ("BTC", "ETH"))
+
+    print("\n\n##################### BEST CONFIG #####################")
+    print(f"BEST = DUALTF_{best_tag}  cfg={best_cfg}")
+    print(f"  params = {WINNER}")
+    print(f"  grade={best_r['grade']} minFull={best_r['minFull']:+.2f} minHold={best_r['minHold']:+.2f}"
+          f" maxDD={best_r['maxDD']*100:.1f}% minTr={best_r['minTr']}")
+    print(f"  fee@0.30% (min asset FULL Sh) = {fee030_min:+.3f}  feeOK={best_r['fee_ok']}")
+    print(f"  causality: BTC={cb} ETH={ce} -> OK={caus_ok}")
+    print(f"  marginal: corr_full={best_mg.get('corr_full')} corr_hold={best_mg.get('corr_hold')}"
+          f" verdict={best_mg.get('marginal_verdict')}")
+    print(f"            has_insample_edge={best_mg.get('has_insample_edge')} is_hedge={best_mg.get('is_hedge')}"
+          f" robust_oos={best_mg.get('robust_oos')} multicut_persistent={best_mg.get('multicut_persistent')}")
+    print(f"            clean_year_uplift={best_mg.get('clean_year_uplift')}"
+          f" jackknife_min_uplift={best_mg.get('jackknife_min_uplift')}"
+          f" cand_insample_sharpe={best_mg.get('cand_insample_sharpe')}")
+    print(f"            blend w25 uplift_hold={w25}  w50={w50}")
+    print(f"  earns_slot={best_earns}  BEATS_WINNER={best_beats}")
+
+    # Emit a compact machine-readable line for the harness.
+    import json
+    out = dict(family="DUALTF_PTN", best_tag=best_tag, best_cfg=best_cfg, winner_params=WINNER,
+               grade=best_r["grade"], minFull=best_r["minFull"], minHold=best_r["minHold"],
+               maxDD=best_r["maxDD"], minTr=best_r["minTr"], fee030_min=fee030_min,
+               causality_ok=caus_ok, marginal_verdict=best_mg.get("marginal_verdict"),
+               corr_full=best_mg.get("corr_full"), has_insample_edge=best_mg.get("has_insample_edge"),
+               is_hedge=best_mg.get("is_hedge"), robust_oos=best_mg.get("robust_oos"),
+               multicut_persistent=best_mg.get("multicut_persistent"),
+               clean_year_uplift=best_mg.get("clean_year_uplift"), w25_uplift_hold=w25, w50=w50,
+               earns_slot=best_earns, beats_winner=best_beats)
+    print("\nJSON " + json.dumps(out, default=str))

scripts/research/skyhook/runs/SKH2_ENS_PARAM.py

@@ -0,0 +1,285 @@
+"""SKH2_ENS_PARAM — within-sleeve PARAM ENSEMBLE for Skyhook DD reduction.
+
+Family: equal-weight the DAILY returns of K diverse skyhook param sets (incl. the V2 winner),
+varying ptn_n {25,45,90}, exits, sl/tp. Diversification across configs smooths equity and cuts
+standalone DD without killing hold-out. We:
+  * build each config's per-asset 230m equity (sk.run_asset) -> daily returns,
+  * equal-weight average the configs' daily returns PER ASSET -> ensemble per-asset equity ->
+    standalone DD (max over BTC/ETH) and per-asset/year/full/hold Sharpe via the SAME _split logic,
+  * fee sweep: re-run each config at fee f, average daily, recompute Sharpe (fee_ok = Sharpe>0 @0.30% RT),
+  * causality: every member is a pure SkyhookParams variant -> sk.causality on each (must be ok),
+  * marginal: feed the 50/50 ensemble daily series to altlib.marginal_vs_tp01.
+
+Standalone max_dd for the ensemble = max-DD of the COMBINED (averaged) per-asset equity curve.
+All causal/leak-free: ensemble is a linear combo of leak-free member equities; no future data used.
+"""
+from __future__ import annotations
+
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+
+import numpy as np
+import pandas as pd
+
+import skyhooklib as sk
+import altlib as al
+from src.strategies.skyhook import SkyhookParams
+
+HOLDOUT = sk.HOLDOUT
+FEE = sk.FEE_RT
+CERTIFIED = ("BTC", "ETH")
+ANN = np.sqrt(365.25)
+
+# The verified V2 winner (must be a member of every ensemble).
+WINNER = SkyhookParams(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16,
+                       vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+
+
+def _sharpe(r: np.ndarray) -> float:
+    r = r[np.isfinite(r)]
+    return float(np.mean(r) / np.std(r) * ANN) if len(r) > 2 and np.std(r) > 0 else 0.0
+
+
+def _dd_from_eq(eq: np.ndarray) -> float:
+    pk = np.maximum.accumulate(eq)
+    return float(np.max((pk - eq) / pk)) if len(eq) else 0.0
+
+
+# ---------------------------------------------------------------------------
+# Per-config DAILY equity-return series per asset, cached by (config-id, asset, fee).
+# We use sk.run_asset to get the leak-free 230m equity, then resample to daily LAST and
+# take pct_change -> daily returns. Aligning all members on the same daily index lets us
+# equal-weight-average their daily returns (an equal-capital rebalanced ensemble).
+# ---------------------------------------------------------------------------
+_DAILY_CACHE: dict = {}
+_NTR_CACHE: dict = {}
+
+
+def _config_daily(p: SkyhookParams, asset: str, fee: float) -> pd.Series:
+    key = (id(p), asset, fee)
+    if key in _DAILY_CACHE:
+        return _DAILY_CACHE[key]
+    r = sk.run_asset(asset, p, fee)
+    s = pd.Series(r["_eq"], index=r["_idx"])
+    daily = s.resample("1D").last().ffill().pct_change().dropna()
+    _DAILY_CACHE[key] = daily
+    _NTR_CACHE[(id(p), asset)] = r["full"]["n_trades"]
+    return daily
+
+
+def _ensemble_daily_asset(members, asset: str, fee: float) -> pd.Series:
+    """Equal-weight average of members' daily returns for one asset (common dates)."""
+    cols = {f"m{i}": _config_daily(p, asset, fee) for i, p in enumerate(members)}
+    J = pd.concat(cols, axis=1, join="inner").fillna(0.0)
+    return J.mean(axis=1)
+
+
+def study_ensemble(name: str, members) -> dict:
+    """FULL+HOLD+fee-sweep+per-year on BOTH assets for the equal-weight param ensemble.
+    Standalone DD = max-DD of the averaged per-asset equity curve."""
+    per_asset = {}
+    fee_ok_all = True
+    for a in CERTIFIED:
+        ens = _ensemble_daily_asset(members, a, FEE)
+        eq = np.cumprod(1.0 + ens.values)
+        idx = ens.index
+        full_sh = _sharpe(ens.values)
+        full_dd = _dd_from_eq(eq)
+        full_ret = float(eq[-1] / eq[0] - 1) if len(eq) else 0.0
+        hmask = idx >= HOLDOUT
+        rh = ens.values[hmask]
+        eqh = np.cumprod(1.0 + rh) if rh.size else np.array([1.0])
+        hold_sh = _sharpe(rh)
+        hold_ret = float(eqh[-1] / eqh[0] - 1) if eqh.size else 0.0
+        hold_dd = _dd_from_eq(eqh)
+        # per-year Sharpe-equivalent return
+        yearly = {}
+        for y in sorted(set(idx.year)):
+            ry = ens.values[idx.year == y]
+            eqy = np.cumprod(1.0 + ry) if ry.size else np.array([1.0])
+            yearly[int(y)] = float(eqy[-1] - 1.0) if eqy.size else 0.0
+        # fee sweep
+        sweep = {}
+        for f in (0.0, 0.001, 0.002, 0.003):
+            ensf = _ensemble_daily_asset(members, a, f)
+            sweep[f"{f*100:.2f}%"] = round(_sharpe(ensf.values), 3)
+        fee_ok_all = fee_ok_all and (sweep["0.30%"] > 0)
+        # n_trades = sum across members (the ensemble trades all of them)
+        ntr = sum(_NTR_CACHE.get((id(p), a), 0) for p in members)
+        per_asset[a] = dict(
+            full=dict(sharpe=round(full_sh, 3), ret=round(full_ret, 4), maxdd=round(full_dd, 4),
+                      n_trades=int(ntr)),
+            hold=dict(sharpe=round(hold_sh, 3), ret=round(hold_ret, 4), maxdd=round(hold_dd, 4)),
+            yearly=yearly, fee_sweep=sweep)
+    mf = min(per_asset[a]["full"]["sharpe"] for a in per_asset)
+    mh = min(per_asset[a]["hold"]["sharpe"] for a in per_asset)
+    mt = min(per_asset[a]["full"]["n_trades"] for a in per_asset)
+    mdd = max(per_asset[a]["full"]["maxdd"] for a in per_asset)
+    grade = "PASS" if (mt >= 20 and mf >= 0.5 and mh >= 0.2 and fee_ok_all) else \
+            ("WEAK" if (mt >= 20 and mf >= 0.3 and mh >= 0.0) else "FAIL")
+    print(f"\n=== {name} -> {grade}  (minFull={mf:+.2f} minHold={mh:+.2f} minTr={mt} maxDD={mdd*100:.0f}% feeOK={fee_ok_all})")
+    for a in per_asset:
+        pa = per_asset[a]
+        yr = " ".join(f"{y}:{r*100:+.0f}%" for y, r in pa["yearly"].items())
+        print(f"  {a}: FULL Sh={pa['full']['sharpe']:+.2f} ret={pa['full']['ret']*100:+.0f}% DD={pa['full']['maxdd']*100:.0f}%"
+              f" n={pa['full']['n_trades']} | HOLD Sh={pa['hold']['sharpe']:+.2f} ret={pa['hold']['ret']*100:+.0f}% DD={pa['hold']['maxdd']*100:.0f}%")
+        print(f"      fee: " + " ".join(f"{k}={v:+.2f}" for k, v in pa["fee_sweep"].items()))
+        print(f"      yr: {yr}")
+    return dict(grade=grade, minFull=mf, minHold=mh, minTr=mt, maxDD=mdd, fee_ok=fee_ok_all,
+                per_asset=per_asset)
+
+
+def ensemble_5050_daily(members, fee: float = FEE) -> pd.Series:
+    """50/50 BTC+ETH ensemble daily series (same convention as altlib baseline) for marginal."""
+    sb = _ensemble_daily_asset(members, "BTC", fee)
+    se = _ensemble_daily_asset(members, "ETH", fee)
+    J = pd.concat({"BTC": sb, "ETH": se}, axis=1, join="inner").fillna(0.0)
+    return 0.5 * J["BTC"] + 0.5 * J["ETH"]
+
+
+def marginal_ensemble(members) -> dict:
+    return al.marginal_vs_tp01(ensemble_5050_daily(members))
+
+
+def report(tag, members, names):
+    r = study_ensemble(tag, members)
+    caus = {}
+    for i, p in enumerate(members):
+        cb = sk.causality(p, "BTC")
+        ce = sk.causality(p, "ETH")
+        caus[names[i]] = (cb["ok"], ce["ok"])
+    caus_ok = all(b and e for b, e in caus.values())
+    mg = marginal_ensemble(members)
+    w25 = mg.get("blends", {}).get("w25", {})
+    w50 = mg.get("blends", {}).get("w50", {})
+    earns = (r["grade"] != "FAIL" and mg.get("marginal_verdict") == "ADDS"
+             and mg.get("robust_oos") is True and mg.get("is_hedge") is False)
+    beats = (earns and r["maxDD"] < 0.30 and (w25.get("uplift_hold") or -9) >= 0.55
+             and r["minHold"] >= 0.65)
+    print(f"\n----- MARGINAL [{tag}] -----")
+    print(f"  members: {names}")
+    print(f"  causality per member (BTC,ETH): {caus}  -> all_ok={caus_ok}")
+    print(f"  corr_full={mg.get('corr_full')} corr_hold={mg.get('corr_hold')} verdict={mg.get('marginal_verdict')}")
+    print(f"  has_insample_edge={mg.get('has_insample_edge')} cand_insample_sharpe={mg.get('cand_insample_sharpe')}"
+          f" is_hedge={mg.get('is_hedge')} robust_oos={mg.get('robust_oos')} multicut_persistent={mg.get('multicut_persistent')}")
+    print(f"  clean_year_uplift={mg.get('clean_year_uplift')} jackknife_min_uplift={mg.get('jackknife_min_uplift')}"
+          f" multicut_uplift={mg.get('multicut_uplift')}")
+    print(f"  w25={w25}")
+    print(f"  w50={w50}")
+    print(f"  => earns_slot={earns}  BEATS_WINNER={beats}  (DD={r['maxDD']*100:.0f}% minHold={r['minHold']:+.2f} w25_up_hold={w25.get('uplift_hold')})")
+    return dict(study=r, marginal=mg, caus_ok=caus_ok, earns=earns, beats=beats, w25=w25, w50=w50)
+
+
+if __name__ == "__main__":
+    # ---- Diverse member pool (all pure SkyhookParams variants, all causal) ----
+    # WINNER (ptn_n=45, sl2.5/tp7.0, exits 24/16, vola 35-95, vol_lo 0)
+    P_WIN = WINNER
+    # Faster pattern, tighter stop, shorter TP (different turnover/regime sensitivity)
+    P_FAST = SkyhookParams(ptn_n=25, sl_atr=2.0, tp_atr=5.0, uscitalong=18, uscitashort=12,
+                           vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+    # Slow pattern, wider stop, longer TP (smoother, fewer trades)
+    P_SLOW = SkyhookParams(ptn_n=90, sl_atr=3.0, tp_atr=9.0, uscitalong=30, uscitashort=20,
+                           vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+    # Mid pattern, percent exits (structurally different exit mode) + tighter vola band
+    P_PCT = SkyhookParams(ptn_n=45, exit_mode="pct", sl_pct=0.04, tp_pct=0.10,
+                          uscitalong=24, uscitashort=16, vola_lo=30.0, vola_hi=90.0, vol_lo=0.0)
+    # Low-vol gate variant: add a vol floor + slightly different vola band (regime diversity)
+    P_GATE = SkyhookParams(ptn_n=45, sl_atr=2.5, tp_atr=6.0, uscitalong=24, uscitashort=16,
+                           vola_lo=40.0, vola_hi=95.0, vol_lo=40.0)
+    # DEFENSIVE members: tight stop cuts losers fast -> shallow per-trade DD (the DD-cutters).
+    P_TIGHT = SkyhookParams(ptn_n=45, sl_atr=1.5, tp_atr=4.5, uscitalong=18, uscitashort=12,
+                            vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+    P_TIGHT2 = SkyhookParams(ptn_n=25, sl_atr=1.3, tp_atr=4.0, uscitalong=14, uscitashort=10,
+                             vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+    # Calm-regime gate (sit out high-vola tails) + tight stop -> lowest DD contributor
+    P_CALM = SkyhookParams(ptn_n=45, sl_atr=1.8, tp_atr=5.0, uscitalong=20, uscitashort=14,
+                           vola_lo=20.0, vola_hi=70.0, vol_lo=0.0)
+    # Calm variants: narrower / different vola windows -> diverse DD timing among defenders.
+    P_CALM2 = SkyhookParams(ptn_n=90, sl_atr=1.8, tp_atr=5.5, uscitalong=24, uscitashort=16,
+                            vola_lo=25.0, vola_hi=65.0, vol_lo=0.0)
+    P_CALM3 = SkyhookParams(ptn_n=45, sl_atr=2.0, tp_atr=6.0, uscitalong=24, uscitashort=16,
+                            vola_lo=15.0, vola_hi=60.0, vol_lo=0.0)
+    # CALM4: strong hold-out defensive (wider TP like winner but calm band) — uplift booster
+    P_CALM4 = SkyhookParams(ptn_n=45, sl_atr=2.2, tp_atr=7.0, uscitalong=24, uscitashort=16,
+                            vola_lo=15.0, vola_hi=65.0, vol_lo=0.0)
+    # CALM5: ptn_n=90 calm + wide TP (smooth, strong) for uplift+DD balance
+    P_CALM5 = SkyhookParams(ptn_n=90, sl_atr=2.0, tp_atr=7.0, uscitalong=28, uscitashort=18,
+                            vola_lo=15.0, vola_hi=62.0, vol_lo=0.0)
+
+    POOL = {"WIN": P_WIN, "FAST": P_FAST, "SLOW": P_SLOW, "PCT": P_PCT, "GATE": P_GATE,
+            "TIGHT": P_TIGHT, "TIGHT2": P_TIGHT2, "CALM": P_CALM,
+            "CALM2": P_CALM2, "CALM3": P_CALM3, "CALM4": P_CALM4, "CALM5": P_CALM5}
+
+    # ---- First: standalone DD of each member (diagnostic) ----
+    print("\n===== STANDALONE per-member DD (max over BTC/ETH) =====")
+    for k, p in POOL.items():
+        dds, fhs, hhs = {}, {}, {}
+        for a in CERTIFIED:
+            r = sk.run_asset(a, p, FEE)
+            dds[a] = r["full"]["maxdd"]; fhs[a] = r["full"]["sharpe"]; hhs[a] = r["holdout"]["sharpe"]
+        print(f"  {k:7s}: maxDD={max(dds.values())*100:.0f}% (BTC {dds['BTC']*100:.0f}/ETH {dds['ETH']*100:.0f})"
+              f"  minFull={min(fhs.values()):+.2f} minHold={min(hhs.values()):+.2f}")
+
+    results = {}
+
+    # Smallest K first: K=3, then K=4, K=5 mixes (winner always included).
+    # Focus on DEFENSIVE-heavy mixes to drive standalone DD below 30%.
+    mixes = {
+        # maximize w25 uplift_hold (>=0.55) while keeping DD<30 -> strong-holdout calm members
+        "K3_WIN_CALM3_CALM4":      ["WIN", "CALM3", "CALM4"],
+        "K3_WIN_CALM4_CALM5":      ["WIN", "CALM4", "CALM5"],
+        "K3_WIN_CALM3_CALM5":      ["WIN", "CALM3", "CALM5"],
+        "K3_WIN_PCT_CALM3":        ["WIN", "PCT", "CALM3"],       # PCT has hold 1.0 (uplift booster) but DD 43
+        "K4_WIN_CALM3_CALM4_CALM5":["WIN", "CALM3", "CALM4", "CALM5"],
+        "K3_WIN_CALM3_CALM2":      ["WIN", "CALM3", "CALM2"],     # prior: DD 24, uplift 0.508
+        "K3_WIN_CALM_CALM3":       ["WIN", "CALM", "CALM3"],      # prior: DD 23, uplift 0.493
+        "K4_WIN_GATE_CALM3_CALM4": ["WIN", "GATE", "CALM3", "CALM4"],
+        "K3_WIN_GATE_CALM3":       ["WIN", "GATE", "CALM3"],
+    }
+
+    for tag, keys in mixes.items():
+        members = [POOL[k] for k in keys]
+        print(f"\n########## {tag}  members={keys} ##########")
+        results[tag] = report(tag, members, keys)
+
+    # ---- pick best: prefer BEATS_WINNER, else best by (earns, low DD, hold) ----
+    def score(res):
+        r, w25 = res["study"], res["w25"]
+        uh = (w25.get("uplift_hold") or -9)
+        dd_ok = 1 if r["maxDD"] < 0.30 else 0          # goal #1 first
+        hold_ok = 1 if r["minHold"] >= 0.65 else 0     # goal #2
+        return (1 if res["beats"] else 0, 1 if res["earns"] else 0,
+                dd_ok, hold_ok, uh, -r["maxDD"])
+    best_tag = max(results, key=lambda t: score(results[t]))
+    bres = results[best_tag]
+    br, bw25, bw50, bmg = bres["study"], bres["w25"], bres["w50"], bres["marginal"]
+
+    print("\n\n##################### BEST ENSEMBLE #####################")
+    print(f"BEST = {best_tag}  members={mixes[best_tag]}")
+    print(f"grade={br['grade']} minFull={br['minFull']:+.3f} minHold={br['minHold']:+.3f}"
+          f" max_dd={br['maxDD']:.4f} n_trades_min={br['minTr']} fee_ok(@0.30%)={br['fee_ok']}")
+    print(f"causality_ok={bres['caus_ok']}")
+    print(f"marginal: corr_full={bmg.get('corr_full')} verdict={bmg.get('marginal_verdict')}"
+          f" has_insample_edge={bmg.get('has_insample_edge')} is_hedge={bmg.get('is_hedge')}"
+          f" robust_oos={bmg.get('robust_oos')} multicut_persistent={bmg.get('multicut_persistent')}"
+          f" clean_year_uplift={bmg.get('clean_year_uplift')}")
+    print(f"blend w25 uplift_hold={bw25.get('uplift_hold')} uplift_full={bw25.get('uplift_full')}")
+    print(f"blend w50 full={bw50.get('full')} hold={bw50.get('hold')} dd={bw50.get('dd')}")
+    print(f"earns_slot={bres['earns']}  BEATS_WINNER={bres['beats']}")
+
+    # Emit a machine-readable line so the agent can lift exact numbers.
+    import json
+    print("\nRESULT_JSON " + json.dumps({
+        "best_tag": best_tag, "members": mixes[best_tag],
+        "grade": br["grade"], "minFull": br["minFull"], "minHold": br["minHold"],
+        "max_dd": br["maxDD"], "n_trades_min": br["minTr"], "fee_ok": br["fee_ok"],
+        "causality_ok": bres["caus_ok"],
+        "corr_full": bmg.get("corr_full"), "verdict": bmg.get("marginal_verdict"),
+        "has_insample_edge": bmg.get("has_insample_edge"), "is_hedge": bmg.get("is_hedge"),
+        "robust_oos": bmg.get("robust_oos"), "multicut_persistent": bmg.get("multicut_persistent"),
+        "clean_year_uplift": bmg.get("clean_year_uplift"),
+        "w25_uplift_hold": bw25.get("uplift_hold"), "w50_full": bw50.get("full"),
+        "w50_hold": bw50.get("hold"), "w50_dd": bw50.get("dd"),
+        "earns_slot": bres["earns"], "beats_winner": bres["beats"],
+        "cand_insample_sharpe": bmg.get("cand_insample_sharpe"),
+    }, default=str))

scripts/research/skyhook/runs/SKH2_ENS_STRUCT.py

@@ -0,0 +1,375 @@
+"""SKH2_ENS_STRUCT — cross-definition ENSEMBLE [ENS_STRUCT].
+
+WAVE GOAL: cut the V2 winner's STANDALONE max-DD below 30% (BTC 34% / ETH 31% is the only
+unmet goal) while keeping min-asset HOLD-OUT >= ~0.70 and earns_slot True.
+
+IDEA: the V2 winner uses ONE regime definition (Chande01 cycle band on ATR + Donchian
+breakout). Its drawdowns come from that one signal source firing into the wrong tape. If we
+ensemble it with STRUCTURALLY DIFFERENT regime definitions — (B) causal PERCENTILE-RANK regime
+(SKH_R_PCTL) and (C) VOLATILITY-EXPANSION regime (SKH_R_EXPAND) — that disagree about WHEN to
+trade, their drawdowns are imperfectly correlated. Equal-weighting the three daily-return
+streams (per asset, then 50/50 across BTC+ETH) should reduce the COMBINED-equity DD below any
+single member's DD, at a modest cost to full Sharpe.
+
+The ensemble is EQUAL-WEIGHT on DAILY RETURNS (a constant-weight rebalanced book of three
+sub-sleeves on the SAME asset). Standalone DD = max-DD of the COMBINED equity curve (per asset,
+max over BTC & ETH). Marginal vs TP01 uses the 50/50 BTC+ETH combined daily series.
+
+All three members are causal/leak-free (winner via sk.causality; PCTL/EXPAND via their own
+truncated-prefix guards, re-run here). Equal-weighting causal streams stays causal.
+"""
+from __future__ import annotations
+
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+
+import importlib.util
+import numpy as np
+import pandas as pd
+
+import skyhooklib as sk
+import altlib as al
+from src.backtest.harness import backtest_signals
+from src.strategies import skyhook as S
+from src.strategies.skyhook import SkyhookParams
+
+HOLDOUT = sk.HOLDOUT
+FEE = sk.FEE_RT
+
+
+# --- import the two structural entry builders without running their __main__ ----------------
+def _load(modname, path):
+    spec = importlib.util.spec_from_file_location(modname, path)
+    mod = importlib.util.module_from_spec(spec)
+    spec.loader.exec_module(mod)
+    return mod
+
+PCTL = _load("skr_pctl", "/opt/docker/PythagorasGoal/scripts/research/skyhook/runs/SKH_R_PCTL.py")
+EXPD = _load("skr_expd", "/opt/docker/PythagorasGoal/scripts/research/skyhook/runs/SKH_R_EXPAND.py")
+
+
+# --- the V2 winner from the prior wave (Chande/Donchian regime) -----------------------------
+WINNER = SkyhookParams(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16,
+                       vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+
+# pattern/exit shared knobs for the structural members. Match the WINNER's exit profile
+# (sl2.5/tp7.0, asym time exits) so the difference is the REGIME, not the exit.
+def member_params(**kw):
+    base = dict(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16)
+    base.update(kw)
+    return SkyhookParams(**base)
+
+
+# ============================================================================================
+# Per-asset equity helpers. We need the FULL bar-level equity curve of each member so we can
+# build the COMBINED equity (for an honest combined max-DD), AND the daily-return series (for
+# Sharpe split + marginal). All on the SAME 230m execution index.
+# ============================================================================================
+def member_equity(asset, kind, p, cfg=None):
+    """Return (eq, idx) bar-level equity for a member. kind in {'winner','pctl','expand'}."""
+    ltf, htf = sk.frames(asset)
+    if kind == "winner":
+        ent = S.skyhook_entries(ltf, htf, p)
+    elif kind == "pctl":
+        ent = PCTL.pctl_entries(ltf, htf, p, **cfg)
+    elif kind == "expand":
+        ent = EXPD.expand_entries(ltf, htf, p, **cfg)
+    else:
+        raise ValueError(kind)
+    m = backtest_signals(ltf, ent, fee_rt=FEE, leverage=1.0, asset=asset, tf="230m")
+    idx = pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True))
+    return np.asarray(m.equity, float), idx, int(m.n_trades)
+
+
+def eq_to_daily_ret(eq, idx):
+    """bar equity -> daily simple returns (last-of-day, ffill, pct_change)."""
+    s = pd.Series(eq, index=idx)
+    return s.resample("1D").last().ffill().pct_change()
+
+
+def combined_daily_ret(asset, members):
+    """Equal-weight DAILY returns of the members on one asset -> combined daily return series.
+    Each member contributes its own daily return; the equal-weight portfolio return is the mean.
+    Members are aligned on the union of daily timestamps; a member that has not started yet
+    (NaN) contributes 0 that day and is dropped from the active-weight denominator (outer-join
+    with renormalized equal weights), so early days where only some members trade are handled."""
+    drs = {}
+    ntr = {}
+    for name, kind, p, cfg in members:
+        eq, idx, nt = member_equity(asset, kind, p, cfg)
+        drs[name] = eq_to_daily_ret(eq, idx)
+        ntr[name] = nt
+    D = pd.concat(drs, axis=1, join="outer")
+    # renormalized equal weight across the members that are ACTIVE (non-NaN) each day
+    active = D.notna()
+    w = active.div(active.sum(axis=1).replace(0, np.nan), axis=0)
+    comb = (D.fillna(0.0) * w.fillna(0.0)).sum(axis=1)
+    comb = comb.dropna()
+    return comb, ntr
+
+
+def combined_metrics(comb):
+    """Sharpe (full + hold-out) and max-DD from a DAILY combined return series."""
+    comb = comb[np.isfinite(comb.values)]
+    eq = (1.0 + comb).cumprod().values
+    idx = comb.index
+
+    def _sh(r):
+        r = r[np.isfinite(r)]
+        return float(np.mean(r) / np.std(r) * np.sqrt(365.25)) if len(r) and np.std(r) > 0 else 0.0
+
+    pk = np.maximum.accumulate(eq)
+    dd = float(np.max((pk - eq) / pk)) if len(eq) else 0.0
+    full_sh = _sh(comb.values)
+    hmask = idx >= HOLDOUT
+    hold_sh = _sh(comb.values[hmask]) if hmask.sum() > 5 else 0.0
+    # hold-out DD on the hold-out slice of the combined equity
+    eqh = eq[hmask]
+    pkh = np.maximum.accumulate(eqh) if len(eqh) else np.array([1.0])
+    ddh = float(np.max((pkh - eqh) / pkh)) if len(eqh) else 0.0
+    yrs = {}
+    for y in sorted(set(idx.year)):
+        rr = comb.values[idx.year == y]
+        yrs[int(y)] = round(float((1 + pd.Series(rr)).prod() - 1), 4)
+    return dict(full_sharpe=round(full_sh, 3), hold_sharpe=round(hold_sh, 3),
+                maxdd=round(dd, 4), hold_maxdd=round(ddh, 4), yearly=yrs)
+
+
+# --- per-member standalone DD (for the correlation/decorrelation diagnostic) ----------------
+def member_standalone_dd(asset, kind, p, cfg=None):
+    eq, idx, nt = member_equity(asset, kind, p, cfg)
+    pk = np.maximum.accumulate(eq)
+    dd = float(np.max((pk - eq) / pk)) if len(eq) else 0.0
+    return dd, nt
+
+
+# ============================================================================================
+# Fee robustness of the ENSEMBLE: re-run every member at fee f, recombine, check Sharpe>0.
+# ============================================================================================
+def ensemble_fee_sweep(members, fees=(0.0, 0.001, 0.002, 0.003)):
+    rows = {}
+    for f in fees:
+        ok = True
+        for asset in ("BTC", "ETH"):
+            drs = {}
+            for name, kind, p, cfg in members:
+                ltf, htf = sk.frames(asset)
+                if kind == "winner":
+                    ent = S.skyhook_entries(ltf, htf, p)
+                elif kind == "pctl":
+                    ent = PCTL.pctl_entries(ltf, htf, p, **cfg)
+                else:
+                    ent = EXPD.expand_entries(ltf, htf, p, **cfg)
+                m = backtest_signals(ltf, ent, fee_rt=f, leverage=1.0, asset=asset, tf="230m")
+                drs[name] = eq_to_daily_ret(np.asarray(m.equity, float),
+                                            pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True)))
+            D = pd.concat(drs, axis=1, join="outer")
+            active = D.notna()
+            w = active.div(active.sum(axis=1).replace(0, np.nan), axis=0)
+            comb = (D.fillna(0.0) * w.fillna(0.0)).sum(axis=1).dropna()
+            sh = combined_metrics(comb)["full_sharpe"]
+            rows[(f, asset)] = sh
+            ok = ok and (sh > 0)
+        rows[(f, "ok")] = ok
+    return rows
+
+
+def ensemble_daily_5050(members):
+    """50/50 BTC+ETH combined daily series for marginal_vs_tp01."""
+    cb, _ = combined_daily_ret("BTC", members)
+    ce, _ = combined_daily_ret("ETH", members)
+    J = pd.concat({"BTC": cb, "ETH": ce}, axis=1, join="inner").fillna(0.0)
+    return 0.5 * J["BTC"] + 0.5 * J["ETH"]
+
+
+# ============================================================================================
+def run_ensemble(tag, members):
+    print(f"\n========== ENSEMBLE: {tag} ==========")
+    print(f"  members: {[m[0] for m in members]}")
+    per = {}
+    for asset in ("BTC", "ETH"):
+        comb, ntr = combined_daily_ret(asset, members)
+        met = combined_metrics(comb)
+        per[asset] = dict(met=met, ntr=ntr)
+        print(f"  {asset}: FULL Sh={met['full_sharpe']:+.2f} HOLD Sh={met['hold_sharpe']:+.2f}"
+              f" maxDD={met['maxdd']*100:.0f}% holdDD={met['hold_maxdd']*100:.0f}%  ntrades={ntr}")
+        print(f"       yearly: " + " ".join(f"{y}:{v*100:+.0f}%" for y, v in met['yearly'].items()))
+    min_full = min(per[a]["met"]["full_sharpe"] for a in per)
+    min_hold = min(per[a]["met"]["hold_sharpe"] for a in per)
+    max_dd = max(per[a]["met"]["maxdd"] for a in per)
+    n_trades_min = min(sum(per[a]["ntr"].values()) for a in per)
+    print(f"  -> minFull={min_full:+.2f} minHold={min_hold:+.2f} maxDD={max_dd*100:.0f}%"
+          f" nTrades(min,sum-of-members)={n_trades_min}")
+    return dict(per=per, min_full=min_full, min_hold=min_hold, max_dd=max_dd,
+                n_trades_min=n_trades_min)
+
+
+if __name__ == "__main__":
+    print("=== SKH2_ENS_STRUCT: cross-definition regime ensemble (DD reduction) ===")
+
+    # --- 0) Baseline: the V2 winner ALONE (reference) -----------------------------------
+    print("\n--- V2 WINNER standalone (Chande/Donchian) ---")
+    w_per = {}
+    for a in ("BTC", "ETH"):
+        r = sk.run_asset(a, WINNER, FEE)
+        w_per[a] = r
+        print(f"  {a}: FULL Sh={r['full']['sharpe']:+.2f} DD={r['full']['maxdd']*100:.0f}%"
+              f" n={r['full']['n_trades']} | HOLD Sh={r['holdout']['sharpe']:+.2f}")
+    w_minfull = min(w_per[a]["full"]["sharpe"] for a in w_per)
+    w_minhold = min(w_per[a]["holdout"]["sharpe"] for a in w_per)
+    w_maxdd = max(w_per[a]["full"]["maxdd"] for a in w_per)
+    print(f"  WINNER: minFull={w_minfull:+.2f} minHold={w_minhold:+.2f} maxDD={w_maxdd*100:.0f}%")
+
+    # --- 1) Structural member configs (different regime definitions) --------------------
+    pP = member_params()
+    pE = member_params()
+    # PCTL: expanding percentile-rank regime. From SKH_R_PCTL_final, the lower-DD / robust
+    # band was the low-vola band; we use a mid band with a modest volume floor so it disagrees
+    # with the winner's HIGH-vola Chande band (different regime -> decorrelated DD).
+    PCTL_CFG = dict(vola_win=None, vol_win=None, vola_lo=0.0, vola_hi=0.70, vol_lo=0.0, vol_hi=1.0)
+    PCTL_CFG2 = dict(vola_win=None, vol_win=None, vola_lo=0.10, vola_hi=0.80, vol_lo=0.30, vol_hi=1.0)
+    # EXPAND: volatility-expansion regime (ATR above its MA + volume elevated).
+    EXP_CFG = dict(w_atr=20, k_atr=1.00, w_vol=20, k_vol=1.00)
+    EXP_CFG2 = dict(w_atr=20, k_atr=1.10, w_vol=20, k_vol=1.20)
+
+    # --- standalone DD + pairwise correlation of the three regime streams (BTC) ---------
+    print("\n--- standalone member DD + pairwise daily-return corr (decorrelation check) ---")
+    for a in ("BTC", "ETH"):
+        dW, nW = member_standalone_dd(a, "winner", WINNER)
+        dP, nP = member_standalone_dd(a, "pctl", pP, PCTL_CFG)
+        dE, nE = member_standalone_dd(a, "expand", pE, EXP_CFG)
+        print(f"  {a} standalone DD: winner={dW*100:.0f}%(n{nW}) pctl={dP*100:.0f}%(n{nP}) expand={dE*100:.0f}%(n{nE})")
+        # corr of daily returns
+        rw = eq_to_daily_ret(*member_equity(a, "winner", WINNER)[:2])
+        rp = eq_to_daily_ret(*member_equity(a, "pctl", pP, PCTL_CFG)[:2])
+        re_ = eq_to_daily_ret(*member_equity(a, "expand", pE, EXP_CFG)[:2])
+        DD = pd.concat({"W": rw, "P": rp, "E": re_}, axis=1, join="inner").fillna(0.0)
+        cc = DD.corr()
+        print(f"     corr W-P={cc.loc['W','P']:.2f} W-E={cc.loc['W','E']:.2f} P-E={cc.loc['P','E']:.2f}")
+
+    # --- 2) Candidate ensembles ---------------------------------------------------------
+    candidates = {
+        "WPE (winner+pctlLo+expand)": [
+            ("winner", "winner", WINNER, None),
+            ("pctl", "pctl", pP, PCTL_CFG),
+            ("expand", "expand", pE, EXP_CFG),
+        ],
+        "WP (winner+pctlLo)": [
+            ("winner", "winner", WINNER, None),
+            ("pctl", "pctl", pP, PCTL_CFG),
+        ],
+        "WE (winner+expand)": [
+            ("winner", "winner", WINNER, None),
+            ("expand", "expand", pE, EXP_CFG),
+        ],
+        "WPE2 (winner+pctlMid+expandStrong)": [
+            ("winner", "winner", WINNER, None),
+            ("pctl", "pctl", pP, PCTL_CFG2),
+            ("expand", "expand", pE, EXP_CFG2),
+        ],
+    }
+
+    results = {}
+    for tag, members in candidates.items():
+        results[tag] = (members, run_ensemble(tag, members))
+
+    # --- 3) Pick best by: max_dd<0.30, then maximize min_hold ---------------------------
+    def score(v):
+        # prefer DD<0.30 AND min_hold>=0.65; among those, maximize min_hold then -DD
+        ok = v["max_dd"] < 0.30 and v["min_hold"] >= 0.65
+        return (1 if ok else 0, v["min_hold"], -v["max_dd"])
+    best_tag = max(results, key=lambda t: score(results[t][1]))
+    best_members, best_v = results[best_tag]
+    print(f"\n*** BEST ENSEMBLE = {best_tag} ***")
+    print(f"    minFull={best_v['min_full']:+.2f} minHold={best_v['min_hold']:+.2f}"
+          f" maxDD={best_v['max_dd']*100:.0f}% nTradesMin={best_v['n_trades_min']}")
+
+    # --- 4) Causality: winner via sk; structural members via their own guards -----------
+    print("\n--- causality (all active members of best ensemble) ---")
+    caus_ok = True
+    kinds = {m[0]: (m[1], m[2], m[3]) for m in best_members}
+    if "winner" in kinds:
+        cb = sk.causality(WINNER, "BTC"); ce = sk.causality(WINNER, "ETH")
+        print(f"  winner: BTC {cb} ETH {ce}")
+        caus_ok = caus_ok and cb["ok"] and ce["ok"]
+    if "pctl" in kinds:
+        _, p, cfg = kinds["pctl"]
+        cb = PCTL.check_causality(cfg, p, "BTC"); ce = PCTL.check_causality(cfg, p, "ETH")
+        print(f"  pctl:   BTC {cb} ETH {ce}")
+        caus_ok = caus_ok and cb["ok"] and ce["ok"]
+    if "expand" in kinds:
+        _, p, cfg = kinds["expand"]
+        cb = EXPD.check_causality(cfg, p, "BTC"); ce = EXPD.check_causality(cfg, p, "ETH")
+        print(f"  expand: BTC {cb} ETH {ce}")
+        caus_ok = caus_ok and cb["ok"] and ce["ok"]
+    print(f"  causality_ok (all members) = {caus_ok}")
+
+    # --- 5) Fee sweep on the ensemble ---------------------------------------------------
+    print("\n--- ensemble fee sweep (FULL Sharpe per asset) ---")
+    fsw = ensemble_fee_sweep(best_members)
+    for f in (0.0, 0.001, 0.002, 0.003):
+        print(f"  {f*100:.2f}%RT: BTC={fsw[(f,'BTC')]:+.2f} ETH={fsw[(f,'ETH')]:+.2f} ok={fsw[(f,'ok')]}")
+    fee_survives = fsw[(0.003, "ok")]
+    print(f"  fee_survives 0.30%RT (both): {fee_survives}")
+
+    # --- 6) Marginal vs TP01 on the ensemble 50/50 series -------------------------------
+    print("\n--- marginal vs TP01 (best ensemble, 50/50 BTC+ETH) ---")
+    cand = ensemble_daily_5050(best_members)
+    marg = al.marginal_vs_tp01(cand)
+    corr_full = marg.get("corr_full")
+    verdict = marg.get("marginal_verdict")
+    has_edge = marg.get("has_insample_edge")
+    is_hedge = marg.get("is_hedge")
+    robust_oos = marg.get("robust_oos")
+    multicut = marg.get("multicut_persistent")
+    clean_year = marg.get("clean_year_uplift")
+    w25 = marg.get("blends", {}).get("w25", {})
+    w50 = marg.get("blends", {}).get("w50", {})
+    up25_hold = w25.get("uplift_hold")
+    print(f"  corr_full={corr_full} corr_hold={marg.get('corr_hold')}")
+    print(f"  marginal_verdict={verdict} robust_oos={robust_oos} multicut_persistent={multicut}")
+    print(f"  has_insample_edge={has_edge} (cand_insample_sharpe={marg.get('cand_insample_sharpe')}) is_hedge={is_hedge}")
+    print(f"  clean_year_uplift={clean_year} jackknife_min_uplift={marg.get('jackknife_min_uplift')}")
+    print(f"  multicut_uplift={marg.get('multicut_uplift')}")
+    print(f"  blend w25: uplift_hold={up25_hold} uplift_full={w25.get('uplift_full')} dd={w25.get('dd')}")
+    print(f"  blend w50: full={w50.get('full')} hold={w50.get('hold')} uplift_hold={w50.get('uplift_hold')} dd={w50.get('dd')}")
+
+    # --- 7) Grade + earns_slot + beats_winner ------------------------------------------
+    n_trades_min = best_v["n_trades_min"]
+    grade = "PASS" if (n_trades_min >= 20 and best_v["min_full"] >= 0.5
+                       and best_v["min_hold"] >= 0.2 and fee_survives) else \
+            ("WEAK" if (n_trades_min >= 20 and best_v["min_full"] >= 0.3
+                        and best_v["min_hold"] >= 0.0) else "FAIL")
+    earns_slot = (grade != "FAIL") and verdict == "ADDS" and robust_oos and (not is_hedge)
+    beats = (earns_slot and best_v["max_dd"] < 0.30
+             and (up25_hold is not None and up25_hold >= 0.55)
+             and best_v["min_hold"] >= 0.65)
+
+    print("\n=========== FINAL ===========")
+    print(f"BEST CONFIG = {best_tag}")
+    print(f"  members:")
+    for m in best_members:
+        print(f"    {m[0]}: kind={m[1]} cfg={m[3]}")
+    print(f"  minFull={best_v['min_full']:+.2f}  minHold={best_v['min_hold']:+.2f}"
+          f"  max_dd={best_v['max_dd']*100:.1f}%  n_trades_min={n_trades_min}")
+    print(f"  fee@0.30%RT survives={fee_survives}  causality_ok={caus_ok}  grade={grade}")
+    print(f"  marginal: corr_full={corr_full} verdict={verdict} insample_edge={has_edge}"
+          f" is_hedge={is_hedge} robust_oos={robust_oos} multicut={multicut}")
+    print(f"            clean_year_uplift={clean_year}  blend_w25_uplift_hold={up25_hold}")
+    print(f"            blend_w50: full={w50.get('full')} hold={w50.get('hold')} dd={w50.get('dd')}")
+    print(f"  earns_slot={earns_slot}  beats_winner={beats}")
+    print(f"\n  (WINNER ref: minFull={w_minfull:+.2f} minHold={w_minhold:+.2f} maxDD={w_maxdd*100:.0f}%)")
+
+    # machine-readable tail for the agent
+    import json
+    print("\nRESULT_JSON=" + json.dumps({
+        "best_tag": best_tag,
+        "best_config": {"members": [{"name": m[0], "kind": m[1], "cfg": m[3]} for m in best_members]},
+        "min_full": best_v["min_full"], "min_hold": best_v["min_hold"],
+        "max_dd": best_v["max_dd"], "n_trades_min": n_trades_min,
+        "fee_survives": bool(fee_survives), "causality_ok": bool(caus_ok), "grade": grade,
+        "corr_full": corr_full, "marginal_verdict": verdict, "has_insample_edge": bool(has_edge),
+        "is_hedge": bool(is_hedge), "robust_oos": bool(robust_oos),
+        "multicut_persistent": bool(multicut), "clean_year_uplift": clean_year,
+        "blend_w25_uplift_hold": up25_hold, "earns_slot": bool(earns_slot), "beats_winner": bool(beats),
+    }, default=str))

scripts/research/skyhook/runs/SKH2_EXPAND_DD.py

@@ -0,0 +1,258 @@
+"""SKH2_EXPAND_DD — DD-reduction wave, vol-EXPANSION family.
+
+Family task: reuse the volatility-EXPANSION regime from SKH_R_EXPAND.py (ATR rising vs its own
+MA AND volume elevated vs its own MA), monkeypatch S.htf_features, run sk.study, and TUNE
+w_atr/k_atr/w_vol/k_vol + winner-style exits to:
+  (1) cut standalone maxDD below 30% (max over BTC&ETH)  <-- the only unmet wave goal
+  (2) keep min-asset HOLD-OUT Sharpe >= ~0.70 and earns_slot == True
+  (3) stretch: lift blend w25 uplift_hold and minHold.
+
+Mechanism / DD theory:
+  * the EXPANSION gate (vol rising + volume elevated) is itself a DD filter: it suppresses
+    entries during quiet/contracting chop where Donchian breakouts whipsaw. Tightening k_atr /
+    k_vol trades trade-count for cleaner regime -> fewer adverse entries.
+  * but per-trade loss size is set by sl_atr; the V2 winner used sl_atr=2.5 (DD 34/31%).
+    Lowering sl_atr is the direct DD lever. We sweep sl_atr in {1.6,1.8,2.0,2.2,2.5} and
+    couple it with the winner exits (uscitalong=24/uscitashort=16) and tp_atr in {5,6,7}.
+  * vola_lo/vola_hi/vol_lo bands are IRRELEVANT here: the expansion regime REPLACES the Chande
+    band gate (htf_features is monkeypatched), so those SkyhookParams fields are dead. Only
+    ptn_n / sl_atr / tp_atr / uscita* / max_per_day / long_only matter through the patched path.
+
+Everything causal: the expansion features use only x[0..i] (causal rolling MA, ATR ewm, donchian
+shift(1)); HTF merged BACKWARD onto LTF on HTF-close ts. We verify with sk.causality (works
+because we patch S.htf_features inside skyhooklib's namespace, so skyhook_entries uses our gate).
+"""
+from __future__ import annotations
+
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+
+import numpy as np
+import pandas as pd
+
+import skyhooklib as sk
+from src.strategies import skyhook as S
+from src.strategies.skyhook import SkyhookParams
+
+# reuse the EXPANSION feature builder verbatim
+from SKH_R_EXPAND import expand_htf_features
+
+ORIG_FEAT = S.htf_features
+HOLDOUT = sk.HOLDOUT
+FEE = sk.FEE_RT
+
+
+def patched(cfg):
+    def _feat(htf, p):
+        return expand_htf_features(htf, p, **cfg)
+    return _feat
+
+
+def study_expand(name, p, cfg, want_marginal=True):
+    """Run sk.study + causality (+ marginal) with htf_features patched to the expansion regime."""
+    S.htf_features = patched(cfg)
+    try:
+        rep = sk.study(name, p)
+        caus_b = sk.causality(p, "BTC")
+        caus_e = sk.causality(p, "ETH")
+        marg = sk.marginal(p) if want_marginal else None
+    finally:
+        S.htf_features = ORIG_FEAT
+    return rep, (caus_b, caus_e), marg
+
+
+def vline(rep):
+    v = rep["verdict"]
+    pa = rep["per_asset"]
+    mdd = max(pa[a]["full"]["maxdd"] for a in pa)
+    return (v["grade"], v["min_asset_full_sharpe"], v["min_asset_holdout_sharpe"],
+            v["min_trades"], mdd, v["fee_survives"])
+
+
+# ---------------------------------------------------------------------------
+# WINNER baseline (Chande band, NOT expansion) for reference — verify the stated DD problem.
+# ---------------------------------------------------------------------------
+def winner_reference():
+    p = SkyhookParams(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16,
+                      vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+    rep = sk.study("WINNER-V2", p)   # uses ORIG_FEAT (Chande band) — not patched
+    g, mf, mh, mt, mdd, fee = vline(rep)
+    print(f"[WINNER-V2 ref] grade={g} minFull={mf:+.2f} minHold={mh:+.2f} minTr={mt} "
+          f"maxDD={mdd*100:.0f}% feeOK={fee}  "
+          f"(BTC DD {rep['per_asset']['BTC']['full']['maxdd']*100:.0f}% / "
+          f"ETH DD {rep['per_asset']['ETH']['full']['maxdd']*100:.0f}%)")
+    return mh
+
+
+def earns_slot(rep, marg):
+    g = rep["verdict"]["grade"] != "FAIL"
+    return bool(g and marg.get("marginal_verdict") == "ADDS"
+                and marg.get("robust_oos") and not marg.get("is_hedge"))
+
+
+if __name__ == "__main__":
+    print("=== SKH2_EXPAND_DD: vol-EXPANSION regime tuned for standalone maxDD < 30% ===\n")
+    win_minhold = winner_reference()
+    print()
+
+    # -------------------------------------------------------------------
+    # PASS 1: coarse grid. Regime gate strength (k_atr,k_vol,windows) x SL size.
+    # Goal: find DD<30% cells that keep minHold high. Marginal computed only for finalists.
+    # -------------------------------------------------------------------
+    # exits: asymmetric time-exits. PASS1 learned that LONGER long-holds (us30/18 vs winner's
+    # 24/16) are what flip the marginal robust_oos gate POSITIVE (clean-2025-year uplift > 0)
+    # while sl_atr=2.4 keeps DD<30. So we sweep exits + sl_atr here, ptn_n fixed near winner.
+    base_kw = dict(ptn_n=45, uscitalong=30, uscitashort=18)
+
+    # The EXPANSION gate REPLACES the Chande band (htf_features monkeypatched): vola_*/vol_* are
+    # dead. DD is cut by (a) the gate itself (only trade rising-vol + elevated-volume regimes) and
+    # (b) sl_atr. The a20/k1.1 gate + sl2.4 + us30/18 is the DD<30 + robust_oos sweet spot found.
+    regimes = {
+        # tag: expansion cfg
+        "r_a20k1.1_v20k1.1": dict(w_atr=20, k_atr=1.10, w_vol=20, k_vol=1.10),
+        "r_a25k1.1_v25k1.1": dict(w_atr=25, k_atr=1.10, w_vol=25, k_vol=1.10),
+        "r_a30k1.1_v30k1.1": dict(w_atr=30, k_atr=1.10, w_vol=30, k_vol=1.10),
+        "r_a18k1.1_v18k1.1": dict(w_atr=18, k_atr=1.10, w_vol=18, k_vol=1.10),
+    }
+    sl_grid = (2.2, 2.4)
+    tp_fixed = 7.0
+
+    print("--- PASS 1 coarse: regime x sl_atr (tp=7.0) ---")
+    pass1 = []
+    for rtag, rcfg in regimes.items():
+        for sl in sl_grid:
+            p = SkyhookParams(sl_atr=sl, tp_atr=tp_fixed, **base_kw)
+            S.htf_features = patched(rcfg)
+            try:
+                rep = sk.study(f"{rtag}_sl{sl}", p)
+            finally:
+                S.htf_features = ORIG_FEAT
+            g, mf, mh, mt, mdd, fee = vline(rep)
+            pass1.append((rtag, rcfg, sl, tp_fixed, g, mf, mh, mt, mdd, fee, p))
+            print(f"  {rtag:22s} sl{sl} -> grade={g:4s} minFull={mf:+.2f} minHold={mh:+.2f}"
+                  f" minTr={mt:3d} maxDD={mdd*100:3.0f}% feeOK={fee}")
+
+    # finalists = DD<30% AND minHold>=0.55 AND grade!=FAIL AND fee survives
+    fin = [r for r in pass1 if r[8] < 0.30 and r[6] >= 0.55 and r[4] != "FAIL" and r[9]]
+    print(f"\n--- PASS1 finalists (DD<30%, minHold>=0.6, !FAIL, feeOK): {len(fin)} ---")
+    for r in fin:
+        print(f"  {r[0]} sl{r[2]} tp{r[3]} : minHold={r[6]:+.2f} DD={r[8]*100:.0f}%")
+
+    # If none, relax to DD<30% AND minHold>=0.5 to still report best-effort.
+    if not fin:
+        fin = [r for r in pass1 if r[8] < 0.30 and r[6] >= 0.50 and r[4] != "FAIL" and r[9]]
+        print(f"  (relaxed minHold>=0.5): {len(fin)}")
+    if not fin:
+        # last resort: lowest DD among non-FAIL fee-surviving with minHold>0
+        cand = [r for r in pass1 if r[4] != "FAIL" and r[9] and r[6] > 0]
+        fin = sorted(cand, key=lambda r: r[8])[:3]
+        print(f"  (last-resort lowest-DD): {len(fin)}")
+
+    # -------------------------------------------------------------------
+    # PASS 2: finalists -> full marginal + tighten tp around best. Pick the BEATS-WINNER one,
+    # else best earns_slot+lowest DD.
+    # -------------------------------------------------------------------
+    # de-dup finalists by (rtag,sl) and cap to keep runtime sane
+    seen = set(); fin2 = []
+    for r in sorted(fin, key=lambda r: (-r[6], r[8])):  # prefer high minHold then low DD
+        key = (r[0], r[2])
+        if key in seen:
+            continue
+        seen.add(key); fin2.append(r)
+    fin2 = fin2[:7]
+
+    print(f"\n--- PASS 2 marginal on {len(fin2)} finalists ---")
+    results = []
+    for r in fin2:
+        rtag, rcfg, sl, tp, g, mf, mh, mt, mdd, fee, p = r
+        rep, (cb, ce), marg = study_expand(f"{rtag}_sl{sl}_tp{tp}", p, rcfg)
+        g, mf, mh, mt, mdd, fee = vline(rep)
+        caus_ok = bool(cb["ok"] and ce["ok"])
+        es = earns_slot(rep, marg)
+        w25 = marg.get("blends", {}).get("w25", {})
+        w50 = marg.get("blends", {}).get("w50", {})
+        uph = w25.get("uplift_hold")
+        beats = bool(es and mdd < 0.30 and (uph is not None and uph >= 0.55) and mh >= 0.65)
+        results.append(dict(tag=f"{rtag}_sl{sl}_tp{tp}", rcfg=rcfg, p=p, rep=rep, marg=marg,
+                            caus_ok=caus_ok, earns=es, beats=beats,
+                            minFull=mf, minHold=mh, minTr=mt, maxDD=mdd, fee=fee,
+                            uph=uph, w25=w25, w50=w50))
+        print(f"  {rtag}_sl{sl} -> grade={g} minFull={mf:+.2f} minHold={mh:+.2f} DD={mdd*100:.0f}%"
+              f" verdict={marg.get('marginal_verdict')} corr={marg.get('corr_full')}"
+              f" w25uplH={uph} earns={es} caus={caus_ok} BEATS={beats}")
+
+    # -------------------------------------------------------------------
+    # PASS 3: around the best finalist, try tp in {5,6} to see if tighter tp helps DD/minHold.
+    # -------------------------------------------------------------------
+    def score(d):
+        # rank: beats first, then earns & DD<30, then minHold, then -DD
+        return (d["beats"], d["earns"] and d["maxDD"] < 0.30, d["minHold"], -d["maxDD"])
+
+    if results:
+        best = max(results, key=score)
+        rtag = best["tag"].rsplit("_sl", 1)[0]
+        rcfg = best["rcfg"]
+        sl = best["p"].sl_atr
+        # PASS3 sweeps the EXIT-BAR dimension: the robust_oos (2025-clean-year uplift) gate is
+        # set by the long-hold length. We probe uscitalong around 30 to confirm the sweet spot
+        # and hunt any DD<30 cell with higher blend uplift.
+        print(f"\n--- PASS 3 exit-bar refine around best regime={rtag} sl{sl} ---")
+        for usL, usS in ((28, 18), (32, 18), (30, 20)):
+            kw = dict(ptn_n=45, uscitalong=usL, uscitashort=usS)
+            p = SkyhookParams(sl_atr=sl, tp_atr=tp_fixed, **kw)
+            rep, (cb, ce), marg = study_expand(f"{rtag}_sl{sl}_us{usL}/{usS}", p, rcfg)
+            g, mf, mh, mt, mdd, fee = vline(rep)
+            caus_ok = bool(cb["ok"] and ce["ok"])
+            es = earns_slot(rep, marg)
+            w25 = marg.get("blends", {}).get("w25", {}); w50 = marg.get("blends", {}).get("w50", {})
+            uph = w25.get("uplift_hold")
+            beats = bool(es and mdd < 0.30 and (uph is not None and uph >= 0.55) and mh >= 0.65)
+            results.append(dict(tag=f"{rtag}_sl{sl}_us{usL}/{usS}", rcfg=rcfg, p=p, rep=rep, marg=marg,
+                                caus_ok=caus_ok, earns=es, beats=beats,
+                                minFull=mf, minHold=mh, minTr=mt, maxDD=mdd, fee=fee,
+                                uph=uph, w25=w25, w50=w50))
+            print(f"  us{usL}/{usS} -> grade={g} minFull={mf:+.2f} minHold={mh:+.2f} DD={mdd*100:.0f}%"
+                  f" verdict={marg.get('marginal_verdict')} robust={marg.get('robust_oos')}"
+                  f" w25uplH={uph} earns={es} BEATS={beats}")
+
+    # -------------------------------------------------------------------
+    # FINAL: pick best config and print full block.
+    # -------------------------------------------------------------------
+    if not results:
+        print("\n!!! no finalists at all — reporting nothing meaningful. !!!")
+        sys.exit(0)
+
+    best = max(results, key=score)
+    m = best["marg"]; rep = best["rep"]
+    print("\n" + "=" * 78)
+    print("FINAL BEST (vol-EXPANSION family)")
+    print("=" * 78)
+    print(f"  tag         = {best['tag']}")
+    print(f"  regime cfg  = {best['rcfg']}")
+    print(f"  params      = ptn_n={best['p'].ptn_n} sl_atr={best['p'].sl_atr} tp_atr={best['p'].tp_atr}"
+          f" uscitalong={best['p'].uscitalong} uscitashort={best['p'].uscitashort}"
+          f" max_per_day={best['p'].max_per_day} long_only={best['p'].long_only}")
+    print(f"  minFull     = {best['minFull']:+.3f}")
+    print(f"  minHold     = {best['minHold']:+.3f}")
+    print(f"  max_dd      = {best['maxDD']:.4f}  ({best['maxDD']*100:.1f}%)")
+    print(f"  n_trades    = {best['minTr']}  (min over BTC&ETH)")
+    print(f"  fee@0.30%RT survives = {best['fee']}")
+    print(f"  causality OK (BTC&ETH) = {best['caus_ok']}")
+    print(f"  earns_slot  = {best['earns']}")
+    print(f"  BEATS_WINNER= {best['beats']}")
+    print("  -- per-asset --")
+    for a in ("BTC", "ETH"):
+        pa = rep["per_asset"][a]
+        print(f"    {a}: FULL Sh={pa['full']['sharpe']:+.2f} DD={pa['full']['maxdd']*100:.0f}%"
+              f" n={pa['full']['n_trades']} | HOLD Sh={pa['holdout']['sharpe']:+.2f}"
+              f" | fee_sweep {pa['fee_sweep']}")
+    print("  -- marginal vs TP01 --")
+    print(f"    corr_full={m.get('corr_full')} corr_hold={m.get('corr_hold')}")
+    print(f"    marginal_verdict={m.get('marginal_verdict')}")
+    print(f"    has_insample_edge={m.get('has_insample_edge')} is_hedge={m.get('is_hedge')}")
+    print(f"    robust_oos={m.get('robust_oos')} multicut_persistent={m.get('multicut_persistent')}")
+    print(f"    clean_year_uplift={m.get('clean_year_uplift')} jackknife_min_uplift={m.get('jackknife_min_uplift')}")
+    print(f"    cand_insample_sharpe={m.get('cand_insample_sharpe')} multicut_uplift={m.get('multicut_uplift')}")
+    print(f"    blend w25={m.get('blends',{}).get('w25')}")
+    print(f"    blend w50={m.get('blends',{}).get('w50')}")
+    print(f"\n  win_minhold(reference)={win_minhold:+.2f}")

scripts/research/skyhook/runs/SKH2_FREQ.py

@@ -0,0 +1,206 @@
+"""SKH2_FREQ — entry cadence / holding-period family for the SKH01 DD-reduction wave.
+
+Goal: cut standalone maxDD below 30% (max over BTC & ETH) while keeping min-asset HOLD-OUT
+Sharpe >= ~0.70 and earns_slot == True. Lever space (all expressible via SkyhookParams):
+  * max_per_day {1, 2}
+  * uscitalong / uscitashort holding windows {12..30}
+  * atr_win (HTF) / ltf_atr_win (exec) windows
+
+Baseline-to-beat (verified V2 winner):
+  SkyhookParams(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16,
+                vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+  minFull +0.83 minHold +0.81 maxDD BTC34%/ETH31% earns_slot True
+  blend w25 uplift_hold +0.58, w50 full1.59/hold1.04/DD12.5%.
+
+A candidate BEATS the winner iff: earns_slot AND max_dd<0.30 AND blend_w25_uplift_hold>=0.55
+AND min_hold_sharpe>=0.65.
+"""
+from __future__ import annotations
+import sys, json
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+import skyhooklib as sk
+from src.strategies.skyhook import SkyhookParams
+
+# Winner base (all FREQ variants share the regime/pattern/stop structure of the winner).
+WINNER = dict(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16,
+              vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+
+
+def mk(**over) -> SkyhookParams:
+    d = dict(WINNER); d.update(over)
+    return SkyhookParams(**d)
+
+
+def quick(name, p) -> dict:
+    """Fast screen: FULL+HOLD on both assets + standalone maxDD. No fee sweep / marginal yet."""
+    rb = sk.run_asset("BTC", p)
+    re = sk.run_asset("ETH", p)
+    minF = min(rb["full"]["sharpe"], re["full"]["sharpe"])
+    minH = min(rb["holdout"]["sharpe"], re["holdout"]["sharpe"])
+    maxdd = max(rb["full"]["maxdd"], re["full"]["maxdd"])
+    minTr = min(rb["full"]["n_trades"], re["full"]["n_trades"])
+    print(f"  {name:38s} minF={minF:+.2f} minH={minH:+.2f} maxDD={maxdd*100:4.0f}% "
+          f"(B{rb['full']['maxdd']*100:.0f}/E{re['full']['maxdd']*100:.0f}) "
+          f"nTr={minTr} | Bh={rb['holdout']['sharpe']:+.2f} Eh={re['holdout']['sharpe']:+.2f}")
+    return dict(name=name, p=p, minF=minF, minH=minH, maxdd=maxdd, minTr=minTr,
+                bdd=rb["full"]["maxdd"], edd=re["full"]["maxdd"])
+
+
+def full_eval(name, p) -> dict:
+    rep = sk.study(name, p)
+    print(sk.fmt(rep))
+    caus = sk.causality(p, "BTC")
+    causE = sk.causality(p, "ETH")
+    caus_ok = bool(caus["ok"] and causE["ok"])
+    mg = sk.marginal(p)
+    v = rep["verdict"]
+    maxdd = max(rep["per_asset"][a]["full"]["maxdd"] for a in rep["per_asset"])
+    w25 = mg.get("blends", {}).get("w25", {})
+    w50 = mg.get("blends", {}).get("w50", {})
+    earns_slot = (v["grade"] != "FAIL" and mg.get("marginal_verdict") == "ADDS"
+                  and bool(mg.get("robust_oos")) and not bool(mg.get("is_hedge")))
+    beats = (earns_slot and maxdd < 0.30
+             and (w25.get("uplift_hold") or -9) >= 0.55
+             and v["min_asset_holdout_sharpe"] >= 0.65)
+    print(f"  causality BTC={caus} ETH={causE} -> ok={caus_ok}")
+    print(f"  marginal: verdict={mg.get('marginal_verdict')} corr_full={mg.get('corr_full')} "
+          f"corr_hold={mg.get('corr_hold')} insample_edge={mg.get('has_insample_edge')} "
+          f"hedge={mg.get('is_hedge')} robust_oos={mg.get('robust_oos')} "
+          f"multicut={mg.get('multicut_persistent')} cleanYr={mg.get('clean_year_uplift')}")
+    print(f"  blends: w25 uplift_hold={w25.get('uplift_hold')} uplift_full={w25.get('uplift_full')} | "
+          f"w50 full={w50.get('full')} hold={w50.get('hold')} dd={w50.get('dd')} "
+          f"uplift_hold={w50.get('uplift_hold')}")
+    print(f"  ==> maxDD={maxdd*100:.1f}% earns_slot={earns_slot} BEATS_WINNER={beats}")
+    return dict(name=name, p=p, rep=rep, mg=mg, caus_ok=caus_ok, maxdd=maxdd,
+                earns_slot=earns_slot, beats=beats, w25=w25, w50=w50, v=v)
+
+
+if __name__ == "__main__":
+    print("="*100)
+    print("PHASE 1 — fast screen of cadence / holding / atr-window variants (FULL+HOLD+DD)")
+    print("="*100)
+
+    screens = []
+    # 0) reproduce the winner as a sanity anchor
+    screens.append(quick("WINNER(uL24/uS16,mpd1,atr14/14)", mk()))
+
+    # --- Holding-window grid (the core DD lever): shorter holds cap single-trade risk.
+    print("\n-- holding windows (uscitalong/uscitashort), mpd=1 --")
+    for uL in (12, 16, 20, 24, 28, 30):
+        for uS in (10, 12, 14, 16, 20):
+            screens.append(quick(f"uL{uL}/uS{uS}", mk(uscitalong=uL, uscitashort=uS)))
+
+    # --- max_per_day: 2 entries/day = more frequent re-entry (more fee, smaller clusters?)
+    print("\n-- max_per_day=2 across a few holds --")
+    for uL in (12, 16, 20, 24):
+        for uS in (10, 12, 16):
+            screens.append(quick(f"mpd2 uL{uL}/uS{uS}", mk(max_per_day=2, uscitalong=uL, uscitashort=uS)))
+
+    # --- atr windows (HTF signal vola & exec stop sizing), at the WINNER hold (uL24/uS16)
+    #     where DD was lowest, not the whipsaw uL16/uS12.
+    print("\n-- atr_win (HTF) x ltf_atr_win (exec), at WINNER hold uL24/uS16 --")
+    for aw in (10, 14, 20):
+        for lw in (10, 14, 20):
+            screens.append(quick(f"atr{aw}/ltf{lw} uL24/uS16", mk(atr_win=aw, ltf_atr_win=lw)))
+
+    # --- targeted DD-reducers: mpd2 at the winner hold (smaller clusters, keep hold) +
+    #     longer ATR for steadier stops; and asymmetric long-bias holds (long crypto = up-drift,
+    #     so a longer long-hold + shorter short-hold protects the worst-asset short DD).
+    print("\n-- targeted DD-reducers (mpd2 @ winner hold; long-bias asym holds) --")
+    for cfg in (
+        dict(max_per_day=2, uscitalong=24, uscitashort=16),
+        dict(max_per_day=2, uscitalong=24, uscitashort=16, ltf_atr_win=20),
+        dict(max_per_day=2, uscitalong=24, uscitashort=16, atr_win=20, ltf_atr_win=20),
+        dict(uscitalong=24, uscitashort=18, ltf_atr_win=20),
+        dict(uscitalong=28, uscitashort=18, ltf_atr_win=20),
+        dict(uscitalong=24, uscitashort=20, ltf_atr_win=20),
+        dict(max_per_day=2, uscitalong=20, uscitashort=16, ltf_atr_win=20),
+        dict(max_per_day=2, uscitalong=20, uscitashort=18, ltf_atr_win=20),
+    ):
+        nm = "DDr " + "/".join(f"{k}={v}" for k, v in cfg.items())
+        screens.append(quick(nm, mk(**cfg)))
+
+    # Rank by: meets DD<30 first, then by minH (we need hold >=0.65), then minF.
+    print("\n" + "="*100)
+    print("PHASE 2 — full eval of best DD-reducing candidates (study+causality+marginal)")
+    print("="*100)
+
+    # candidates: prioritize LOWEST DD with a still-usable hold-out (minH>=0.55), but ALSO
+    # always include the global lowest-DD configs that keep minH>=0.5 (DD is the unmet goal).
+    pool = [s for s in screens if s["minTr"] >= 20]
+    a_set = [s for s in pool if s["maxdd"] < 0.36 and s["minH"] >= 0.55]
+    a_set.sort(key=lambda s: (s["maxdd"], -s["minH"]))
+    b_set = [s for s in pool if s["minH"] >= 0.50]
+    b_set.sort(key=lambda s: s["maxdd"])  # lowest DD overall (usable hold)
+    picked = []
+    seen = set()
+    for s in a_set[:6] + b_set[:6]:
+        if s["name"] in seen:
+            continue
+        seen.add(s["name"])
+        picked.append(s)
+        if len(picked) >= 9:
+            break
+    if not picked:
+        pool.sort(key=lambda s: s["maxdd"])
+        picked = pool[:6]
+
+    print("Picked for full eval (DD<0.32, minH>=0.55, nTr>=20), sorted by DD:")
+    for s in picked:
+        print(f"   {s['name']:38s} maxDD={s['maxdd']*100:.0f}% minH={s['minH']:+.2f} minF={s['minF']:+.2f}")
+
+    results = []
+    for s in picked:
+        print("\n" + "-"*90)
+        results.append(full_eval(s["name"], s["p"]))
+
+    # also full-eval the winner as the reference
+    print("\n" + "-"*90 + "\n[REFERENCE] WINNER full eval:")
+    rwin = full_eval("WINNER", mk())
+
+    # ---- pick the best config: prefer beats_winner, else lowest DD with earns_slot & best hold
+    print("\n" + "="*100)
+    print("FINAL RANKING")
+    print("="*100)
+    def score(r):
+        return (not r["beats"], not r["earns_slot"], r["maxdd"], -r["v"]["min_asset_holdout_sharpe"])
+    allr = results + [rwin]
+    allr.sort(key=score)
+    for r in allr:
+        print(f"  {r['name']:38s} beats={r['beats']} earns={r['earns_slot']} maxDD={r['maxdd']*100:.0f}% "
+              f"minF={r['v']['min_asset_full_sharpe']:+.2f} minH={r['v']['min_asset_holdout_sharpe']:+.2f} "
+              f"w25uH={r['w25'].get('uplift_hold')} caus={r['caus_ok']}")
+
+    best = allr[0]
+    print("\n" + "="*100)
+    print("BEST CONFIG")
+    print("="*100)
+    bp = best["p"]
+    cfg = {k: getattr(bp, k) for k in bp.__dataclass_fields__}
+    print(f"name={best['name']}")
+    print(f"config={json.dumps(cfg)}")
+    print(f"minFull={best['v']['min_asset_full_sharpe']:+.3f}")
+    print(f"minHold={best['v']['min_asset_holdout_sharpe']:+.3f}")
+    print(f"max_dd={best['maxdd']:.4f}")
+    print(f"n_trades_min={best['v']['min_trades']}")
+    print(f"fee_survives_0.30%={best['v']['fee_survives']}")
+    print(f"causality_ok={best['caus_ok']}")
+    mg = best["mg"]
+    print(f"MARGINAL DICT: corr_full={mg.get('corr_full')} corr_hold={mg.get('corr_hold')} "
+          f"verdict={mg.get('marginal_verdict')} has_insample_edge={mg.get('has_insample_edge')} "
+          f"is_hedge={mg.get('is_hedge')} robust_oos={mg.get('robust_oos')} "
+          f"multicut_persistent={mg.get('multicut_persistent')} clean_year_uplift={mg.get('clean_year_uplift')}")
+    print(f"blend w25 uplift_hold={best['w25'].get('uplift_hold')} | "
+          f"w50 full={best['w50'].get('full')} hold={best['w50'].get('hold')} dd={best['w50'].get('dd')}")
+    print(f"earns_slot={best['earns_slot']}  BEATS_WINNER={best['beats']}")
+    # dump machine-readable for final structured output
+    print("\nJSON_BEST=" + json.dumps(dict(
+        name=best["name"], config=cfg, minFull=best["v"]["min_asset_full_sharpe"],
+        minHold=best["v"]["min_asset_holdout_sharpe"], max_dd=best["maxdd"],
+        n_trades_min=best["v"]["min_trades"], fee_survives=best["v"]["fee_survives"],
+        causality_ok=best["caus_ok"], earns_slot=best["earns_slot"], beats=best["beats"],
+        corr_full=mg.get("corr_full"), marginal_verdict=mg.get("marginal_verdict"),
+        has_insample_edge=mg.get("has_insample_edge"), is_hedge=mg.get("is_hedge"),
+        robust_oos=mg.get("robust_oos"), multicut_persistent=mg.get("multicut_persistent"),
+        clean_year_uplift=mg.get("clean_year_uplift"),
+        w25_uplift_hold=best["w25"].get("uplift_hold"))))

scripts/research/skyhook/runs/SKH2_KELTNER_PTN.py

@@ -0,0 +1,277 @@
+"""SKH2_KELTNER_PTN — KELTNER/ATR-channel breakout pattern (replaces Donchian).
+
+FAMILY: KELTNER_PTN. Goal of this wave = CUT standalone maxDD below 30% while keeping
+hold-out Sharpe high and earns_slot True.
+
+Idea: the V1/V2 Skyhook pattern is a Donchian breakout (close > rolling-high of n bars).
+Donchian highs/lows are driven by single wicks -> a fast spike can set a fresh extreme that
+the next close pokes through, firing a false breakout that mean-reverts -> drawdown. An
+ATR-CHANNEL (Keltner) breakout instead requires close to clear EMA(n) +/- k*ATR(n), a
+SMOOTHED reference that ignores isolated wicks. Steadier reference -> fewer wick-driven false
+entries -> potentially lower DD for similar exposure.
+
+We keep EVERYTHING ELSE identical to the verified V2 winner (regime Chande01 bands
+vola_lo=35/vola_hi=95/vol_lo=0, exits sl_atr=2.5/tp_atr=7.0/uscitalong=24/uscitashort=16) and
+ONLY swap the pattern from Donchian to Keltner. We do this by monkeypatching S.htf_features
+inside skyhooklib's namespace (same safe technique as SKH_R_EXPAND_study.py) so sk.study /
+sk.causality / sk.marginal run the EXACT honest machinery unchanged.
+
+CAUSALITY: EMA and ATR are causal ewm (use x[0..i] inclusive of the current, already-closed
+HTF bar); the channel for breakout-comparison is shift(1) (strictly prior bar's channel) so
+close[i] is compared against a band known BEFORE bar i closes -> leak-free. We verify with
+sk.causality (truncated-prefix guard) on BOTH assets.
+"""
+from __future__ import annotations
+
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+
+import numpy as np
+import pandas as pd
+
+import skyhooklib as sk
+from src.strategies import skyhook as S
+from src.strategies.skyhook import SkyhookParams, HTF_MIN
+
+ORIG_FEAT = S.htf_features
+FEE = sk.FEE_RT
+
+
+# ---------------------------------------------------------------------------
+# Keltner channel breakout on HTF (causal, shift-safe).
+#   mid   = EMA(close, n)
+#   width = k * ATR(n)        (ATR over the same n window, ewm)
+#   upper = mid + width ; lower = mid - width
+#   ptn_long  = close[i] > upper[i-1]   (clears the PRIOR bar's upper channel)
+#   ptn_short = close[i] < lower[i-1]
+# The shift(1) on the channel makes the comparison strictly causal: the band the close must
+# clear is fully determined by bars <= i-1 (Donchian uses shift(1) on the rolling extreme for
+# the same reason). EMA/ATR ewm themselves use only past+current data.
+# ---------------------------------------------------------------------------
+def keltner_breakout(htf: pd.DataFrame, n: int, k: float, atr_win: int) -> tuple[np.ndarray, np.ndarray]:
+    c = htf["close"].values.astype(float)
+    mid = pd.Series(c).ewm(span=n, adjust=False, min_periods=n).mean().values
+    a = S.atr(htf, atr_win)
+    upper = mid + k * a
+    lower = mid - k * a
+    # compare current close vs the PRIOR bar's channel (shift 1) -> strictly causal
+    upper_prev = pd.Series(upper).shift(1).values
+    lower_prev = pd.Series(lower).shift(1).values
+    ptn_long = np.where(np.isfinite(upper_prev), c > upper_prev, False)
+    ptn_short = np.where(np.isfinite(lower_prev), c < lower_prev, False)
+    return ptn_long.astype(bool), ptn_short.astype(bool)
+
+
+def make_keltner_features(n: int, k: float, kelt_atr_win: int):
+    """Return an htf_features replacement: V1 Chande01 regime + Keltner pattern."""
+    def _feat(htf: pd.DataFrame, p: SkyhookParams) -> pd.DataFrame:
+        buz_vola = S.chande01(S.atr(htf, p.atr_win), p.n_vola)
+        buz_volume = S.chande01(htf["volume"].values, p.n_volume)
+        ptn_long, ptn_short = keltner_breakout(htf, n, k, kelt_atr_win)
+        regime_ok = ((buz_vola >= p.vola_lo) & (buz_vola <= p.vola_hi)
+                     & (buz_volume >= p.vol_lo) & (buz_volume <= p.vol_hi))
+        comp_long = regime_ok & ptn_long
+        comp_short = regime_ok & ptn_short
+        if p.long_only:
+            comp_short = np.zeros_like(comp_short, dtype=bool)
+        close_ts = htf["timestamp"].astype("int64").values + HTF_MIN * 60 * 1000
+        return pd.DataFrame({
+            "close_ts": close_ts,
+            "buz_vola": buz_vola, "buz_volume": buz_volume,
+            "comp_long": comp_long.astype(bool), "comp_short": comp_short.astype(bool),
+        })
+    return _feat
+
+
+def study_keltner(name, p, n, k, kelt_atr_win):
+    """sk.study + causality + marginal with htf_features patched to Keltner."""
+    S.htf_features = make_keltner_features(n, k, kelt_atr_win)
+    try:
+        rep = sk.study(name, p)
+        caus = sk.causality(p, "BTC")
+        caus_eth = sk.causality(p, "ETH")
+        marg = sk.marginal(p)
+    finally:
+        S.htf_features = ORIG_FEAT
+    return rep, (caus, caus_eth), marg
+
+
+def earns_slot(rep, marg):
+    grade_ok = rep["verdict"]["grade"] != "FAIL"
+    return bool(grade_ok and marg.get("marginal_verdict") == "ADDS"
+                and marg.get("robust_oos") is True and marg.get("is_hedge") is False)
+
+
+if __name__ == "__main__":
+    # Winner exits/regime (the verified V2 winner) — only the pattern changes to Keltner.
+    p = SkyhookParams(sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16,
+                      vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+
+    # n in {13,20,34}, k in {1.0,1.5,2.0}; ATR window for the channel width = winner default 14.
+    grid = []
+    for n in (13, 20, 34):
+        for k in (1.0, 1.5, 2.0):
+            grid.append((n, k))
+
+    KELT_ATR = 14
+    print("=== SKH2_KELTNER_PTN: ATR-channel (Keltner) breakout sweep (regime+exits = V2 winner) ===\n")
+    print(f"grid n x k = {grid}  kelt_atr_win={KELT_ATR}\n")
+
+    # ---- Sweep (cheap pass: FULL/HOLD/DD/trades + fee survival via study) ----
+    rows = []
+    for (n, k) in grid:
+        tag = f"KELT_n{n}_k{k}"
+        rep, (cb, ce), marg = study_keltner(tag, p, n, k, KELT_ATR)
+        v = rep["verdict"]
+        # standalone DD = max over BTC&ETH FULL maxdd
+        dd = max(rep["per_asset"][a]["full"]["maxdd"] for a in rep["per_asset"])
+        es = earns_slot(rep, marg)
+        w25 = marg.get("blends", {}).get("w25", {}).get("uplift_hold")
+        rows.append(dict(tag=tag, n=n, k=k, rep=rep, caus=(cb, ce), marg=marg,
+                         minFull=v["min_asset_full_sharpe"], minHold=v["min_asset_holdout_sharpe"],
+                         minTr=v["min_trades"], feeOK=v["fee_survives"], grade=v["grade"],
+                         dd=dd, es=es, w25=w25, caus_ok=(cb["ok"] and ce["ok"])))
+        beats = (es and dd < 0.30 and (w25 is not None and w25 >= 0.55) and v["min_asset_holdout_sharpe"] >= 0.65)
+        print(f"  [{tag}] grade={v['grade']} minFull={v['min_asset_full_sharpe']:+.2f}"
+              f" minHold={v['min_asset_holdout_sharpe']:+.2f} minTr={v['min_trades']}"
+              f" DD={dd*100:.0f}% feeOK={v['fee_survives']} caus={cb['ok'] and ce['ok']}"
+              f" | verdict={marg.get('marginal_verdict')} corr={marg.get('corr_full')}"
+              f" w25={w25} robust={marg.get('robust_oos')} hedge={marg.get('is_hedge')}"
+              f" earns_slot={es} BEATS={beats}")
+
+    # =======================================================================
+    # REFINEMENT PASS: the plain swap keeps DD>30% (too many entries / wick pokes).
+    # DD is driven by (a) wide vola band letting in blow-off breakouts, (b) loose SL,
+    # (c) shorts bleeding in a structural bull. Sweep regime-tightening + SL + long_only
+    # around the best earns_slot region (n13/n20, k1.5-2.0) to push DD under 30%.
+    # =======================================================================
+    print("\n--- REFINEMENT: tighten regime / SL / long_only to cut DD<30% ---")
+    refine = [
+        # (n, k, sl_atr, tp_atr, vola_lo, vola_hi, vol_lo, long_only, tag)
+        (13, 2.0, 2.0, 7.0, 35.0, 95.0, 0.0, False, "n13k2_sl2.0"),
+        (13, 2.0, 2.5, 7.0, 45.0, 90.0, 0.0, False, "n13k2_vola45-90"),
+        (13, 2.0, 2.5, 7.0, 35.0, 85.0, 0.0, False, "n13k2_volaHi85"),
+        (13, 2.0, 2.5, 7.0, 35.0, 95.0, 40.0, False, "n13k2_volFloor40"),
+        (13, 2.0, 2.5, 7.0, 35.0, 95.0, 0.0, True, "n13k2_longOnly"),
+        (13, 2.0, 2.0, 7.0, 45.0, 90.0, 40.0, False, "n13k2_tight_all"),
+        (20, 2.0, 2.0, 7.0, 35.0, 95.0, 0.0, False, "n20k2_sl2.0"),
+        (20, 2.0, 2.5, 7.0, 45.0, 90.0, 40.0, False, "n20k2_tight_all"),
+        (13, 2.5, 2.5, 7.0, 35.0, 95.0, 0.0, False, "n13k2.5"),
+        (13, 2.5, 2.0, 7.0, 45.0, 90.0, 0.0, False, "n13k2.5_sl2_vola45-90"),
+        # ---- pass 3: sl2.0 was the DD/hold winner; push SL tighter + lower TP (cut tail) ----
+        (13, 2.0, 1.5, 6.0, 35.0, 95.0, 0.0, False, "n13k2_sl1.5_tp6"),
+        (13, 2.0, 1.5, 7.0, 35.0, 95.0, 40.0, False, "n13k2_sl1.5_volFloor40"),
+        (13, 2.0, 2.0, 5.0, 35.0, 95.0, 0.0, False, "n13k2_sl2_tp5"),
+        (13, 2.0, 2.0, 6.0, 35.0, 95.0, 40.0, False, "n13k2_sl2_tp6_volFloor40"),
+        (13, 2.0, 1.5, 5.0, 35.0, 95.0, 0.0, False, "n13k2_sl1.5_tp5"),
+        (13, 1.5, 2.0, 7.0, 35.0, 95.0, 0.0, False, "n13k1.5_sl2.0"),
+    ]
+    for (n, k, sl, tp, vlo, vhi, vol_lo, lo, tag) in refine:
+        pr = SkyhookParams(sl_atr=sl, tp_atr=tp, uscitalong=24, uscitashort=16,
+                           vola_lo=vlo, vola_hi=vhi, vol_lo=vol_lo, long_only=lo)
+        rep, (cb, ce), marg = study_keltner(tag, pr, n, k, KELT_ATR)
+        v = rep["verdict"]
+        dd = max(rep["per_asset"][a]["full"]["maxdd"] for a in rep["per_asset"])
+        es = earns_slot(rep, marg)
+        w25 = marg.get("blends", {}).get("w25", {}).get("uplift_hold")
+        rows.append(dict(tag=tag, n=n, k=k, sl=sl, tp=tp, vlo=vlo, vhi=vhi, vol_lo=vol_lo, lo=lo,
+                         rep=rep, caus=(cb, ce), marg=marg,
+                         minFull=v["min_asset_full_sharpe"], minHold=v["min_asset_holdout_sharpe"],
+                         minTr=v["min_trades"], feeOK=v["fee_survives"], grade=v["grade"],
+                         dd=dd, es=es, w25=w25, caus_ok=(cb["ok"] and ce["ok"])))
+        b2 = (es and dd < 0.30 and (w25 is not None and w25 >= 0.55) and v["min_asset_holdout_sharpe"] >= 0.65)
+        print(f"  [{tag}] grade={v['grade']} minFull={v['min_asset_full_sharpe']:+.2f}"
+              f" minHold={v['min_asset_holdout_sharpe']:+.2f} minTr={v['min_trades']}"
+              f" DD={dd*100:.0f}% feeOK={v['fee_survives']} caus={cb['ok'] and ce['ok']}"
+              f" | verdict={marg.get('marginal_verdict')} w25={w25} robust={marg.get('robust_oos')}"
+              f" hedge={marg.get('is_hedge')} earns_slot={es} BEATS={b2}")
+
+    # ---- Pick best: prefer DD<30% with earns_slot, then by (minHold, then w25) ----
+    def beats_winner(r):
+        return bool(r["es"] and r["dd"] < 0.30
+                    and (r["w25"] is not None and r["w25"] >= 0.55)
+                    and r["minHold"] >= 0.65)
+
+    winners = [r for r in rows if beats_winner(r)]
+    if winners:
+        best = max(winners, key=lambda r: (r["minHold"], r["w25"] or -9))
+        pool = "BEATS-WINNER"
+    else:
+        # objective priority: DD<30 + earns_slot first; else best DD among earns_slot;
+        # else best DD among fee-surviving non-FAIL; else lowest DD overall.
+        cand1 = [r for r in rows if r["dd"] < 0.30 and r["es"]]
+        # secondary-quality: earns_slot AND meets the two NON-DD beats gates (w25>=0.55, minHold>=0.65)
+        candQ = [r for r in rows if r["es"] and (r["w25"] is not None and r["w25"] >= 0.55)
+                 and r["minHold"] >= 0.65]
+        cand2 = [r for r in rows if r["es"]]
+        cand3 = [r for r in rows if r["grade"] != "FAIL" and r["feeOK"]]
+        if cand1:
+            best = max(cand1, key=lambda r: (r["minHold"], r["w25"] or -9)); pool = "DD<30+earns_slot"
+        elif candQ:
+            # best DD among configs that already clear the other two beats gates
+            best = min(candQ, key=lambda r: r["dd"]); pool = "earns_slot+w25>=.55+minHold>=.65 (DD>=30)"
+        elif cand2:
+            best = min(cand2, key=lambda r: r["dd"]); pool = "earns_slot (DD>=30)"
+        elif cand3:
+            best = min(cand3, key=lambda r: r["dd"]); pool = "fee-surviving non-FAIL"
+        else:
+            best = min(rows, key=lambda r: r["dd"]); pool = "lowest-DD overall"
+
+    rep, marg = best["rep"], best["marg"]
+    cb, ce = best["caus"]
+    v = rep["verdict"]
+    bl = marg.get("blends", {})
+    w25 = bl.get("w25", {})
+    w50 = bl.get("w50", {})
+
+    print("\n" + "=" * 78)
+    print(f"BEST CONFIG ({pool}): {best['tag']}  (n={best['n']}, k={best['k']}, kelt_atr_win={KELT_ATR})")
+    print("=" * 78)
+    print(sk.fmt(rep))
+    print(f"\nstandalone max_dd (max BTC&ETH FULL) = {best['dd']:.4f}  ({best['dd']*100:.1f}%)")
+    print(f"causality BTC={cb}  ETH={ce}  -> ok={cb['ok'] and ce['ok']}")
+    print(f"minFull={v['min_asset_full_sharpe']:+.3f}  minHold={v['min_asset_holdout_sharpe']:+.3f}"
+          f"  minTrades={v['min_trades']}  fee_survives_0.30%={v['fee_survives']}")
+    print("\n--- MARGINAL vs TP01 ---")
+    print(f"  marginal_verdict   = {marg.get('marginal_verdict')}")
+    print(f"  corr_full          = {marg.get('corr_full')}")
+    print(f"  corr_hold          = {marg.get('corr_hold')}")
+    print(f"  has_insample_edge  = {marg.get('has_insample_edge')}")
+    print(f"  is_hedge           = {marg.get('is_hedge')}")
+    print(f"  robust_oos         = {marg.get('robust_oos')}")
+    print(f"  multicut_persistent= {marg.get('multicut_persistent')}")
+    print(f"  clean_year_uplift  = {marg.get('clean_year_uplift')}")
+    print(f"  jackknife_min_uplift= {marg.get('jackknife_min_uplift')}")
+    print(f"  multicut_uplift    = {marg.get('multicut_uplift')}")
+    print(f"  cand_insample_sharpe= {marg.get('cand_insample_sharpe')}")
+    print(f"  blend w25 uplift_hold = {w25.get('uplift_hold')}  uplift_full={w25.get('uplift_full')}")
+    print(f"  blend w50          = full={w50.get('full')} hold={w50.get('hold')}"
+          f" uplift_hold={w50.get('uplift_hold')} dd={w50.get('dd')}")
+
+    es = best["es"]
+    beats = beats_winner(best)
+    print(f"\n  earns_slot   = {es}")
+    print(f"  BEATS_WINNER = {beats}  "
+          f"(need: earns_slot AND max_dd<0.30 AND w25_uplift_hold>=0.55 AND minHold>=0.65)")
+
+    # ---- machine-readable final line for the orchestrator/agent to parse ----
+    import json
+    out = dict(
+        family="KELTNER_PTN", tag=best["tag"],
+        best_config=dict(ptn_kind="keltner", n=best["n"], k=best["k"], kelt_atr_win=KELT_ATR,
+                         sl_atr=best.get("sl", 2.5), tp_atr=best.get("tp", 7.0),
+                         uscitalong=24, uscitashort=16,
+                         vola_lo=best.get("vlo", 35.0), vola_hi=best.get("vhi", 95.0),
+                         vol_lo=best.get("vol_lo", 0.0), long_only=best.get("lo", False)),
+        min_full_sharpe=v["min_asset_full_sharpe"], min_hold_sharpe=v["min_asset_holdout_sharpe"],
+        max_dd=best["dd"], n_trades_min=v["min_trades"], fee_survives_030=bool(v["fee_survives"]),
+        causality_ok=bool(cb["ok"] and ce["ok"]),
+        marginal_verdict=marg.get("marginal_verdict"),
+        has_insample_edge=bool(marg.get("has_insample_edge")),
+        is_hedge=bool(marg.get("is_hedge")), robust_oos=bool(marg.get("robust_oos")),
+        multicut_persistent=bool(marg.get("multicut_persistent")),
+        clean_year_uplift=marg.get("clean_year_uplift"), corr_full=marg.get("corr_full"),
+        blend_w25_uplift_hold=w25.get("uplift_hold"),
+        earns_slot=bool(es), beats_winner=bool(beats),
+    )
+    print("\nFINAL_JSON=" + json.dumps(out, default=str))

scripts/research/skyhook/runs/SKH2_PATTERN_CONF.py

@@ -0,0 +1,298 @@
+"""SKH2_PATTERN_CONF — breakout CONFIRMATION filter family (DD-reduction wave).
+
+GOAL of the wave: cut standalone maxDD < 30% (max over BTC&ETH) while keeping
+min-asset HOLD-OUT Sharpe >= ~0.70 and earns_slot == True.
+
+FAMILY = breakout confirmation. The main DD source is FALSE breakouts (whipsaws).
+We require CONFIRMATION before allowing the composer to fire, via a STRUCTURAL
+htf_features patch (causal, shift-safe). Confirmation modes (all use data <= close[i]):
+
+  persist2  : the breakout must PERSIST -> the *previous* HTF close also broke the
+              donchian level that was active one bar earlier (2 consecutive breakouts).
+  close_loc : the breakout close must sit in the upper/lower `loc_thr` of the HTF
+              bar range (close near the high for a long, near the low for a short)
+              -> rejects exhaustion wicks that close back inside the bar.
+  roc_agree : HTF ROC (close/close[-roc_n]-1) sign must agree with the breakout dir.
+  combos    : AND-combinations of the above.
+
+We monkeypatch S.htf_features INSIDE skyhooklib's namespace for the duration of each
+study (same safe pattern as SKH_R_EXPAND_study.py): only the feature/composer builder
+changes; pattern donchian, regime bands, entry/exit and ALL eval code are unchanged.
+
+Baseline regime/exit params = the verified V2 WINNER:
+  SkyhookParams(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16,
+                vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+
+CAUSALITY: every confirmation feature is built from HTF columns and SHIFTED so that
+only data up to (and including) the breakout-bar close is used; donchian itself is
+shift(1) (strictly prior bars). We verify with sk.causality (truncated-prefix) which
+re-runs skyhook_entries on a prefix of BOTH frames -> our patched htf_features is
+exercised on the prefix and must match the full run.
+"""
+from __future__ import annotations
+
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+
+import numpy as np
+import pandas as pd
+
+import skyhooklib as sk
+from src.strategies import skyhook as S
+from src.strategies.skyhook import SkyhookParams, HTF_MIN
+from src.strategies.skyhook import atr as _atr, chande01 as _chande01
+
+ORIG_FEAT = S.htf_features
+
+
+# ---------------------------------------------------------------------------
+# Confirmed htf_features builder (STRUCTURAL). Same shape/columns as the engine's
+# htf_features, but the composer's pattern leg is gated by a CONFIRMATION mask.
+# ---------------------------------------------------------------------------
+def conf_htf_features(htf: pd.DataFrame, p: SkyhookParams, *,
+                      modes=("persist2",), loc_thr: float = 0.34, roc_n: int = 1):
+    """Causal regime+CONFIRMED-pattern features indexed by HTF close.
+
+    modes: subset of {"persist2","close_loc","roc_agree"} ANDed together as the
+    confirmation requirement on top of the raw donchian breakout.
+    """
+    h = htf["high"].values.astype(float)
+    l = htf["low"].values.astype(float)
+    c = htf["close"].values.astype(float)
+    n = len(c)
+
+    # --- regime (unchanged) ---
+    buz_vola = _chande01(_atr(htf, p.atr_win), p.n_vola)
+    buz_volume = _chande01(htf["volume"].values, p.n_volume)
+    regime_ok = ((buz_vola >= p.vola_lo) & (buz_vola <= p.vola_hi)
+                 & (buz_volume >= p.vol_lo) & (buz_volume <= p.vol_hi))
+
+    # --- raw donchian breakout (leak-free: close[i] vs max/min of n PRIOR bars) ---
+    hi_n = pd.Series(h).rolling(p.ptn_n, min_periods=p.ptn_n).max().shift(1).values
+    lo_n = pd.Series(l).rolling(p.ptn_n, min_periods=p.ptn_n).min().shift(1).values
+    ptn_long = c > hi_n
+    ptn_short = c < lo_n
+
+    # --- CONFIRMATION masks (all causal: built from data <= close[i]) ---
+    conf_long = np.ones(n, dtype=bool)
+    conf_short = np.ones(n, dtype=bool)
+
+    if "persist2" in modes:
+        # previous bar's close also broke the donchian level active ONE bar earlier.
+        # ptn_long shifted by 1 == "did the prior bar break out?" (its own causal level).
+        prev_long = np.concatenate(([False], ptn_long[:-1]))
+        prev_short = np.concatenate(([False], ptn_short[:-1]))
+        conf_long &= prev_long
+        conf_short &= prev_short
+
+    if "close_loc" in modes:
+        rng = h - l
+        with np.errstate(divide="ignore", invalid="ignore"):
+            pos = np.where(rng > 0, (c - l) / rng, 0.5)   # 0=at low, 1=at high; current bar only
+        conf_long &= (pos >= (1.0 - loc_thr))
+        conf_short &= (pos <= loc_thr)
+
+    if "roc_agree" in modes:
+        cprev = pd.Series(c).shift(roc_n).values            # close roc_n bars ago (causal)
+        with np.errstate(divide="ignore", invalid="ignore"):
+            roc = np.where(np.isfinite(cprev) & (cprev != 0), c / cprev - 1.0, 0.0)
+        conf_long &= (roc > 0.0)
+        conf_short &= (roc < 0.0)
+
+    comp_long = regime_ok & ptn_long & conf_long
+    comp_short = regime_ok & ptn_short & conf_short
+    if p.long_only:
+        comp_short = np.zeros_like(comp_short, dtype=bool)
+
+    close_ts = htf["timestamp"].astype("int64").values + HTF_MIN * 60 * 1000
+    return pd.DataFrame({
+        "close_ts": close_ts,
+        "buz_vola": buz_vola, "buz_volume": buz_volume,
+        "comp_long": comp_long.astype(bool), "comp_short": comp_short.astype(bool),
+    })
+
+
+def make_patched(**cfg):
+    def _feat(htf, p):
+        return conf_htf_features(htf, p, **cfg)
+    return _feat
+
+
+def study_conf(name, p, cfg, do_marginal=True):
+    """Run sk.study/causality/marginal with htf_features patched to the confirmed builder."""
+    S.htf_features = make_patched(**cfg)
+    try:
+        rep = sk.study(name, p)
+        caus_btc = sk.causality(p, "BTC")
+        caus_eth = sk.causality(p, "ETH")
+        marg = sk.marginal(p) if do_marginal else None
+    finally:
+        S.htf_features = ORIG_FEAT
+    return rep, (caus_btc, caus_eth), marg
+
+
+def _earns_slot(rep, marg):
+    grade_ok = rep["verdict"]["grade"] != "FAIL"
+    adds = marg.get("marginal_verdict") == "ADDS"
+    robust = bool(marg.get("robust_oos"))
+    hedge = bool(marg.get("is_hedge"))
+    return bool(grade_ok and adds and robust and (not hedge))
+
+
+def _beats_winner(rep, marg, max_dd):
+    es = _earns_slot(rep, marg)
+    w25 = marg.get("blends", {}).get("w25", {}).get("uplift_hold")
+    mh = rep["verdict"]["min_asset_holdout_sharpe"]
+    return bool(es and max_dd < 0.30 and (w25 is not None and w25 >= 0.55) and mh >= 0.65)
+
+
+# Baseline regime/exit = verified V2 winner
+WIN = dict(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16,
+           vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+
+
+def win_params():
+    return SkyhookParams(**WIN)
+
+
+def win_params_ex(**over):
+    d = dict(WIN); d.update(over); return SkyhookParams(**d)
+
+
+if __name__ == "__main__":
+    p = win_params()
+
+    # ---- PHASE 1: scan confirmation modes (quick, no marginal yet) ----
+    # winner's exit = sl_atr 2.5 / tp_atr 7.0. close_loc+roc was the clear leader (DD 31.4%,
+    # minHold +1.13). Now drive DD<30% via STRONGER close-location confirmation (tighter
+    # loc_thr) and TIGHTER stops (sl_atr down), and an upper-vola cap to skip blow-offs.
+    CL = ("close_loc",)   # roc_agree is ~collinear with close_loc (no-op); drop it -> cleaner
+    C = dict(modes=CL, loc_thr=0.40)
+    scan = {
+        "RAW (no conf, =winner)": (p, dict(modes=())),
+        "cl 0.40 vH90": (win_params_ex(vola_hi=90.0), C),
+        "cl 0.40 vH88": (win_params_ex(vola_hi=88.0), C),
+        "cl 0.40 vH85": (win_params_ex(vola_hi=85.0), C),
+        # volume floor: skip thin (low-volume-cycle) breakouts -> fewer false ETH whipsaws
+        "cl 0.40 vH90 volLo20": (win_params_ex(vola_hi=90.0, vol_lo=20.0), C),
+        "cl 0.40 vH90 volLo35": (win_params_ex(vola_hi=90.0, vol_lo=35.0), C),
+        # raise vola_lo floor (skip dead-vol regimes) + cap top
+        "cl 0.40 vL45 vH90": (win_params_ex(vola_lo=45.0, vola_hi=90.0), C),
+        # tighten long hold to cut give-back on the trend reversals
+        "cl 0.40 vH90 uL20": (win_params_ex(vola_hi=90.0, uscitalong=20), C),
+        "cl 0.40 vH90 uL20 uS14": (win_params_ex(vola_hi=90.0, uscitalong=20, uscitashort=14), C),
+        # tp tighter to bank wins sooner (less round-trip give-back) — keeps DD lower
+        "cl 0.40 vH90 tp6": (win_params_ex(vola_hi=90.0, tp_atr=6.0), C),
+        "cl 0.40 vH90 tp6 uL20": (win_params_ex(vola_hi=90.0, tp_atr=6.0, uscitalong=20), C),
+        "cl 0.40 vH90 volLo20 uL20": (win_params_ex(vola_hi=90.0, vol_lo=20.0, uscitalong=20), C),
+    }
+    print("=" * 78)
+    print("PHASE 1 SCAN — confirmation modes on the V2 winner (DD-focus)")
+    print("=" * 78)
+    rows = []
+    for name, (pp, cfg) in scan.items():
+        rep, caus, _ = study_conf(name, pp, cfg, do_marginal=False)
+        v = rep["verdict"]
+        mdd = max(rep["per_asset"][a]["full"]["maxdd"] for a in rep["per_asset"])
+        cb = caus[0]["ok"] and caus[1]["ok"]
+        nmin = min(rep["per_asset"][a]["full"]["n_trades"] for a in rep["per_asset"])
+        rows.append((name, (pp, cfg), v["grade"], v["min_asset_full_sharpe"],
+                     v["min_asset_holdout_sharpe"], mdd, nmin, v["fee_survives"], cb))
+        print(f"  {name:30s} grade={v['grade']:4s} minFull={v['min_asset_full_sharpe']:+.2f} "
+              f"minHold={v['min_asset_holdout_sharpe']:+.2f} maxDD={mdd*100:4.0f}% "
+              f"nMin={nmin:4d} feeOK={v['fee_survives']} caus={cb}")
+
+    # pick candidates: grade != FAIL, maxDD lowest, holdout decent. Take all with DD<32% & minHold>0.4.
+    cands = [r for r in rows if r[2] != "FAIL" and r[7] and r[8]
+             and r[5] < 0.33 and r[4] >= 0.40 and r[0] != "RAW (no conf, =winner)"]
+    # sort: sub-30% DD first (the wave goal), then highest hold
+    cands.sort(key=lambda r: (r[5] >= 0.30, r[5], -r[4]))
+    print("\nPHASE 1 candidates (DD<35%, minHold>=0.40, feeOK, causal), best DD first:")
+    for r in cands:
+        print(f"  {r[0]:24s} DD={r[5]*100:.0f}% minHold={r[4]:+.2f} minFull={r[3]:+.2f}")
+
+    # ---- PHASE 2: full marginal on the top few ----
+    top = cands[:5] if cands else []
+    if not top:
+        # fall back to the lowest-DD non-FAIL configs regardless of hold threshold
+        fb = [r for r in rows if r[2] != "FAIL" and r[7] and r[8]]
+        fb.sort(key=lambda r: r[5])
+        top = fb[:3]
+
+    print("\n" + "=" * 78)
+    print("PHASE 2 — full marginal vs TP01 on top confirmation candidates")
+    print("=" * 78)
+
+    best = None  # (beats, earns, max_dd, minHold, name, cfg, rep, marg)
+    for r in top:
+        name, (pp, cfg) = r[0], r[1]
+        rep, caus, marg = study_conf(name, pp, cfg, do_marginal=True)
+        v = rep["verdict"]
+        mdd = max(rep["per_asset"][a]["full"]["maxdd"] for a in rep["per_asset"])
+        cb = caus[0]["ok"] and caus[1]["ok"]
+        es = _earns_slot(rep, marg)
+        bw = _beats_winner(rep, marg, mdd)
+        w25 = marg.get("blends", {}).get("w25", {})
+        w50 = marg.get("blends", {}).get("w50", {})
+        print(f"\n----- {name}  cfg={cfg} -----")
+        print(sk.fmt(rep))
+        print(f"causality: BTC={caus[0]} ETH={caus[1]}  -> ok={cb}")
+        print(f"marginal: verdict={marg.get('marginal_verdict')} corr_full={marg.get('corr_full')} "
+              f"corr_hold={marg.get('corr_hold')}")
+        print(f"          has_insample_edge={marg.get('has_insample_edge')} "
+              f"cand_insample_sharpe={marg.get('cand_insample_sharpe')} "
+              f"is_hedge={marg.get('is_hedge')} robust_oos={marg.get('robust_oos')} "
+              f"multicut_persistent={marg.get('multicut_persistent')}")
+        print(f"          clean_year_uplift={marg.get('clean_year_uplift')} "
+              f"jackknife_min_uplift={marg.get('jackknife_min_uplift')} "
+              f"multicut_uplift={marg.get('multicut_uplift')}")
+        print(f"          blend w25: uplift_hold={w25.get('uplift_hold')} uplift_full={w25.get('uplift_full')} "
+              f"hold={w25.get('hold')} full={w25.get('full')}")
+        print(f"          blend w50: full={w50.get('full')} hold={w50.get('hold')} "
+              f"uplift_hold={w50.get('uplift_hold')} dd={w50.get('dd')}")
+        print(f"  EARNS_SLOT={es}  BEATS_WINNER={bw}")
+        key = (bw, es, -mdd, v["min_asset_holdout_sharpe"])
+        cur = dict(beats=bw, earns=es, max_dd=mdd, minHold=v["min_asset_holdout_sharpe"],
+                   minFull=v["min_asset_full_sharpe"], name=name, cfg=cfg, rep=rep, marg=marg,
+                   base=sk._params_dict(pp),
+                   caus=cb, nmin=min(rep["per_asset"][a]["full"]["n_trades"] for a in rep["per_asset"]),
+                   feeOK=v["fee_survives"], grade=v["grade"])
+        if best is None or key > best[0]:
+            best = (key, cur)
+
+    # ---- FINAL BLOCK ----
+    print("\n" + "#" * 78)
+    print("FINAL — BEST PATTERN_CONF CONFIG")
+    print("#" * 78)
+    if best is None:
+        print("NO non-FAIL candidate found.")
+    else:
+        b = best[1]
+        m = b["marg"]
+        w25 = m.get("blends", {}).get("w25", {})
+        w50 = m.get("blends", {}).get("w50", {})
+        print(f"name        : {b['name']}")
+        print(f"cfg         : {b['cfg']}")
+        print(f"base params : {b['base']}")
+        print(f"grade       : {b['grade']}")
+        print(f"minFull     : {b['minFull']:+.3f}")
+        print(f"minHold     : {b['minHold']:+.3f}")
+        print(f"max_dd      : {b['max_dd']:.4f}  ({b['max_dd']*100:.1f}%)")
+        print(f"n_trades_min: {b['nmin']}")
+        print(f"fee@0.30%   : survives={b['feeOK']}")
+        print(f"causality_ok: {b['caus']}")
+        print(f"--- marginal dict ---")
+        print(f"  corr_full          : {m.get('corr_full')}")
+        print(f"  corr_hold          : {m.get('corr_hold')}")
+        print(f"  marginal_verdict   : {m.get('marginal_verdict')}")
+        print(f"  has_insample_edge  : {m.get('has_insample_edge')}")
+        print(f"  cand_insample_sh   : {m.get('cand_insample_sharpe')}")
+        print(f"  is_hedge           : {m.get('is_hedge')}")
+        print(f"  robust_oos         : {m.get('robust_oos')}")
+        print(f"  multicut_persistent: {m.get('multicut_persistent')}")
+        print(f"  clean_year_uplift  : {m.get('clean_year_uplift')}")
+        print(f"  blend w25 uplift_hold: {w25.get('uplift_hold')}")
+        print(f"  blend w25 uplift_full: {w25.get('uplift_full')}")
+        print(f"  blend w50          : full={w50.get('full')} hold={w50.get('hold')} dd={w50.get('dd')}")
+        print(f"EARNS_SLOT  : {b['earns']}")
+        print(f"BEATS_WINNER: {b['beats']}")

scripts/research/skyhook/runs/SKH2_PCTL_DD.py

@@ -0,0 +1,289 @@
+"""SKH2_PCTL_DD — DD-reduction wave, family [PCTL_DD].
+
+GOAL: cut STANDALONE maxDD below 30% (max over BTC & ETH) while keeping minHold>=~0.70
+and earns_slot==True, using the CAUSAL expanding/rolling PERCENTILE-RANK regime from
+SKH_R_PCTL.py (reuse pctl_entries), tuned together with the winner's exits.
+
+Baseline to beat (V2 winner, Chande regime):
+  SkyhookParams(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16,
+                vola_lo=35, vola_hi=95, vol_lo=0.0)
+  minFull +0.83, minHold +0.81, standalone DD BTC34%/ETH31% (THE PROBLEM),
+  marginal ADDS, blend w25 uplift_hold +0.58, blend 50/50 full1.59/hold1.04/DD12.5%.
+
+LEVERS FOR DD CUT (all causal, expressed through pctl_entries cfg + the SkyhookParams exits):
+  * percentile-rank regime bands (where ATR/volume sit in their own causal history):
+      - cap the upper vola band (avoid blow-off-vol entries that cluster losses)
+      - add a volume floor (live tape only) OR keep vol open
+  * tighter hard stop (sl_atr) caps per-trade loss -> shrinks DD
+  * the winner's wider tp_atr=7.0 + asym time exits (24/16) carried over.
+
+A candidate BEATS THE WINNER iff: earns_slot AND max_dd<0.30 AND blend_w25_uplift_hold>=0.55
+AND min_hold_sharpe>=0.65.  We report TRUE numbers regardless.
+"""
+from __future__ import annotations
+
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+
+import importlib.util
+import numpy as np
+import pandas as pd
+
+import skyhooklib as sk
+import altlib as al
+from src.backtest.harness import backtest_signals
+from src.strategies import skyhook as S
+from src.strategies.skyhook import SkyhookParams
+
+# import the structural pctl builder (pctl_entries, pctl_rank, _split) from the sweep script
+spec = importlib.util.spec_from_file_location(
+    "skr", "/opt/docker/PythagorasGoal/scripts/research/skyhook/runs/SKH_R_PCTL.py")
+skr = importlib.util.module_from_spec(spec)
+spec.loader.exec_module(skr)   # __main__ guard prevents the sweep from running
+
+HOLDOUT = sk.HOLDOUT
+FEE = sk.FEE_RT
+
+
+# ---------------------------------------------------------------------------
+# Build a SkyhookParams holding the WINNER's exits; only regime comes from pctl cfg.
+# pctl_entries reads: ptn_n, sl_atr, tp_atr, uscitalong, uscitashort, exit_mode, ltf_atr_win,
+#                     max_per_day, long_only  (the regime bands come from the cfg kwargs).
+# ---------------------------------------------------------------------------
+def winner_exit_params(**kw):
+    base = dict(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16)
+    base.update(kw)
+    return SkyhookParams(**base)
+
+
+# ---------------------------------------------------------------------------
+# Eval a (cfg, params) pair on both assets: FULL + HOLD via the honest engine.
+# ---------------------------------------------------------------------------
+def eval_pair(cfg, p):
+    out = {}
+    for a in ("BTC", "ETH"):
+        ltf, htf = sk.frames(a)
+        ent = skr.pctl_entries(ltf, htf, p, **cfg)
+        m = backtest_signals(ltf, ent, fee_rt=FEE, leverage=1.0, asset=a, tf="230m")
+        eq = m.equity
+        idx = pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True))
+        hmask = np.asarray(idx >= HOLDOUT)
+        full = dict(sharpe=round(m.sharpe, 3), ret=round(m.net_return, 4), maxdd=round(m.max_dd, 4),
+                    n_trades=int(m.n_trades), win_rate=round(m.win_rate, 1))
+        hold = skr._split(eq, idx, hmask)
+        out[a] = dict(full=full, hold=hold,
+                      yearly={int(y): round(v, 4) for y, v in m.yearly.items()})
+    return out
+
+
+def summarize(res):
+    mf = min(res[a]["full"]["sharpe"] for a in res)
+    mh = min(res[a]["hold"]["sharpe"] for a in res)
+    mt = min(res[a]["full"]["n_trades"] for a in res)
+    mdd = max(res[a]["full"]["maxdd"] for a in res)
+    return dict(minFull=mf, minHold=mh, minTr=mt, maxDD=mdd, res=res)
+
+
+def line(tag, v):
+    r = v["res"]
+    print(f"  [{tag:30s}] minFull={v['minFull']:+.2f} minHold={v['minHold']:+.2f} "
+          f"minTr={v['minTr']:3d} maxDD={v['maxDD']*100:4.0f}% | "
+          f"BTC F{r['BTC']['full']['sharpe']:+.2f}/H{r['BTC']['hold']['sharpe']:+.2f}/DD{r['BTC']['full']['maxdd']*100:.0f}% "
+          f"ETH F{r['ETH']['full']['sharpe']:+.2f}/H{r['ETH']['hold']['sharpe']:+.2f}/DD{r['ETH']['full']['maxdd']*100:.0f}%")
+
+
+# ---------------------------------------------------------------------------
+# Causality (truncated-prefix) on the structural pctl entries.
+# ---------------------------------------------------------------------------
+def check_causality(cfg, p, asset="BTC", tail=200):
+    return skr.check_causality(cfg, p, asset, tail=tail)
+
+
+# ---------------------------------------------------------------------------
+# Marginal vs TP01 on a (cfg, params) pair (50/50 daily, same convention as skyhooklib).
+# ---------------------------------------------------------------------------
+def marginal_struct(cfg, p):
+    def daily(a):
+        ltf, htf = sk.frames(a)
+        ent = skr.pctl_entries(ltf, htf, p, **cfg)
+        m = backtest_signals(ltf, ent, fee_rt=FEE, leverage=1.0, asset=a, tf="230m")
+        s = pd.Series(m.equity, index=pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True)))
+        return s.resample("1D").last().ffill().pct_change().dropna()
+    sb, se = daily("BTC"), daily("ETH")
+    J = pd.concat({"BTC": sb, "ETH": se}, axis=1, join="inner").fillna(0.0)
+    cand = 0.5 * J["BTC"] + 0.5 * J["ETH"]
+    return al.marginal_vs_tp01(cand)
+
+
+def fee_sweep(cfg, p):
+    ok = True
+    rows = {}
+    for a in ("BTC", "ETH"):
+        ltf, htf = sk.frames(a)
+        ent = skr.pctl_entries(ltf, htf, p, **cfg)
+        row = []
+        for f in (0.0, 0.001, 0.002, 0.003):
+            m = backtest_signals(ltf, ent, fee_rt=f, leverage=1.0, asset=a, tf="230m")
+            row.append((f, round(m.sharpe, 3)))
+        rows[a] = row
+        ok = ok and (dict(row)[0.003] > 0)
+    return ok, rows
+
+
+if __name__ == "__main__":
+    print("=== SKH2_PCTL_DD : percentile-rank regime tuned for DD<30 ===\n")
+
+    # -----------------------------------------------------------------------
+    # STAGE 1 — coarse sweep: regime bands (pctl space) x stop tightness.
+    # Winner exits (tp7/24/16) carried; we vary sl_atr and the regime cfg.
+    # Intuition for DD cut:
+    #   - cap vola_hi (drop blow-off-vol entries) ; modest vol floor (live tape)
+    #   - tighter sl_atr (2.0/1.8) caps per-trade loss.
+    # -----------------------------------------------------------------------
+    print("--- STAGE 1: regime band x stop sweep (exits tp7/24/16) ---")
+    band_cfgs = {
+        # name: pctl regime cfg (expanding unless _r)
+        "volaHi95_vol0":   dict(vola_win=None, vol_win=None, vola_lo=0.35, vola_hi=0.95, vol_lo=0.0,  vol_hi=1.0),
+        "volaHi90_vol0":   dict(vola_win=None, vol_win=None, vola_lo=0.35, vola_hi=0.90, vol_lo=0.0,  vol_hi=1.0),
+        "volaMid_vol0":    dict(vola_win=None, vol_win=None, vola_lo=0.25, vola_hi=0.85, vol_lo=0.0,  vol_hi=1.0),
+        "volaMid_volFlr":  dict(vola_win=None, vol_win=None, vola_lo=0.25, vola_hi=0.85, vol_lo=0.30, vol_hi=1.0),
+        "volaHi90_volFlr": dict(vola_win=None, vol_win=None, vola_lo=0.35, vola_hi=0.90, vol_lo=0.30, vol_hi=1.0),
+        "volaCap80_volFlr":dict(vola_win=None, vol_win=None, vola_lo=0.30, vola_hi=0.80, vol_lo=0.30, vol_hi=1.0),
+    }
+    sls = [2.5, 2.0, 1.8]
+    stage1 = {}
+    for bname, cfg in band_cfgs.items():
+        for sl in sls:
+            p = winner_exit_params(sl_atr=sl)
+            tag = f"{bname}|sl{sl}"
+            v = summarize(eval_pair(cfg, p))
+            stage1[tag] = (cfg, p, v)
+            line(tag, v)
+
+    # Pick DD<30 candidates with the best minHold (need minHold>=~0.7).
+    sub30 = {t: tup for t, tup in stage1.items() if tup[2]["maxDD"] < 0.30}
+    print(f"\n--- STAGE 1: configs with maxDD<30%: {len(sub30)} ---")
+    for t, (_, _, v) in sorted(sub30.items(), key=lambda kv: -kv[1][2]["minHold"]):
+        line(t, v)
+
+    # -----------------------------------------------------------------------
+    # STAGE 2 — refine: take best DD<30 (and near-30 with high hold) candidates,
+    # fine-tune bands/stop to push minHold up while keeping DD<30.
+    # -----------------------------------------------------------------------
+    print("\n--- STAGE 2: refinement around best DD<30 / high-hold cells ---")
+    refine = {
+        # tighter blow-off cap + small vol floor, sl 1.8-2.0
+        "R1": (dict(vola_win=None, vol_win=None, vola_lo=0.30, vola_hi=0.85, vol_lo=0.20, vol_hi=1.0), winner_exit_params(sl_atr=2.0)),
+        "R2": (dict(vola_win=None, vol_win=None, vola_lo=0.30, vola_hi=0.85, vol_lo=0.20, vol_hi=1.0), winner_exit_params(sl_atr=1.8)),
+        "R3": (dict(vola_win=None, vol_win=None, vola_lo=0.30, vola_hi=0.88, vol_lo=0.30, vol_hi=1.0), winner_exit_params(sl_atr=2.0)),
+        "R4": (dict(vola_win=None, vol_win=None, vola_lo=0.35, vola_hi=0.85, vol_lo=0.30, vol_hi=1.0), winner_exit_params(sl_atr=2.0)),
+        # rolling-window regime (recent), which reacts faster to regime shift -> may cut DD
+        "R5": (dict(vola_win=120, vol_win=120, vola_lo=0.30, vola_hi=0.85, vol_lo=0.20, vol_hi=1.0), winner_exit_params(sl_atr=2.0)),
+        "R6": (dict(vola_win=120, vol_win=120, vola_lo=0.30, vola_hi=0.85, vol_lo=0.30, vol_hi=1.0), winner_exit_params(sl_atr=1.8)),
+        # tighter tp to bank faster (lower DD) with tight sl
+        "R7": (dict(vola_win=None, vol_win=None, vola_lo=0.30, vola_hi=0.85, vol_lo=0.20, vol_hi=1.0), winner_exit_params(sl_atr=2.0, tp_atr=6.0)),
+        "R8": (dict(vola_win=None, vol_win=None, vola_lo=0.30, vola_hi=0.85, vol_lo=0.20, vol_hi=1.0), winner_exit_params(sl_atr=1.6)),
+    }
+    stage2 = {}
+    for t, (cfg, p) in refine.items():
+        v = summarize(eval_pair(cfg, p))
+        stage2[t] = (cfg, p, v)
+        line(t, v)
+
+    # -----------------------------------------------------------------------
+    # PICK BEST: among ALL cells, prefer maxDD<0.30 AND minHold>=0.65; rank by
+    # (DD<30) then minHold then -DD. Fall back to best minHold if none sub-30.
+    # -----------------------------------------------------------------------
+    allcells = {**stage1, **stage2}
+    def score(tup):
+        v = tup[2]
+        dd_ok = v["maxDD"] < 0.30
+        hold_ok = v["minHold"] >= 0.65
+        full_ok = v["minFull"] >= 0.5
+        tr_ok = v["minTr"] >= 20
+        # primary: meets all gates; secondary: minHold; tertiary: lower DD
+        return (dd_ok and hold_ok and full_ok and tr_ok, dd_ok, v["minHold"], -v["maxDD"])
+    best_tag = max(allcells, key=lambda t: score(allcells[t]))
+    best_cfg, best_p, best_v = allcells[best_tag]
+    print(f"\n*** SELECTED = {best_tag} ***")
+    line(best_tag, best_v)
+
+    # -----------------------------------------------------------------------
+    # FULL VERIFICATION on selected: causality + fee sweep + marginal.
+    # -----------------------------------------------------------------------
+    print("\n--- causality (truncated-prefix) ---")
+    cB = check_causality(best_cfg, best_p, "BTC")
+    cE = check_causality(best_cfg, best_p, "ETH")
+    causality_ok = bool(cB["ok"] and cE["ok"])
+    print(f"  BTC={cB}  ETH={cE}  -> causality_ok={causality_ok}")
+
+    print("\n--- fee sweep (FULL sharpe) ---")
+    fee_ok, frows = fee_sweep(best_cfg, best_p)
+    for a, row in frows.items():
+        print(f"  {a}: " + " ".join(f"{f*100:.2f}%={s:+.2f}" for f, s in row))
+    print(f"  fee_survives 0.30%RT (both): {fee_ok}")
+
+    print("\n--- marginal vs TP01 (selected) ---")
+    marg = marginal_struct(best_cfg, best_p)
+    corr_full = marg.get("corr_full")
+    verdict = marg.get("marginal_verdict")
+    has_edge = marg.get("has_insample_edge")
+    is_hedge = marg.get("is_hedge")
+    robust_oos = marg.get("robust_oos")
+    multicut = marg.get("multicut_persistent")
+    clean_yr = marg.get("clean_year_uplift")
+    w25 = marg.get("blends", {}).get("w25", {})
+    w50 = marg.get("blends", {}).get("w50", {})
+    up_h = w25.get("uplift_hold")
+    print(f"  corr_full={corr_full} corr_hold={marg.get('corr_hold')}")
+    print(f"  marginal_verdict={verdict} robust_oos={robust_oos} multicut_persistent={multicut}")
+    print(f"  has_insample_edge={has_edge} is_hedge={is_hedge} cand_insample_sharpe={marg.get('cand_insample_sharpe')}")
+    print(f"  clean_year_uplift={clean_yr} jackknife_min_uplift={marg.get('jackknife_min_uplift')}")
+    print(f"  blend w25: uplift_hold={up_h} uplift_full={w25.get('uplift_full')}")
+    print(f"  blend w50: full={w50.get('full')} hold={w50.get('hold')} uplift_hold={w50.get('uplift_hold')} dd={w50.get('dd')}")
+
+    # grade (mirror sk verdict thresholds): PASS if minTr>=20 & minFull>=0.5 & minHold>=0.2 & feeOK
+    mf, mh, mt, mdd = best_v["minFull"], best_v["minHold"], best_v["minTr"], best_v["maxDD"]
+    if mt >= 20 and mf >= 0.5 and mh >= 0.2 and fee_ok:
+        grade = "PASS"
+    elif mt >= 20 and mf >= 0.3 and mh >= 0.0:
+        grade = "WEAK"
+    else:
+        grade = "FAIL"
+    earns_slot = (grade != "FAIL") and (verdict == "ADDS") and bool(robust_oos) and (not bool(is_hedge))
+    beats_winner = bool(earns_slot and mdd < 0.30 and (up_h is not None and up_h >= 0.55) and mh >= 0.65)
+
+    print("\n" + "=" * 70)
+    print("FINAL BLOCK — SKH2_PCTL_DD")
+    print("=" * 70)
+    print(f"best_cfg(regime) = {best_cfg}")
+    print(f"best_params      = ptn_n={best_p.ptn_n} sl_atr={best_p.sl_atr} tp_atr={best_p.tp_atr} "
+          f"uscitalong={best_p.uscitalong} uscitashort={best_p.uscitashort} exit_mode={best_p.exit_mode}")
+    print(f"grade={grade}")
+    print(f"minFull={mf:+.3f}  minHold={mh:+.3f}  max_dd={mdd:.4f} ({mdd*100:.0f}%)  n_trades_min={mt}")
+    print(f"fee@0.30%RT survives={fee_ok}  causality_ok={causality_ok}")
+    print(f"marginal: verdict={verdict} corr_full={corr_full} has_insample_edge={has_edge} "
+          f"is_hedge={is_hedge} robust_oos={robust_oos} multicut_persistent={multicut} clean_year_uplift={clean_yr}")
+    print(f"blend w25 uplift_hold={up_h}  blend w50 full={w50.get('full')}/hold={w50.get('hold')}/dd={w50.get('dd')}")
+    print(f"earns_slot={earns_slot}")
+    print(f"beats_winner={beats_winner}")
+    print("=" * 70)
+
+    # machine-readable echo for the agent
+    import json
+    print("RESULT_JSON=" + json.dumps({
+        "best_cfg": best_cfg,
+        "best_params": {"ptn_n": best_p.ptn_n, "sl_atr": best_p.sl_atr, "tp_atr": best_p.tp_atr,
+                        "uscitalong": best_p.uscitalong, "uscitashort": best_p.uscitashort,
+                        "exit_mode": best_p.exit_mode,
+                        "vola_lo": best_cfg["vola_lo"], "vola_hi": best_cfg["vola_hi"],
+                        "vol_lo": best_cfg["vol_lo"], "vol_hi": best_cfg["vol_hi"],
+                        "vola_win": best_cfg["vola_win"], "vol_win": best_cfg["vol_win"]},
+        "grade": grade, "minFull": mf, "minHold": mh, "max_dd": mdd, "n_trades_min": mt,
+        "fee_ok": fee_ok, "causality_ok": causality_ok,
+        "marginal_verdict": verdict, "corr_full": corr_full, "has_insample_edge": has_edge,
+        "is_hedge": is_hedge, "robust_oos": robust_oos, "multicut_persistent": multicut,
+        "clean_year_uplift": clean_yr, "blend_w25_uplift_hold": up_h,
+        "w50_full": w50.get("full"), "w50_hold": w50.get("hold"), "w50_dd": w50.get("dd"),
+        "earns_slot": earns_slot, "beats_winner": beats_winner,
+    }, default=str))

scripts/research/skyhook/runs/SKH2_REGIME_TIGHT.py

@@ -0,0 +1,163 @@
+"""SKH2_REGIME_TIGHT — DD-reduction wave, family: tighter regime selectivity.
+
+Hypothesis: make the regime band MORE selective (narrow vola band, add a volume floor)
+so only the cleanest setups trade -> fewer, higher-quality entries -> lower standalone DD,
+while keeping the winner's asymmetric exits (sl 2.5 / tp 7.0, uscitalong 24 / short 16).
+
+Baseline winner to beat (V2):
+  SkyhookParams(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16,
+                vola_lo=35, vola_hi=95, vol_lo=0)
+  minFull +0.83 minHold +0.81 ; DD BTC 34% / ETH 31% (the unmet goal: DD<30%) ;
+  marginal ADDS, corr_full 0.05, blend w25 uplift_hold +0.58, w50 full 1.59 / hold 1.04 / DD 12.5%.
+
+BEATS-THE-WINNER iff: earns_slot AND max_dd<0.30 AND blend_w25_uplift_hold>=0.55 AND min_hold>=0.65.
+
+Everything is param-based (no new feature) -> sk.study/sk.marginal/sk.causality are directly valid.
+"""
+from __future__ import annotations
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+import itertools
+import skyhooklib as sk
+from src.strategies.skyhook import SkyhookParams
+
+# --- the winner's exits, frozen ----------------------------------------------------------
+EXITS = dict(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16)
+
+
+def mk(vola_lo, vola_hi, vol_lo, **extra):
+    return SkyhookParams(vola_lo=vola_lo, vola_hi=vola_hi, vol_lo=vol_lo, **EXITS, **extra)
+
+
+def earns_slot(rep, mg):
+    return (rep["verdict"]["grade"] != "FAIL"
+            and mg.get("marginal_verdict") == "ADDS"
+            and bool(mg.get("robust_oos"))
+            and not bool(mg.get("is_hedge")))
+
+
+def beats(rep, mg, max_dd):
+    es = earns_slot(rep, mg)
+    upl = mg.get("blends", {}).get("w25", {}).get("uplift_hold")
+    mh = rep["verdict"]["min_asset_holdout_sharpe"]
+    return (es and max_dd < 0.30 and (upl is not None and upl >= 0.55) and mh >= 0.65)
+
+
+def summarize(name, p):
+    rep = sk.study(name, p)
+    v = rep["verdict"]
+    dd = max(rep["per_asset"][a]["full"]["maxdd"] for a in rep["per_asset"])
+    nt = v["min_trades"]
+    print(sk.fmt(rep))
+    print(f"   >> maxDD(BTC,ETH) = {dd*100:.1f}%   minTrades={nt}")
+    return rep, dd
+
+
+if __name__ == "__main__":
+    # ---- STAGE 1: coarse grid over the family levers -----------------------------------
+    vola_los = [35.0, 40.0, 45.0, 50.0]       # 35 = winner; 40/45/50 = tighter floor
+    vola_his = [85.0, 90.0, 95.0]             # 95 = winner; 85/90 = clip blow-off harder
+    vol_los  = [0.0, 30.0, 40.0, 50.0]        # 0 = winner; floor = require live volume
+
+    rows = []
+    print("########## STAGE 1: coarse DD scan (study, both assets) ##########")
+    for vlo, vhi, vol in itertools.product(vola_los, vola_his, vol_los):
+        p = mk(vlo, vhi, vol)
+        rep = sk.study(f"vlo{vlo:.0f}_vhi{vhi:.0f}_vol{vol:.0f}", p)
+        v = rep["verdict"]
+        dd = max(rep["per_asset"][a]["full"]["maxdd"] for a in rep["per_asset"])
+        nt = v["min_trades"]
+        ddb = rep["per_asset"]["BTC"]["full"]["maxdd"]
+        dde = rep["per_asset"]["ETH"]["full"]["maxdd"]
+        rows.append(dict(vlo=vlo, vhi=vhi, vol=vol, grade=v["grade"],
+                         mf=v["min_asset_full_sharpe"], mh=v["min_asset_holdout_sharpe"],
+                         dd=dd, ddb=ddb, dde=dde, nt=nt, fee=v["fee_survives"]))
+        print(f"  vlo{vlo:.0f} vhi{vhi:.0f} vol{vol:.0f}: {v['grade']:4s} "
+              f"mF={v['min_asset_full_sharpe']:+.2f} mH={v['min_asset_holdout_sharpe']:+.2f} "
+              f"DD={dd*100:4.1f}% (B{ddb*100:.0f}/E{dde*100:.0f}) nT={nt:3d} feeOK={v['fee_survives']}")
+
+    # ---- rank: candidates with DD<30%, minTrades>=20, fee OK, holdout>=0.65 -------------
+    print("\n########## STAGE 1 RANK (filter DD<30, nT>=20, fee OK, mH>=0.65, not FAIL) ##########")
+    cand = [r for r in rows if r["dd"] < 0.30 and r["nt"] >= 20 and r["fee"]
+            and r["mh"] >= 0.65 and r["grade"] != "FAIL"]
+    cand.sort(key=lambda r: (-r["mh"], r["dd"]))     # prefer high holdout then low DD
+    if not cand:
+        print("  (none met all hard filters; falling back to lowest-DD with nT>=20 & not FAIL)")
+        cand = [r for r in rows if r["nt"] >= 20 and r["grade"] != "FAIL"]
+        cand.sort(key=lambda r: (r["dd"], -r["mh"]))
+    for r in cand[:8]:
+        print(f"  vlo{r['vlo']:.0f} vhi{r['vhi']:.0f} vol{r['vol']:.0f}: {r['grade']} "
+              f"mF={r['mf']:+.2f} mH={r['mh']:+.2f} DD={r['dd']*100:.1f}% nT={r['nt']}")
+
+    # ---- STAGE 2: full diligence (causality + marginal) on top few ---------------------
+    print("\n########## STAGE 2: causality + marginal on top candidates ##########")
+    best = None
+    top = cand[:4]
+    for r in top:
+        p = mk(r["vlo"], r["vhi"], r["vol"])
+        name = f"TIGHT vlo{r['vlo']:.0f}_vhi{r['vhi']:.0f}_vol{r['vol']:.0f}"
+        print(f"\n----- {name} -----")
+        rep, dd = summarize(name, p)
+        cau = sk.causality(p, "BTC")
+        cau_e = sk.causality(p, "ETH")
+        cau_ok = bool(cau["ok"] and cau_e["ok"])
+        mg = sk.marginal(p)
+        es = earns_slot(rep, mg)
+        bw = beats(rep, mg, dd) and cau_ok
+        w25 = mg.get("blends", {}).get("w25", {})
+        w50 = mg.get("blends", {}).get("w50", {})
+        print(f"   causality BTC={cau} ETH={cau_e} -> ok={cau_ok}")
+        print(f"   marginal: verdict={mg.get('marginal_verdict')} corr_full={mg.get('corr_full')} "
+              f"corr_hold={mg.get('corr_hold')} insample_edge={mg.get('has_insample_edge')} "
+              f"(cand_is_sh={mg.get('cand_insample_sharpe')}) hedge={mg.get('is_hedge')} "
+              f"robust_oos={mg.get('robust_oos')} multicut={mg.get('multicut_persistent')} "
+              f"clean_year_uplift={mg.get('clean_year_uplift')}")
+        print(f"   blend w25: uplift_hold={w25.get('uplift_hold')} uplift_full={w25.get('uplift_full')} "
+              f"| w50: full={w50.get('full')} hold={w50.get('hold')} dd={w50.get('dd')}")
+        print(f"   >> earns_slot={es}  beats_winner={bw}")
+        cand_obj = dict(name=name, p=p, rep=rep, dd=dd, mg=mg, cau_ok=cau_ok,
+                        es=es, bw=bw, w25=w25, w50=w50,
+                        mf=rep["verdict"]["min_asset_full_sharpe"],
+                        mh=rep["verdict"]["min_asset_holdout_sharpe"],
+                        nt=rep["verdict"]["min_trades"],
+                        fee=rep["verdict"]["fee_survives"])
+        # pick best: prefer beats_winner, then earns_slot+lowDD, then lowDD
+        def keyf(c):
+            return (c["bw"], c["es"], -c["dd"], c["mh"])
+        if best is None or keyf(cand_obj) > keyf(best):
+            best = cand_obj
+
+    # ---- FINAL BLOCK -------------------------------------------------------------------
+    print("\n\n==================== FINAL — REGIME_TIGHT BEST ====================")
+    if best is None:
+        print("NO CANDIDATE — no config passed even the soft filter.")
+        sys.exit(0)
+    b = best
+    pf = {k: getattr(b["p"], k) for k in b["p"].__dataclass_fields__}
+    mg = b["mg"]
+    print(f"config: vola_lo={b['p'].vola_lo} vola_hi={b['p'].vola_hi} vol_lo={b['p'].vol_lo} "
+          f"ptn_n={b['p'].ptn_n} sl_atr={b['p'].sl_atr} tp_atr={b['p'].tp_atr} "
+          f"uscitalong={b['p'].uscitalong} uscitashort={b['p'].uscitashort}")
+    print(f"minFull={b['mf']:+.3f}  minHold={b['mh']:+.3f}  max_dd={b['dd']*100:.1f}%  "
+          f"n_trades_min={b['nt']}  fee@0.30%OK={b['fee']}  causality_ok={b['cau_ok']}")
+    print(f"earns_slot={b['es']}  beats_winner={b['bw']}")
+    print("FULL marginal dict:")
+    for k in ("corr_full", "corr_hold", "marginal_verdict", "has_insample_edge",
+              "cand_insample_sharpe", "is_hedge", "robust_oos", "multicut_persistent",
+              "clean_year_uplift", "jackknife_min_uplift", "multicut_uplift"):
+        print(f"    {k} = {mg.get(k)}")
+    print(f"    blend w25: {mg.get('blends', {}).get('w25')}")
+    print(f"    blend w50: {mg.get('blends', {}).get('w50')}")
+    print("PARAMS:", pf)
+    # machine-readable tail for me to parse
+    import json
+    print("\nRESULT_JSON " + json.dumps(dict(
+        family="REGIME_TIGHT", params=pf,
+        minFull=b["mf"], minHold=b["mh"], max_dd=b["dd"], n_trades_min=b["nt"],
+        fee_ok=b["fee"], causality_ok=b["cau_ok"], earns_slot=b["es"], beats=b["bw"],
+        corr_full=mg.get("corr_full"), marginal_verdict=mg.get("marginal_verdict"),
+        has_insample_edge=mg.get("has_insample_edge"), is_hedge=mg.get("is_hedge"),
+        robust_oos=mg.get("robust_oos"), multicut_persistent=mg.get("multicut_persistent"),
+        clean_year_uplift=mg.get("clean_year_uplift"),
+        w25_uplift_hold=mg.get("blends", {}).get("w25", {}).get("uplift_hold"),
+    ), default=str))

scripts/research/skyhook/runs/SKH2_TPSL_DD.py

@@ -0,0 +1,191 @@
+"""SKH2_TPSL_DD: RR / stop fine grid around the V2 winner to push standalone maxDD < 30%
+while holding min-asset HOLD-OUT >= ~0.70 and earns_slot True.
+
+Winner baseline:
+  SkyhookParams(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16,
+                vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+  -> minFull +0.83, minHold +0.81, DD BTC 34% / ETH 31% (THE PROBLEM), earns_slot True.
+
+Family task: sl_atr in {1.75,2.0,2.25,2.5}, tp_atr in {5,6,7,8}, exit_mode 'pct' vs 'atr'.
+Tighter SL cuts DD but can lower hold-out. Find DD<30 cell + minHold>=0.7 + plateau.
+"""
+from __future__ import annotations
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+import numpy as np  # noqa: E402
+import skyhooklib as sk  # noqa: E402
+from src.strategies.skyhook import SkyhookParams  # noqa: E402
+
+# Winner fixed (non-RR) fields:
+BASE = dict(ptn_n=45, uscitalong=24, uscitashort=16, vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+
+
+def mk(sl_atr=None, tp_atr=None, exit_mode="atr", sl_pct=None, tp_pct=None):
+    kw = dict(BASE)
+    kw["exit_mode"] = exit_mode
+    if exit_mode == "atr":
+        kw["sl_atr"] = sl_atr
+        kw["tp_atr"] = tp_atr
+    else:
+        kw["sl_pct"] = sl_pct
+        kw["tp_pct"] = tp_pct
+    return SkyhookParams(**kw)
+
+
+def metrics(p):
+    """FULL/HOLD/DD min-asset + fee@0.30 + trades, both assets."""
+    pa = {}
+    fee_ok = True
+    for a in ("BTC", "ETH"):
+        r = sk.run_asset(a, p)
+        # fee sweep at 0.30%
+        rf = sk.run_asset(a, p, fee_rt=0.003)
+        fee_ok = fee_ok and (rf["full"]["sharpe"] > 0)
+        pa[a] = dict(full=r["full"], hold=r["holdout"], yearly=r["yearly"],
+                     fee30=rf["full"]["sharpe"])
+    minFull = min(pa[a]["full"]["sharpe"] for a in pa)
+    minHold = min(pa[a]["hold"]["sharpe"] for a in pa)
+    minTr = min(pa[a]["full"]["n_trades"] for a in pa)
+    maxDD = max(pa[a]["full"]["maxdd"] for a in pa)
+    return dict(pa=pa, minFull=minFull, minHold=minHold, minTr=minTr, maxDD=maxDD, fee_ok=fee_ok)
+
+
+def grade_of(m):
+    enough = m["minTr"] >= 20
+    if enough and m["minFull"] >= 0.5 and m["minHold"] >= 0.2 and m["fee_ok"]:
+        return "PASS"
+    if enough and m["minFull"] >= 0.3 and m["minHold"] >= 0.0:
+        return "WEAK"
+    return "FAIL"
+
+
+if __name__ == "__main__":
+    # ---- STAGE 1: coarse ATR grid (the family core) ----
+    sl_grid = [1.75, 2.0, 2.25, 2.5]
+    tp_grid = [5.0, 6.0, 7.0, 8.0]
+    rows = []
+    print("##### STAGE 1: ATR grid (sl_atr x tp_atr) #####")
+    print(f"{'sl':>5} {'tp':>4} | {'minFull':>8} {'minHold':>8} {'maxDD%':>7} {'minTr':>6} "
+          f"{'BTC_DD':>7} {'ETH_DD':>7} {'feeOK':>5} {'grade':>5}")
+    for sl in sl_grid:
+        for tp in tp_grid:
+            p = mk(sl_atr=sl, tp_atr=tp, exit_mode="atr")
+            m = metrics(p)
+            g = grade_of(m)
+            bdd = m["pa"]["BTC"]["full"]["maxdd"]
+            edd = m["pa"]["ETH"]["full"]["maxdd"]
+            rows.append((sl, tp, "atr", m, g))
+            print(f"{sl:>5} {tp:>4} | {m['minFull']:>+8.2f} {m['minHold']:>+8.2f} "
+                  f"{m['maxDD']*100:>7.1f} {m['minTr']:>6} {bdd*100:>7.1f} {edd*100:>7.1f} "
+                  f"{str(m['fee_ok']):>5} {g:>5}")
+
+    # ---- STAGE 2: pct exit grid (a few sensible RR pairs ~ matching ATR ratios) ----
+    # ATR LTF ~ a few % of price; pct exit gives a HARD dollar cap on DD per trade.
+    print("\n##### STAGE 2: PCT exit grid (sl_pct x tp_pct) #####")
+    print(f"{'slP':>6} {'tpP':>6} | {'minFull':>8} {'minHold':>8} {'maxDD%':>7} {'minTr':>6} "
+          f"{'BTC_DD':>7} {'ETH_DD':>7} {'feeOK':>5} {'grade':>5}")
+    pct_pairs = [(0.02, 0.06), (0.025, 0.07), (0.03, 0.075), (0.03, 0.09),
+                 (0.035, 0.10), (0.04, 0.10),
+                 # dense neighbourhood around the DD<30 winner (0.025,0.07) to prove a plateau:
+                 (0.0225, 0.065), (0.0225, 0.07), (0.0225, 0.075),
+                 (0.025, 0.0625), (0.025, 0.065), (0.025, 0.075), (0.025, 0.08),
+                 (0.0275, 0.065), (0.0275, 0.07), (0.0275, 0.075)]
+    for slp, tpp in pct_pairs:
+        p = mk(exit_mode="pct", sl_pct=slp, tp_pct=tpp)
+        m = metrics(p)
+        g = grade_of(m)
+        bdd = m["pa"]["BTC"]["full"]["maxdd"]
+        edd = m["pa"]["ETH"]["full"]["maxdd"]
+        rows.append((slp, tpp, "pct", m, g))
+        print(f"{slp:>6} {tpp:>6} | {m['minFull']:>+8.2f} {m['minHold']:>+8.2f} "
+              f"{m['maxDD']*100:>7.1f} {m['minTr']:>6} {bdd*100:>7.1f} {edd*100:>7.1f} "
+              f"{str(m['fee_ok']):>5} {g:>5}")
+
+    # ---- Pick best: DD<30, minHold>=0.7, grade!=FAIL; tie-break by minHold then minFull ----
+    def ok_dd(r):
+        return r[3]["maxDD"] < 0.30 and r[3]["minHold"] >= 0.70 and r[4] != "FAIL"
+    cands = [r for r in rows if ok_dd(r)]
+    if not cands:
+        # relax: DD<30 and minHold>=0.65
+        cands = [r for r in rows if r[3]["maxDD"] < 0.30 and r[3]["minHold"] >= 0.65 and r[4] != "FAIL"]
+        relaxed = True
+    else:
+        relaxed = False
+    if not cands:
+        # fall back to lowest DD among non-FAIL with decent hold
+        cands = [r for r in rows if r[4] != "FAIL"]
+    # rank: among DD<30 cells, maximize a balanced score (minHold + minFull) so we don't pick a
+    # low-DD-but-weak-Sharpe corner. DD is already gated < 0.30 above, so optimise value next.
+    cands_sorted = sorted(cands, key=lambda r: -(r[3]["minHold"] + r[3]["minFull"]))
+    best = cands_sorted[0]
+    print(f"\n##### BEST PICK (relaxed={relaxed if cands else 'fallback'}): "
+          f"{'sl/tp' if best[2]=='atr' else 'slP/tpP'}=({best[0]},{best[1]}) mode={best[2]} #####")
+
+    # Build best params
+    if best[2] == "atr":
+        bp = mk(sl_atr=best[0], tp_atr=best[1], exit_mode="atr")
+        best_cfg = dict(ptn_n=45, sl_atr=best[0], tp_atr=best[1], uscitalong=24,
+                        uscitashort=16, vola_lo=35.0, vola_hi=95.0, vol_lo=0.0,
+                        exit_mode="atr")
+    else:
+        bp = mk(exit_mode="pct", sl_pct=best[0], tp_pct=best[1])
+        best_cfg = dict(ptn_n=45, sl_pct=best[0], tp_pct=best[1], uscitalong=24,
+                        uscitashort=16, vola_lo=35.0, vola_hi=95.0, vol_lo=0.0,
+                        exit_mode="pct")
+
+    # ---- Full verification of best: study + causality + marginal ----
+    print("\n##### FULL STUDY of BEST #####")
+    rep = sk.study("SKH2_TPSL_DD-BEST", bp)
+    print(sk.fmt(rep))
+    caus = sk.causality(bp, "BTC")
+    caus_eth = sk.causality(bp, "ETH")
+    print(f"\ncausality BTC: {caus}")
+    print(f"causality ETH: {caus_eth}")
+    mg = sk.marginal(bp)
+
+    m = best[3]
+    g = rep["verdict"]["grade"]
+    earns = (g != "FAIL" and mg.get("marginal_verdict") == "ADDS"
+             and bool(mg.get("robust_oos")) and not bool(mg.get("is_hedge")))
+    w25 = mg.get("blends", {}).get("w25", {})
+    w50 = mg.get("blends", {}).get("w50", {})
+    beats = (earns and m["maxDD"] < 0.30
+             and (w25.get("uplift_hold") or -9) >= 0.55 and m["minHold"] >= 0.65)
+
+    print("\n========== FINAL BLOCK ==========")
+    print(f"best_cfg          = {best_cfg}")
+    print(f"exit_mode         = {best[2]}")
+    print(f"minFull           = {m['minFull']:+.3f}")
+    print(f"minHold           = {m['minHold']:+.3f}")
+    print(f"max_dd (BTC/ETH)  = {m['maxDD']:.4f} (BTC {m['pa']['BTC']['full']['maxdd']:.4f} / "
+          f"ETH {m['pa']['ETH']['full']['maxdd']:.4f})")
+    print(f"n_trades_min      = {m['minTr']}")
+    print(f"fee@0.30 OK       = {m['fee_ok']}  (BTC {m['pa']['BTC']['fee30']:+.2f} / "
+          f"ETH {m['pa']['ETH']['fee30']:+.2f})")
+    print(f"causality_ok      = {caus['ok'] and caus_eth['ok']} "
+          f"(BTC mism={caus['mismatches']} ETH mism={caus_eth['mismatches']})")
+    print(f"grade             = {g}")
+    print("--- marginal vs TP01 ---")
+    print(f"corr_full         = {mg.get('corr_full')}")
+    print(f"corr_hold         = {mg.get('corr_hold')}")
+    print(f"marginal_verdict  = {mg.get('marginal_verdict')}")
+    print(f"has_insample_edge = {mg.get('has_insample_edge')}")
+    print(f"is_hedge          = {mg.get('is_hedge')}")
+    print(f"robust_oos        = {mg.get('robust_oos')}")
+    print(f"multicut_persist  = {mg.get('multicut_persistent')}")
+    print(f"clean_year_uplift = {mg.get('clean_year_uplift')}")
+    print(f"jackknife_min_upl = {mg.get('jackknife_min_uplift')}")
+    print(f"cand_insample_sh  = {mg.get('cand_insample_sharpe')}")
+    print(f"blend w25         = {w25}")
+    print(f"blend w50         = {w50}")
+    print(f"earns_slot        = {earns}")
+    print(f"BEATS_WINNER      = {beats}")
+
+    # ---- plateau report: neighbors of best in the same mode ----
+    print("\n##### PLATEAU (neighbors of best) #####")
+    nbrs = [r for r in rows if r[2] == best[2]]
+    nbrs_sorted = sorted(nbrs, key=lambda r: (r[3]["maxDD"]))
+    for r in nbrs_sorted[:8]:
+        tag = f"({r[0]},{r[1]})"
+        print(f"  {r[2]} {tag:>14}: DD={r[3]['maxDD']*100:5.1f}% minFull={r[3]['minFull']:+.2f} "
+              f"minHold={r[3]['minHold']:+.2f} grade={r[4]}")

scripts/research/skyhook/runs/SKH2_VOLTGT.py

@@ -0,0 +1,322 @@
+"""SKH2_VOLTGT — CAUSAL vol-target overlay on the V2 winner's daily return series.
+
+Family: vol-target overlay [VOLTGT]. Wave goal: cut standalone maxDD < 30% while keeping
+min-asset hold-out Sharpe >= ~0.70 and earns_slot True.
+
+Method:
+  * Build the winner's daily return series per asset (from the honest intrabar equity).
+  * Scale each day t by  lev_t = min(cap, target_vol / rv_{t-1})  where rv_{t-1} is the
+    trailing realized vol KNOWN AT t-1 (rolling window of past daily returns, .shift(1)).
+    -> strictly causal: the scaler at day t uses returns up to and including day t-1 only.
+  * scaled_ret_t = lev_t * ret_t.  Rebuild scaled equity, measure DD per asset + combined.
+  * Run altlib.marginal_vs_tp01 on the 50/50 scaled-combined daily series.
+
+We sweep target_vol in {15%,20%,25%}, cap in {1.5,2.0}, and a couple of vol windows.
+We prove causality of the scaler two ways:
+  (1) construction (shift(1) -> rv known at t-1),
+  (2) an explicit truncated-prefix recompute: lev_t computed on the full history must equal
+      lev_t recomputed from only the returns up to t-1.
+The underlying winner entries are param-only -> their causality is sk.causality (0 mismatches).
+"""
+from __future__ import annotations
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+import numpy as np
+import pandas as pd
+import skyhooklib as sk
+import altlib as al
+from src.strategies.skyhook import SkyhookParams
+
+HOLDOUT = sk.HOLDOUT
+FEE = sk.FEE_RT
+ANN = np.sqrt(365.25)
+
+# ---- the verified V2 winner (baseline to beat) ----------------------------------------
+WINNER = SkyhookParams(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16,
+                       vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+
+
+def _sh(r: np.ndarray) -> float:
+    r = np.asarray(r, float)
+    r = r[np.isfinite(r)]
+    if len(r) < 2 or np.std(r) == 0:
+        return 0.0
+    return float(np.mean(r) / np.std(r) * ANN)
+
+
+def _maxdd_from_returns(r: pd.Series) -> float:
+    eq = (1.0 + r.fillna(0.0)).cumprod().values
+    pk = np.maximum.accumulate(eq)
+    return float(np.max((pk - eq) / pk))
+
+
+def _split_sharpe(r: pd.Series, mask) -> float:
+    return _sh(r[mask].values)
+
+
+def _trailing_rv(daily: pd.Series, win: int, mode: str) -> pd.Series:
+    """Annualized trailing realized vol KNOWN AT t-1 (shift(1) -> strictly past data)."""
+    if mode == "ewma":
+        # EWMA std of past returns, reacts faster to a vol spike than a flat window
+        v = daily.ewm(span=win, min_periods=max(10, win // 2)).std().shift(1)
+    else:
+        v = daily.rolling(win, min_periods=max(10, win // 2)).std().shift(1)
+    return v * ANN
+
+
+def vol_target_lev(daily: pd.Series, target_vol: float, cap: float, win: int,
+                   floor: float = 0.0, mode: str = "roll") -> pd.Series:
+    """CAUSAL leverage series. rv_{t-1} = annualized trailing realized vol KNOWN at t-1.
+    lev_t = clip(target/rv_{t-1}, floor, cap).  cap<=1.0 => de-risk only (never lever up)."""
+    rv = _trailing_rv(daily, win, mode)
+    lev = (target_vol / rv).clip(lower=floor, upper=cap)
+    # before we have enough history -> stay at min(1.0, cap) (no scaling, no look-ahead)
+    lev = lev.where(rv.notna(), min(1.0, cap)).fillna(min(1.0, cap))
+    return lev
+
+
+def prove_scaler_causal(daily: pd.Series, target_vol: float, cap: float, win: int,
+                        mode: str = "roll", n_checks: int = 60) -> dict:
+    """Truncated-prefix recompute: lev_t built on the FULL series must equal lev_t rebuilt
+    from only returns up to t-1. Any leak (un-shifted vol) would break this."""
+    full = vol_target_lev(daily, target_vol, cap, win, mode=mode)
+    n = len(daily)
+    bad = 0
+    checked = 0
+    mp = max(10, win // 2)
+    idxs = np.linspace(int(n * 0.5), n - 1, n_checks).astype(int)
+    for t in sorted(set(idxs)):
+        if t < 1:
+            continue
+        prefix = daily.iloc[:t]              # returns up to and including day t-1 ONLY
+        if mode == "ewma":
+            rv_prev = prefix.ewm(span=win, min_periods=mp).std().iloc[-1] * ANN
+        else:
+            rv_prev = prefix.rolling(win, min_periods=mp).std().iloc[-1] * ANN
+        if (not np.isfinite(rv_prev)) or rv_prev == 0:
+            lev_t = min(1.0, cap)
+        else:
+            lev_t = float(np.clip(target_vol / rv_prev, 0.0, cap))
+        checked += 1
+        if abs(float(full.iloc[t]) - lev_t) > 1e-9:
+            bad += 1
+    return dict(ok=bool(bad == 0), mismatches=int(bad), checked=int(checked))
+
+
+def winner_daily(asset: str) -> pd.Series:
+    """Winner's RAW daily return series for one asset (from honest intrabar equity)."""
+    return sk.daily_returns(asset, WINNER, FEE)
+
+
+def run_overlay(target_vol: float, cap: float, win: int, floor: float = 0.0,
+                mode: str = "roll") -> dict:
+    """Apply the causal vol-target overlay per asset, combine 50/50, report DD + marginal."""
+    raw = {a: winner_daily(a) for a in ("BTC", "ETH")}
+    scaled = {}
+    lev_stats = {}
+    scaler_caus = {}
+    per_asset_dd_raw = {}
+    per_asset_dd_scaled = {}
+    per_asset_full_sh = {}
+    per_asset_hold_sh = {}
+    for a in ("BTC", "ETH"):
+        d = raw[a]
+        lev = vol_target_lev(d, target_vol, cap, win, floor, mode)
+        s = (lev * d)
+        scaled[a] = s
+        lev_stats[a] = (round(float(lev.mean()), 3), round(float(lev.median()), 3),
+                        round(float((lev >= cap - 1e-9).mean()), 3))
+        scaler_caus[a] = prove_scaler_causal(d, target_vol, cap, win, mode)
+        per_asset_dd_raw[a] = _maxdd_from_returns(d)
+        per_asset_dd_scaled[a] = _maxdd_from_returns(s)
+        hmask = s.index >= HOLDOUT
+        per_asset_full_sh[a] = _sh(s.values)
+        per_asset_hold_sh[a] = _split_sharpe(s, hmask)
+
+    # combined 50/50 scaled series (same convention as sk.skyhook_daily_5050)
+    J = pd.concat(scaled, axis=1, join="inner").fillna(0.0)
+    comb = 0.5 * J["BTC"] + 0.5 * J["ETH"]
+    comb_dd = _maxdd_from_returns(comb)
+    comb_full_sh = _sh(comb.values)
+    comb_hold_sh = _split_sharpe(comb, comb.index >= HOLDOUT)
+
+    mg = al.marginal_vs_tp01(comb)
+
+    max_dd = max(per_asset_dd_scaled.values())   # max over BTC & ETH (per-asset scaled DD)
+    min_full = min(per_asset_full_sh.values())
+    min_hold = min(per_asset_hold_sh.values())
+    return dict(target_vol=target_vol, cap=cap, win=win, floor=floor, mode=mode,
+                lev_stats=lev_stats, scaler_caus=scaler_caus,
+                dd_raw=per_asset_dd_raw, dd_scaled=per_asset_dd_scaled,
+                full_sh=per_asset_full_sh, hold_sh=per_asset_hold_sh,
+                comb_dd=comb_dd, comb_full=comb_full_sh, comb_hold=comb_hold_sh,
+                max_dd=max_dd, min_full=min_full, min_hold=min_hold, marginal=mg)
+
+
+def fee_survives_winner() -> bool:
+    """The vol-target overlay does NOT change trade count/turnover materially (it scales an
+    already-net daily series), so fee survival is the WINNER's fee survival. Report it."""
+    rep = sk.study("WINNER-fee", WINNER)
+    ok = True
+    for a, pa in rep["per_asset"].items():
+        ok = ok and (pa["fee_sweep"].get("0.30%RT", -9) > 0)
+    return ok, rep
+
+
+def winner_min_trades() -> int:
+    rep = sk.study("WINNER-tr", WINNER)
+    return min(pa["full"]["n_trades"] for pa in rep["per_asset"].values())
+
+
+if __name__ == "__main__":
+    print("=" * 90)
+    print("SKH2_VOLTGT — causal vol-target overlay on the V2 winner")
+    print("Winner:", WINNER)
+    print("=" * 90)
+
+    # baseline reference: winner raw (no overlay) for context
+    raw = {a: winner_daily(a) for a in ("BTC", "ETH")}
+    Jr = pd.concat(raw, axis=1, join="inner").fillna(0.0)
+    raw_comb = 0.5 * Jr["BTC"] + 0.5 * Jr["ETH"]
+    print("\n--- WINNER RAW (no overlay), daily-return view ---")
+    for a in ("BTC", "ETH"):
+        print(f"  {a}: dailyFullSh={_sh(raw[a].values):+.2f} "
+              f"holdSh={_split_sharpe(raw[a], raw[a].index>=HOLDOUT):+.2f} "
+              f"DD(daily)={_maxdd_from_returns(raw[a])*100:.0f}%")
+    print(f"  COMB: fullSh={_sh(raw_comb.values):+.2f} "
+          f"holdSh={_split_sharpe(raw_comb, raw_comb.index>=HOLDOUT):+.2f} "
+          f"DD={_maxdd_from_returns(raw_comb)*100:.0f}%")
+    print("  NB daily-view Sharpe != intrabar headline Sharpe (winner minFull +0.83/minHold +0.81 "
+          "are the intrabar numbers). The overlay's job is the DD; we judge marginal+DD on the daily series.")
+
+    # KEY LESSON from v1 grid: cap>1.0 levers UP in low-vol regimes that precede crashes ->
+    # per-asset BTC DD got WORSE (34%->43-55%). To CUT standalone per-asset DD<30% the cap must
+    # be <=1.0 (DE-RISK ONLY: never amplify). We also test EWMA vol (reacts faster to spikes).
+    GRID = []
+    for mode in ("roll", "ewma"):
+        for tv in (0.15, 0.20, 0.25):
+            for cap in (0.8, 1.0):           # de-risk only
+                for win in (20, 30):
+                    GRID.append((tv, cap, win, mode))
+    # FRONTIER scan: how much lever-up (cap) can we allow at tv=25 before BTC DD breaks 30%?
+    # (rolling vol drove the highest uplift; we want max w25 uplift_hold subject to DD<0.30)
+    for cap in (1.1, 1.2, 1.3):
+        for win in (20, 30):
+            GRID.append((0.25, cap, win, "roll"))
+            GRID.append((0.25, cap, win, "ewma"))
+    # plus a couple of cap=1.5 references to show the lever-up failure explicitly
+    for tv in (0.20, 0.25):
+        GRID.append((tv, 1.5, 20, "roll"))
+
+    results = []
+    for tv, cap, win, mode in GRID:
+        r = run_overlay(tv, cap, win, mode=mode)
+        results.append(r)
+        mg = r["marginal"]
+        w25 = mg.get("blends", {}).get("w25", {})
+        print(f"\n[{mode} tv={tv:.0%} cap={cap} win={win}] "
+              f"minFull(daily)={r['min_full']:+.2f} minHold={r['min_hold']:+.2f} "
+              f"max_dd={r['max_dd']*100:.0f}% (BTC {r['dd_scaled']['BTC']*100:.0f}%/"
+              f"ETH {r['dd_scaled']['ETH']*100:.0f}%  raw BTC {r['dd_raw']['BTC']*100:.0f}%/"
+              f"ETH {r['dd_raw']['ETH']*100:.0f}%) combDD={r['comb_dd']*100:.0f}%")
+        print(f"      lev BTC mean/med/atcap={r['lev_stats']['BTC']}  ETH={r['lev_stats']['ETH']}  "
+              f"scalerCausal BTC={r['scaler_caus']['BTC']['ok']} ETH={r['scaler_caus']['ETH']['ok']}")
+        print(f"      marginal: corr_full={mg.get('corr_full')} verdict={mg.get('marginal_verdict')} "
+              f"insample_edge={mg.get('has_insample_edge')} hedge={mg.get('is_hedge')} "
+              f"robust_oos={mg.get('robust_oos')} multicut={mg.get('multicut_persistent')} "
+              f"cleanYr={mg.get('clean_year_uplift')} w25upHold={w25.get('uplift_hold')}")
+
+    # ---- pick the best config: prioritize (1) max_dd<0.30, then (2) min_hold, then (3) w25 uplift_hold
+    def beats(r):
+        mg = r["marginal"]
+        w25 = mg.get("blends", {}).get("w25", {})
+        es = (mg.get("marginal_verdict") == "ADDS" and mg.get("robust_oos") is True
+              and mg.get("is_hedge") is False)
+        return (es and r["max_dd"] < 0.30
+                and (w25.get("uplift_hold") or -9) >= 0.55 and r["min_hold"] >= 0.65)
+
+    def _earns(r):
+        mg = r["marginal"]
+        return (mg.get("marginal_verdict") == "ADDS" and mg.get("robust_oos") is True
+                and mg.get("is_hedge") is False)
+
+    def score(r):
+        mg = r["marginal"]
+        w25 = mg.get("blends", {}).get("w25", {})
+        uh = w25.get("uplift_hold") or -9
+        # priority: (1) full BEATS, (2) DD<0.30 AND earns_slot AND hold>=0.65 (deployable DD-cut),
+        # (3) max w25 uplift_hold, (4) max min_hold
+        deployable = 1 if (r["max_dd"] < 0.30 and _earns(r) and r["min_hold"] >= 0.65) else 0
+        return (1 if beats(r) else 0,
+                deployable,
+                round(uh, 3),
+                round(r["min_hold"], 3))
+
+    best = max(results, key=score)
+    mg = best["marginal"]
+    w25 = mg.get("blends", {}).get("w25", {})
+    w50 = mg.get("blends", {}).get("w50", {})
+
+    fee_ok, _ = fee_survives_winner()
+    caus = sk.causality(WINNER, "BTC")
+    caus_e = sk.causality(WINNER, "ETH")
+    min_tr = winner_min_trades()
+    scaler_ok = all(best["scaler_caus"][a]["ok"] for a in ("BTC", "ETH"))
+    causality_ok = bool(caus["ok"] and caus_e["ok"] and scaler_ok)
+
+    es = (mg.get("marginal_verdict") == "ADDS" and mg.get("robust_oos") is True
+          and mg.get("is_hedge") is False)
+    earns_slot = bool(es)   # study grade for winner is PASS (it's the verified winner)
+    beats_winner = bool(earns_slot and best["max_dd"] < 0.30
+                        and (w25.get("uplift_hold") or -9) >= 0.55 and best["min_hold"] >= 0.65)
+
+    print("\n" + "=" * 90)
+    print("FINAL — BEST VOL-TARGET OVERLAY CONFIG")
+    print("=" * 90)
+    print(f"  config: mode={best['mode']} target_vol={best['target_vol']:.0%} cap={best['cap']} win={best['win']}")
+    print(f"  minFull(daily) = {best['min_full']:+.3f}")
+    print(f"  minHold(daily) = {best['min_hold']:+.3f}  (BTC {best['hold_sh']['BTC']:+.2f} / ETH {best['hold_sh']['ETH']:+.2f})")
+    print(f"  standalone max_dd (max BTC&ETH scaled) = {best['max_dd']:.4f} "
+          f"(BTC {best['dd_scaled']['BTC']:.3f} / ETH {best['dd_scaled']['ETH']:.3f})")
+    print(f"  RAW winner daily DD (no overlay)        = BTC {best['dd_raw']['BTC']:.3f} / ETH {best['dd_raw']['ETH']:.3f}")
+    print(f"  combined scaled equity max_dd           = {best['comb_dd']:.4f}")
+    print(f"  n_trades_min (winner) = {min_tr}")
+    print(f"  fee@0.30%RT survives (winner)           = {fee_ok}")
+    print(f"  causality_ok (winner entries + scaler)  = {causality_ok}  "
+          f"[winner BTC {caus} ETH {caus_e}; scaler BTC {best['scaler_caus']['BTC']} ETH {best['scaler_caus']['ETH']}]")
+    print(f"\n  MARGINAL vs TP01 (on scaled 50/50 daily):")
+    print(f"    corr_full           = {mg.get('corr_full')}")
+    print(f"    corr_hold           = {mg.get('corr_hold')}")
+    print(f"    verdict             = {mg.get('marginal_verdict')}")
+    print(f"    has_insample_edge   = {mg.get('has_insample_edge')}  cand_insample_sharpe={mg.get('cand_insample_sharpe')}")
+    print(f"    is_hedge            = {mg.get('is_hedge')}")
+    print(f"    robust_oos          = {mg.get('robust_oos')}")
+    print(f"    multicut_persistent = {mg.get('multicut_persistent')}  multicut={mg.get('multicut_uplift')}")
+    print(f"    clean_year_uplift   = {mg.get('clean_year_uplift')}  jackknife_min={mg.get('jackknife_min_uplift')}")
+    print(f"    blend w25: uplift_hold={w25.get('uplift_hold')} uplift_full={w25.get('uplift_full')} "
+          f"hold={w25.get('hold')} full={w25.get('full')} dd={w25.get('dd')}")
+    print(f"    blend w50: full={w50.get('full')} hold={w50.get('hold')} "
+          f"uplift_hold={w50.get('uplift_hold')} dd={w50.get('dd')}")
+    print(f"\n  earns_slot   = {earns_slot}")
+    print(f"  BEATS_WINNER = {beats_winner}")
+    print("=" * 90)
+
+    # machine-readable tail for the harness
+    import json
+    out = dict(
+        family="voltgt",
+        best_config=dict(strategy="winner+voltgt", mode=best["mode"],
+                         target_vol=best["target_vol"], cap=best["cap"], win=best["win"],
+                         winner=dict(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24,
+                                     uscitashort=16, vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)),
+        min_full=round(best["min_full"], 3), min_hold=round(best["min_hold"], 3),
+        max_dd=round(best["max_dd"], 4), comb_dd=round(best["comb_dd"], 4),
+        n_trades_min=min_tr, fee_ok=fee_ok, causality_ok=causality_ok,
+        corr_full=mg.get("corr_full"), verdict=mg.get("marginal_verdict"),
+        has_insample_edge=mg.get("has_insample_edge"), is_hedge=mg.get("is_hedge"),
+        robust_oos=mg.get("robust_oos"), multicut=mg.get("multicut_persistent"),
+        clean_year_uplift=mg.get("clean_year_uplift"),
+        w25_uplift_hold=w25.get("uplift_hold"),
+        earns_slot=earns_slot, beats_winner=beats_winner,
+    )
+    print("JSON " + json.dumps(out, default=str))

scripts/research/skyhook/runs/SKH_P_CHANDE.py

@@ -0,0 +1,102 @@
+"""SKH_P_CHANDE — Chande-window sweep on the V1 base.
+
+TASK: on the V1 base, sweep n_vola and n_volume (the Chande momentum windows that drive
+BuzVola=Chande01(ATR) and BuzVolume=Chande01(volume)) in {8,13,21,34,55}. Does a different
+vol/volume CYCLE window help the REGIME gate out-of-sample? Maximize min HOLD-OUT Sharpe.
+
+V1 reference: SkyhookParams(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35, vola_hi=95, vol_lo=0.0)
+  with n_vola=n_volume=13 (defaults) -> minFull +0.69, minHold +0.64 (BTC 0.64 / ETH 0.64).
+
+We keep EVERYTHING from V1 fixed (bands, pattern, exits) and vary only the two Chande windows.
+Rank by min HOLD-OUT subject to minFull>=0.5 and >=20 trades both assets, then plateau-check
+the winner (neighbors in the n_vola x n_volume grid must also be good), then full study +
+causality + marginal.
+"""
+from __future__ import annotations
+import sys, itertools
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+import skyhooklib as sk
+from src.strategies.skyhook import SkyhookParams
+
+ASSETS = ("BTC", "ETH")
+
+# V1 base: everything except the two Chande windows stays fixed.
+BASE = dict(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35, vola_hi=95, vol_lo=0.0)
+
+WINDOWS = [8, 13, 21, 34, 55]  # Fibonacci-ish cycle lengths for the Chande momentum
+
+
+def mk(n_vola, n_volume):
+    return SkyhookParams(n_vola=n_vola, n_volume=n_volume, **BASE)
+
+
+def eval_combo(n_vola, n_volume):
+    p = mk(n_vola, n_volume)
+    res = {a: sk.run_asset(a, p, sk.FEE_RT) for a in ASSETS}
+    min_full = min(res[a]["full"]["sharpe"] for a in ASSETS)
+    min_hold = min(res[a]["holdout"]["sharpe"] for a in ASSETS)
+    min_tr   = min(res[a]["full"]["n_trades"] for a in ASSETS)
+    max_dd   = max(res[a]["full"]["maxdd"] for a in ASSETS)
+    return dict(n_vola=n_vola, n_volume=n_volume,
+                min_full=min_full, min_hold=min_hold, min_tr=min_tr, max_dd=max_dd,
+                btc_hold=res["BTC"]["holdout"]["sharpe"], eth_hold=res["ETH"]["holdout"]["sharpe"],
+                btc_full=res["BTC"]["full"]["sharpe"], eth_full=res["ETH"]["full"]["sharpe"])
+
+
+def main():
+    rows = []
+    combos = list(itertools.product(WINDOWS, WINDOWS))
+    print(f"Sweeping {len(combos)} (n_vola x n_volume) combos x 2 assets = {len(combos)*2} run_asset calls")
+    for nv, nvol in combos:
+        rows.append(eval_combo(nv, nvol))
+
+    # V1 reference cell (n_vola=n_volume=13) for explicit comparison
+    v1 = next(r for r in rows if r["n_vola"] == 13 and r["n_volume"] == 13)
+    print(f"\nV1 cell (n_vola=13,n_volume=13): minF={v1['min_full']:+.2f} minH={v1['min_hold']:+.2f} "
+          f"tr={v1['min_tr']} DD={v1['max_dd']*100:.0f}%  (btcH={v1['btc_hold']:+.2f} ethH={v1['eth_hold']:+.2f})")
+
+    valid = [r for r in rows if r["min_full"] >= 0.5 and r["min_tr"] >= 20]
+    valid.sort(key=lambda r: r["min_hold"], reverse=True)
+
+    print("\n=== ALL combos by min HOLD-OUT (minF>=0.5 & tr>=20 marked OK) ===")
+    print(f"{'n_vola':>7}{'n_vol':>6}  {'minF':>6}{'minH':>6}{'minTr':>6}{'maxDD':>7}  {'btcH':>6}{'ethH':>6}  ok")
+    for r in sorted(rows, key=lambda r: r["min_hold"], reverse=True):
+        ok = "OK" if (r["min_full"] >= 0.5 and r["min_tr"] >= 20) else "."
+        mark = "  <-V1" if (r["n_vola"] == 13 and r["n_volume"] == 13) else ""
+        print(f"{r['n_vola']:>7}{r['n_volume']:>6}  {r['min_full']:>6.2f}{r['min_hold']:>6.2f}"
+              f"{r['min_tr']:>6}{r['max_dd']*100:>6.0f}%  {r['btc_hold']:>6.2f}{r['eth_hold']:>6.2f}  {ok}{mark}")
+
+    if not valid:
+        print("\nNo valid combo (minFull>=0.5 & >=20 trades). Best raw by minHold:")
+        print(sorted(rows, key=lambda r: r["min_hold"], reverse=True)[0])
+        return
+
+    top = valid[0]
+    print(f"\n=== WINNER: n_vola={top['n_vola']} n_volume={top['n_volume']} ===")
+    print(f"  minFull={top['min_full']:+.2f} minHold={top['min_hold']:+.2f} minTr={top['min_tr']} maxDD={top['max_dd']*100:.0f}%")
+
+    # plateau grid (full minHold table laid out as n_vola rows x n_volume cols)
+    def find(nv, nvol):
+        return next((r for r in rows if r["n_vola"] == nv and r["n_volume"] == nvol), None)
+    print("\n  Plateau grid (minHold; rows=n_vola, cols=n_volume):")
+    print("        " + "".join(f"{nvol:>7}" for nvol in WINDOWS))
+    for nv in WINDOWS:
+        cells = []
+        for nvol in WINDOWS:
+            r = find(nv, nvol)
+            tag = "*" if (nv == top['n_vola'] and nvol == top['n_volume']) else " "
+            cells.append(f"{r['min_hold']:>6.2f}{tag}")
+        print(f"  nv={nv:>3} " + "".join(cells))
+
+    # final study + causality + marginal on the winner
+    p = mk(top['n_vola'], top['n_volume'])
+    print("\n=== STUDY (winner) ===")
+    rep = sk.study(f"SKH_P_CHANDE_nv{top['n_vola']}_nvol{top['n_volume']}", p)
+    print(sk.fmt(rep))
+    print("\ncausality:", sk.causality(p))
+    print("\nmarginal:", sk.marginal(p))
+    print("\nas_json:", sk.as_json(rep))
+
+
+if __name__ == "__main__":
+    main()

scripts/research/skyhook/runs/SKH_P_EXITBARS.py

@@ -0,0 +1,100 @@
+"""SKH_P_EXITBARS — sweep the asymmetric time-exit horizons on the V1 base.
+
+V1 base: SkyhookParams(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35, vola_hi=95, vol_lo=0.0)
+  defaults uscitalong=24, uscitashort=18  -> minFull +0.69, HOLD +0.64 (BTC 0.64/ETH 0.64),
+  DD ~40-49% (HIGH).
+
+The asymmetry (long held longer than short) is core to Skyhook. Sweep:
+  uscitalong  in {16,20,24,30,40}  (LTF 230m bars max-hold for longs)
+  uscitashort in {10,14,18,24}     (LTF 230m bars max-hold for shorts)
+
+Objective (priority): maximize min-asset HOLD-OUT subject to minFull>=0.5, minTrades>=20 BOTH
+assets, fee survives 0.30%RT, causality ok. Secondary: cut standalone DD toward <30%.
+Compare to V1 (minHold +0.64, DD ~40-49%).
+"""
+from __future__ import annotations
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+import skyhooklib as sk
+from src.strategies.skyhook import SkyhookParams
+
+# V1 base (everything except the two exit horizons we sweep)
+BASE = dict(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+
+UL_GRID = [16, 20, 24, 30, 40]   # uscitalong
+US_GRID = [10, 14, 18, 24]       # uscitashort
+
+
+def cell(ul, us):
+    p = SkyhookParams(uscitalong=ul, uscitashort=us, **BASE)
+    out = {}
+    for a in ("BTC", "ETH"):
+        out[a] = sk.run_asset(a, p, fee_rt=sk.FEE_RT)
+    min_full = min(out[a]["full"]["sharpe"] for a in out)
+    min_hold = min(out[a]["holdout"]["sharpe"] for a in out)
+    min_tr = min(out[a]["full"]["n_trades"] for a in out)
+    max_dd = max(out[a]["full"]["maxdd"] for a in out)
+    return dict(ul=ul, us=us, min_full=min_full, min_hold=min_hold,
+                min_tr=min_tr, max_dd=max_dd,
+                btc_full=out["BTC"]["full"]["sharpe"], eth_full=out["ETH"]["full"]["sharpe"],
+                btc_hold=out["BTC"]["holdout"]["sharpe"], eth_hold=out["ETH"]["holdout"]["sharpe"],
+                btc_dd=out["BTC"]["full"]["maxdd"], eth_dd=out["ETH"]["full"]["maxdd"])
+
+
+print("=== SKH_P_EXITBARS sweep (V1 base ptn_n=55 sl2.5 tp6) — fee=0.10%RT ===")
+print(f"{'uL':>4} {'uS':>4} | {'minFull':>7} {'minHold':>7} {'minTr':>5} {'maxDD':>6} | "
+      f"{'btcF':>5} {'ethF':>5} {'btcH':>5} {'ethH':>5} {'btcDD':>5} {'ethDD':>5}")
+results = []
+for ul in UL_GRID:
+    for us in US_GRID:
+        c = cell(ul, us)
+        results.append(c)
+        flag = " *" if (c["min_full"] >= 0.5 and c["min_tr"] >= 20) else ""
+        marker = " <V1" if (ul == 24 and us == 18) else ""
+        print(f"{ul:>4} {us:>4} | {c['min_full']:>+7.2f} {c['min_hold']:>+7.2f} "
+              f"{c['min_tr']:>5} {c['max_dd']*100:>5.0f}% | "
+              f"{c['btc_full']:>+5.2f} {c['eth_full']:>+5.2f} "
+              f"{c['btc_hold']:>+5.2f} {c['eth_hold']:>+5.2f} "
+              f"{c['btc_dd']*100:>4.0f}% {c['eth_dd']*100:>4.0f}%{flag}{marker}")
+
+# Eligible = minFull>=0.5 AND minTrades>=20. Rank by min_hold, tie-break lower maxDD.
+elig = [c for c in results if c["min_full"] >= 0.5 and c["min_tr"] >= 20]
+print(f"\nEligible cells (minFull>=0.5, minTr>=20): {len(elig)}")
+if not elig:
+    print("No eligible cell — V1 exit horizons may already be at/near the frontier.")
+    sys.exit(0)
+
+elig_hold = sorted(elig, key=lambda c: (-round(c["min_hold"], 3), c["max_dd"]))
+print("Top by minHold (tie-break lower maxDD):")
+for c in elig_hold[:6]:
+    print(f"  uL={c['ul']} uS={c['us']}: minHold={c['min_hold']:+.2f} "
+          f"minFull={c['min_full']:+.2f} maxDD={c['max_dd']*100:.0f}% minTr={c['min_tr']}")
+
+dd_cands = sorted(elig, key=lambda c: (c["max_dd"], -round(c["min_hold"], 3)))
+print("\nTop by lowest maxDD (DD-cut objective):")
+for c in dd_cands[:6]:
+    print(f"  uL={c['ul']} uS={c['us']}: maxDD={c['max_dd']*100:.0f}% "
+          f"minHold={c['min_hold']:+.2f} minFull={c['min_full']:+.2f} minTr={c['min_tr']}")
+
+best = elig_hold[0]
+print(f"\n=== STUDY on best-by-minHold (uL={best['ul']} uS={best['us']}) ===")
+pbest = SkyhookParams(uscitalong=best["ul"], uscitashort=best["us"], **BASE)
+rep = sk.study(f"P_EXITBARS_uL{best['ul']}_uS{best['us']}", pbest)
+print(sk.fmt(rep))
+caus = sk.causality(pbest)
+print("causality:", caus)
+mg = sk.marginal(pbest)
+print("marginal:", {k: v for k, v in mg.items()
+                    if k in ("corr_full", "marginal_verdict", "has_insample_edge",
+                             "is_hedge", "robust_oos")})
+print("blend w25 uplift_hold:", mg.get("blends", {}).get("w25", {}).get("uplift_hold"))
+print("\nAS_JSON_STUDY:", sk.as_json(rep))
+
+# If the DD-cut frontier differs from the headline pick, study it too (cheap, one config).
+ddbest = dd_cands[0]
+if (ddbest["ul"], ddbest["us"]) != (best["ul"], best["us"]) and ddbest["min_hold"] >= 0.2:
+    print(f"\n=== STUDY on lowest-DD eligible (uL={ddbest['ul']} uS={ddbest['us']}) ===")
+    pdd = SkyhookParams(uscitalong=ddbest["ul"], uscitashort=ddbest["us"], **BASE)
+    repdd = sk.study(f"P_EXITBARS_DDcut_uL{ddbest['ul']}_uS{ddbest['us']}", pdd)
+    print(sk.fmt(repdd))
+    print("causality:", sk.causality(pdd))

scripts/research/skyhook/runs/SKH_P_LOCAL.py

@@ -0,0 +1,76 @@
+"""SKH_P_LOCAL — coordinate/local search around SKH01-V1.
+
+V1: SkyhookParams(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35, vola_hi=95, vol_lo=0.0)
+    -> minFull +0.69, minHold +0.64 (BTC0.64/ETH0.64), maxDD ~49% (BTC), clean_year(2025)=0.014
+       robust_oos=False ONLY because clean_year_uplift 0.014 < 0.02.
+
+GOAL: (a) push 2025 clean-year uplift > 0.02 (-> robust_oos True, fully earns slot),
+      (b) cut DD toward <35%, keeping minHold>=0.5, minFull>=0.5, fee survives 0.30%RT, >=20 trades.
+
+Strategy: V1's 2025 is weak (BTC+2/ETH-2). Cleaner regime gating + tighter SL can both lift the
+2025 contribution AND cut the BTC DD. Local coordinate sweep on the high-leverage knobs, each near V1.
+"""
+from __future__ import annotations
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+import skyhooklib as sk
+from src.strategies.skyhook import SkyhookParams
+
+V1 = dict(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+
+def evalp(**over):
+    p = SkyhookParams(**{**V1, **over})
+    out = {}
+    for a in ("BTC", "ETH"):
+        out[a] = sk.run_asset(a, p, fee_rt=sk.FEE_RT)
+    min_full = min(out[a]["full"]["sharpe"] for a in out)
+    min_hold = min(out[a]["holdout"]["sharpe"] for a in out)
+    min_tr   = min(out[a]["full"]["n_trades"] for a in out)
+    max_dd   = max(out[a]["full"]["maxdd"] for a in out)
+    return dict(p=p, over=over, min_full=min_full, min_hold=min_hold, min_tr=min_tr, max_dd=max_dd,
+                btc_dd=out["BTC"]["full"]["maxdd"], eth_dd=out["ETH"]["full"]["maxdd"],
+                btc_h=out["BTC"]["holdout"]["sharpe"], eth_h=out["ETH"]["holdout"]["sharpe"])
+
+def row(tag, r):
+    elig = (r["min_full"] >= 0.5 and r["min_tr"] >= 20)
+    print(f"{tag:<28} minFull={r['min_full']:+.2f} minHold={r['min_hold']:+.2f} "
+          f"maxDD={r['max_dd']*100:>3.0f}% (btc{r['btc_dd']*100:.0f}/eth{r['eth_dd']*100:.0f}) "
+          f"minTr={r['min_tr']:>3} {'OK' if elig else 'x'}")
+    return r
+
+print("=== SKH_P_LOCAL coordinate search around V1 (fee 0.10%RT) ===")
+base = row("V1", evalp())
+cands = [("V1", base)]
+
+# Axis 1: SL tighter to cut DD (V1 sl=2.5). Lower sl -> lower DD, but may cut hold.
+for sl in (1.75, 2.0, 2.25, 2.5):
+    for tp in (5.0, 6.0, 7.0):
+        cands.append((f"sl{sl}_tp{tp}", row(f"sl{sl}_tp{tp}", evalp(sl_atr=sl, tp_atr=tp))))
+
+# Axis 2: regime band — tighten vola top (avoid blow-off) & raise vola_lo to skip dead vol.
+for vlo, vhi in ((40,90),(45,90),(40,85),(35,90),(45,85),(50,90)):
+    cands.append((f"vola{vlo}-{vhi}", row(f"vola{vlo}-{vhi}", evalp(vola_lo=float(vlo), vola_hi=float(vhi)))))
+
+# Axis 3: add a volume floor (V1 vol_lo=0 = no vol gate). A floor concentrates into live regimes.
+for vol_lo in (30.0, 40.0, 50.0):
+    cands.append((f"vol_lo{vol_lo}", row(f"vol_lo{vol_lo}", evalp(vol_lo=vol_lo))))
+
+# Axis 4: ptn_n around 55.
+for ptn in (45, 50, 60, 65):
+    cands.append((f"ptn{ptn}", row(f"ptn{ptn}", evalp(ptn_n=ptn))))
+
+# Axis 5: exit bars (asymmetry).
+for ul, us in ((24,18),(30,18),(20,14),(28,14)):
+    cands.append((f"ex{ul}/{us}", row(f"ex{ul}/{us}", evalp(uscitalong=ul, uscitashort=us))))
+
+# Filter eligible (the constraints), rank by min_hold then lower DD.
+elig = [(t,r) for (t,r) in cands if r["min_full"] >= 0.5 and r["min_tr"] >= 20 and r["min_hold"] >= 0.5]
+print(f"\nEligible (minFull>=0.5, minHold>=0.5, minTr>=20): {len(elig)}")
+elig.sort(key=lambda tr: (-round(tr[1]["min_hold"],3), tr[1]["max_dd"]))
+for t,r in elig[:10]:
+    print(f"  {t:<22} minHold={r['min_hold']:+.2f} minFull={r['min_full']:+.2f} maxDD={r['max_dd']*100:.0f}% over={r['over']}")
+
+# Low-DD subset
+print("\nLowest-DD eligible:")
+for t,r in sorted(elig, key=lambda tr:(tr[1]["max_dd"], -tr[1]["min_hold"]))[:8]:
+    print(f"  {t:<22} maxDD={r['max_dd']*100:.0f}% minHold={r['min_hold']:+.2f} minFull={r['min_full']:+.2f} over={r['over']}")

scripts/research/skyhook/runs/SKH_P_LOCAL2.py

@@ -0,0 +1,65 @@
+"""SKH_P_LOCAL2 — refine: combine the two winning axes (ptn_n DD-cut + sl/tp hold-lift)
+and CHECK MARGINAL clean-year(2025) uplift on the top few, since that is the true gate
+(robust_oos requires clean_year_uplift>0.02 AND multicut_persistent)."""
+from __future__ import annotations
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+import skyhooklib as sk
+from src.strategies.skyhook import SkyhookParams
+
+V1 = dict(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+
+def evalp(**over):
+    p = SkyhookParams(**{**V1, **over})
+    out = {a: sk.run_asset(a, p, fee_rt=sk.FEE_RT) for a in ("BTC","ETH")}
+    return dict(p=p, over=over,
+                min_full=min(out[a]["full"]["sharpe"] for a in out),
+                min_hold=min(out[a]["holdout"]["sharpe"] for a in out),
+                min_tr=min(out[a]["full"]["n_trades"] for a in out),
+                max_dd=max(out[a]["full"]["maxdd"] for a in out),
+                btc_dd=out["BTC"]["full"]["maxdd"], eth_dd=out["ETH"]["full"]["maxdd"])
+
+def row(tag, r):
+    elig = (r["min_full"]>=0.5 and r["min_tr"]>=20 and r["min_hold"]>=0.5)
+    print(f"{tag:<26} minFull={r['min_full']:+.2f} minHold={r['min_hold']:+.2f} "
+          f"maxDD={r['max_dd']*100:>3.0f}% (b{r['btc_dd']*100:.0f}/e{r['eth_dd']*100:.0f}) "
+          f"minTr={r['min_tr']:>3} {'OK' if elig else 'x'}")
+    return r
+
+print("=== SKH_P_LOCAL2 combine ptn_n x sl/tp (fee 0.10%RT) ===")
+cands=[]
+for ptn in (40, 45, 48):
+    for sl,tp in ((2.0,6.0),(2.0,7.0),(2.25,6.0),(2.25,7.0),(2.5,7.0)):
+        over=dict(ptn_n=ptn, sl_atr=sl, tp_atr=tp)
+        cands.append((f"ptn{ptn}_sl{sl}_tp{tp}", row(f"ptn{ptn}_sl{sl}_tp{tp}", evalp(**over))))
+# also ptn45 with exit-bar asymmetry that lifted hold
+for ul,us in ((30,18),(24,18)):
+    over=dict(ptn_n=45, uscitalong=ul, uscitashort=us)
+    cands.append((f"ptn45_ex{ul}/{us}", row(f"ptn45_ex{ul}/{us}", evalp(**over))))
+
+elig=[(t,r) for (t,r) in cands if r["min_full"]>=0.5 and r["min_tr"]>=20 and r["min_hold"]>=0.5]
+elig.sort(key=lambda tr:(tr[1]["max_dd"], -round(tr[1]["min_hold"],3)))
+print(f"\nEligible: {len(elig)} (sorted by lowest DD)")
+for t,r in elig:
+    print(f"  {t:<24} maxDD={r['max_dd']*100:.0f}% minHold={r['min_hold']:+.2f} minFull={r['min_full']:+.2f} over={r['over']}")
+
+# Check MARGINAL clean-year uplift on the lowest-DD eligible + the best-hold eligible.
+def marg_check(tag, over):
+    p = SkyhookParams(**{**V1, **over})
+    mg = sk.marginal(p)
+    print(f"\n--- MARGINAL {tag} over={over} ---")
+    print(f"  verdict={mg.get('marginal_verdict')} corr_full={mg.get('corr_full')} "
+          f"robust_oos={mg.get('robust_oos')} multicut_persistent={mg.get('multicut_persistent')}")
+    print(f"  clean_year_uplift={mg.get('clean_year_uplift')} jackknife_min={mg.get('jackknife_min_uplift')} "
+          f"has_insample_edge={mg.get('has_insample_edge')} is_hedge={mg.get('is_hedge')}")
+    print(f"  blend w25 uplift_hold={mg.get('blends',{}).get('w25',{}).get('uplift_hold')} "
+          f"uplift_full={mg.get('blends',{}).get('w25',{}).get('uplift_full')}")
+    return mg
+
+# pick up to 4 distinct configs to marginal-check
+seen=set(); checked=0
+for t,r in elig:
+    key=tuple(sorted(r["over"].items()))
+    if key in seen: continue
+    seen.add(key); marg_check(t, r["over"]); checked+=1
+    if checked>=5: break

scripts/research/skyhook/runs/SKH_P_LOCAL_final.py

@@ -0,0 +1,48 @@
+"""SKH_P_LOCAL_final — full study + causality + marginal on the top local-search winners,
+plus a small extra pass trying to push DD<35% while keeping minHold high (ptn45 + tighter exits)."""
+from __future__ import annotations
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+import skyhooklib as sk
+from src.strategies.skyhook import SkyhookParams
+
+V1 = dict(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+
+def evalp(**over):
+    p = SkyhookParams(**{**V1, **over})
+    out = {a: sk.run_asset(a, p, fee_rt=sk.FEE_RT) for a in ("BTC","ETH")}
+    return dict(over=over,
+                min_full=min(out[a]["full"]["sharpe"] for a in out),
+                min_hold=min(out[a]["holdout"]["sharpe"] for a in out),
+                min_tr=min(out[a]["full"]["n_trades"] for a in out),
+                max_dd=max(out[a]["full"]["maxdd"] for a in out))
+
+# Small extra: ptn45 + tp7 + tighter SL or exit bars to chase DD<35 with hold>=0.5
+print("=== extra DD-chase around ptn45 ===")
+for over in (dict(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16),
+             dict(ptn_n=45, sl_atr=2.25, tp_atr=7.0, uscitalong=24, uscitashort=16),
+             dict(ptn_n=45, sl_atr=2.5, tp_atr=7.0, vola_lo=40.0),
+             dict(ptn_n=45, sl_atr=2.5, tp_atr=8.0)):
+    r=evalp(**over)
+    print(f"  {over} -> minFull={r['min_full']:+.2f} minHold={r['min_hold']:+.2f} maxDD={r['max_dd']*100:.0f}% minTr={r['min_tr']}")
+
+# WINNER candidates -> full study
+WINNERS = {
+  "P_LOCAL_ptn45_sl2.5_tp7.0":  dict(ptn_n=45, sl_atr=2.5,  tp_atr=7.0),   # best balanced (DD36, Full0.83, Hold0.69)
+  "P_LOCAL_ptn45_sl2.25_tp7.0": dict(ptn_n=45, sl_atr=2.25, tp_atr=7.0),   # best hold/clean-year (DD36, Hold0.77)
+}
+for name, over in WINNERS.items():
+    p = SkyhookParams(**{**V1, **over})
+    print(f"\n################ STUDY {name} over={over} ################")
+    rep = sk.study(name, p)
+    print(sk.fmt(rep))
+    print("causality:", sk.causality(p))
+    mg = sk.marginal(p)
+    keys=("corr_full","corr_hold","marginal_verdict","has_insample_edge","is_hedge",
+          "robust_oos","multicut_persistent","clean_year_uplift","jackknife_min_uplift",
+          "cand_insample_sharpe","cand_full_sharpe")
+    print("marginal:", {k: mg.get(k) for k in keys})
+    print("blend w25 uplift_hold:", mg.get("blends",{}).get("w25",{}).get("uplift_hold"),
+          "uplift_full:", mg.get("blends",{}).get("w25",{}).get("uplift_full"))
+    print("multicut:", mg.get("multicut_uplift"))
+    print("AS_JSON:", sk.as_json(rep))

scripts/research/skyhook/runs/SKH_P_LOCAL_v1ref.py

@@ -0,0 +1,18 @@
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+import skyhooklib as sk
+from src.strategies.skyhook import SkyhookParams
+
+V1 = SkyhookParams(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+print("=== V1 reference ===")
+rep = sk.study("SKH01-V1", V1)
+print(sk.fmt(rep))
+print("causality:", sk.causality(V1))
+mg = sk.marginal(V1)
+keys = ("corr_full","corr_hold","marginal_verdict","has_insample_edge","is_hedge",
+        "robust_oos","multicut_persistent","clean_year_uplift","jackknife_min_uplift",
+        "cand_insample_sharpe","cand_full_sharpe")
+print("marginal:", {k: mg.get(k) for k in keys})
+print("blend w25 uplift_hold:", mg.get("blends",{}).get("w25",{}).get("uplift_hold"))
+print("blend w25 uplift_full:", mg.get("blends",{}).get("w25",{}).get("uplift_full"))
+print("multicut:", mg.get("multicut_uplift"))

scripts/research/skyhook/runs/SKH_P_LOCAL_winner.py

@@ -0,0 +1,26 @@
+"""SKH_P_LOCAL winner: ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16
+(rest = V1). Beats V1 on DD (34% vs 49%), minHold (+0.81 vs +0.64), minFull (+0.83 vs +0.69),
+and pushes clean-year(2025) uplift well over 0.02 -> robust_oos True (fully earns a slot)."""
+from __future__ import annotations
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+import skyhooklib as sk
+from src.strategies.skyhook import SkyhookParams
+
+V1 = dict(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+WIN = SkyhookParams(**{**V1, **dict(ptn_n=45, sl_atr=2.5, tp_atr=7.0, uscitalong=24, uscitashort=16)})
+
+rep = sk.study("SKH_P_LOCAL_winner", WIN)
+print(sk.fmt(rep))
+print("causality:", sk.causality(WIN, asset="BTC"), sk.causality(WIN, asset="ETH"))
+mg = sk.marginal(WIN)
+keys=("corr_full","corr_hold","marginal_verdict","has_insample_edge","is_hedge",
+      "robust_oos","multicut_persistent","clean_year_uplift","jackknife_min_uplift",
+      "cand_insample_sharpe","cand_full_sharpe","beta_to_tp01","alpha_ann")
+print("marginal:", {k: mg.get(k) for k in keys})
+print("blend w25 uplift_hold:", mg.get("blends",{}).get("w25",{}).get("uplift_hold"),
+      "uplift_full:", mg.get("blends",{}).get("w25",{}).get("uplift_full"))
+print("blend w50:", mg.get("blends",{}).get("w50"))
+print("multicut:", mg.get("multicut_uplift"))
+v=rep["verdict"]
+print("\nVERDICT:", v)

scripts/research/skyhook/runs/SKH_P_PTN.py

@@ -0,0 +1,82 @@
+"""SKH_P_PTN  (FAMILY=param)
+On the SKH01-V1 base, sweep ptn_n in {34,45,55,70,89,110} x atr_win in {10,14,21}.
+Slower Donchian breakouts may generalize better OOS. Maximize min-asset HOLD-OUT
+subject to minFull>=0.5, fee survives 0.30%RT, >=20 trades BOTH assets, causality ok.
+Note standalone DD. Always compare vs V1 (ptn_n=55, atr_win=14).
+"""
+import sys
+import itertools
+from dataclasses import replace
+
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+import skyhooklib as sk
+from src.strategies.skyhook import SkyhookParams
+
+# SKH01-V1 reference base
+V1 = SkyhookParams(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35, vola_hi=95, vol_lo=0.0)
+
+
+def quick(p: SkyhookParams) -> dict:
+    rs = {a: sk.run_asset(a, p, sk.FEE_RT) for a in sk.CERTIFIED}
+    mf = min(rs[a]["full"]["sharpe"] for a in rs)
+    mh = min(rs[a]["holdout"]["sharpe"] for a in rs)
+    mt = min(rs[a]["full"]["n_trades"] for a in rs)
+    avg_dd = sum(rs[a]["full"]["maxdd"] for a in rs) / 2
+    return dict(minFull=mf, minHold=mh, minTr=mt, dd=round(avg_dd, 4),
+                btc=rs["BTC"]["full"]["sharpe"], eth=rs["ETH"]["full"]["sharpe"],
+                btcH=rs["BTC"]["holdout"]["sharpe"], ethH=rs["ETH"]["holdout"]["sharpe"],
+                btcDD=rs["BTC"]["full"]["maxdd"], ethDD=rs["ETH"]["full"]["maxdd"])
+
+
+PTN_GRID = (34, 45, 55, 70, 89, 110)
+ATR_GRID = (10, 14, 21)
+
+print("=== SKH_P_PTN sweep: ptn_n x atr_win on SKH01-V1 base ===")
+qv1 = quick(V1)
+print(f"V1 (ptn55/atr14): minF={qv1['minFull']:+.2f} minH={qv1['minHold']:+.2f} "
+      f"btc/eth F={qv1['btc']:+.2f}/{qv1['eth']:+.2f} H={qv1['btcH']:+.2f}/{qv1['ethH']:+.2f} "
+      f"tr={qv1['minTr']} dd~{qv1['dd']*100:.0f}% (btc{qv1['btcDD']*100:.0f}/eth{qv1['ethDD']*100:.0f})")
+print("-" * 108)
+print(f"{'ptn':>4s}{'atr':>4s}  {'minF':>6s}{'minH':>6s}  {'btcF/ethF':>13s} {'btcH/ethH':>13s} "
+      f"{'tr':>4s} {'avgDD':>6s} {'btcDD/ethDD':>12s}  gate")
+
+rows = []
+for ptn_n, atr_win in itertools.product(PTN_GRID, ATR_GRID):
+    p = replace(V1, ptn_n=ptn_n, atr_win=atr_win)
+    q = quick(p)
+    # gate per task: minFull>=0.5 AND minHold>=0.2 AND minTr>=20
+    gate = (q["minFull"] >= 0.5 and q["minHold"] >= 0.2 and q["minTr"] >= 20)
+    rows.append((q["minHold"], q["minFull"], q["minTr"], q["dd"], ptn_n, atr_win, q, gate))
+    tag = "PASS" if gate else ""
+    print(f"{ptn_n:>4d}{atr_win:>4d}  {q['minFull']:>+6.2f}{q['minHold']:>+6.2f}  "
+          f"{q['btc']:>+5.2f}/{q['eth']:>+5.2f} {q['btcH']:>+5.2f}/{q['ethH']:>+5.2f} "
+          f"{q['minTr']:>4d} {q['dd']*100:>5.0f}% {q['btcDD']*100:>4.0f}/{q['ethDD']*100:>4.0f}%  {tag}")
+
+# winner = max min-asset HOLD-OUT among gate-passers (minFull>=0.5, minTr>=20); fallback best minHold
+passers = [r for r in rows if r[7]]
+pool = passers if passers else [r for r in rows if r[1] >= 0.5 and r[2] >= 20]
+if not pool:
+    pool = rows
+# rank by minHold, tiebreak lower avgDD then higher minFull
+pool.sort(key=lambda r: (r[0], -r[3], r[1]), reverse=True)
+best = pool[0]
+b_ptn, b_atr = best[4], best[5]
+print("-" * 108)
+print(f"WINNER: ptn_n={b_ptn} atr_win={b_atr}  minH={best[0]:+.2f} minF={best[1]:+.2f} "
+      f"tr={best[2]} avgDD={best[3]*100:.0f}%")
+
+# Full study + causality + marginal on winner (and re-confirm V1 alongside)
+WIN = replace(V1, ptn_n=b_ptn, atr_win=b_atr)
+print("\n=== STUDY winner ===")
+rep = sk.study(f"SKH_P_PTN ptn{b_ptn}/atr{b_atr}", WIN)
+print(sk.fmt(rep))
+caus = sk.causality(WIN, "BTC")
+caus_eth = sk.causality(WIN, "ETH")
+print(f"causality BTC: {caus}  ETH: {caus_eth}")
+mg = sk.marginal(WIN)
+print(f"marginal: corr_full={mg.get('corr_full')} "
+      f"blend_w25_uplift_hold={mg.get('blends', {}).get('w25', {}).get('uplift_hold')} "
+      f"verdict={mg.get('marginal_verdict')} has_insample_edge={mg.get('has_insample_edge')} "
+      f"is_hedge={mg.get('is_hedge')}")
+print("\nJSON_STUDY:", sk.as_json(rep))
+print("MARGINAL:", mg)

scripts/research/skyhook/runs/SKH_P_REGIME.py

@@ -0,0 +1,107 @@
+"""SKH_P_REGIME — regime-band sweep on the V1 base.
+
+Search bands: vola_lo in {20,30,35,45}, vola_hi in {88,95,100},
+              vol_lo in {0,30,45,55},  vol_hi in {80,100}.
+Find the combo that lifts min HOLD-OUT and is a PLATEAU (neighbors also good).
+Compare to V1: SkyhookParams(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35, vola_hi=95, vol_lo=0.0)
+  -> minFull +0.69, minHold +0.64.
+"""
+from __future__ import annotations
+import sys, itertools
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+import skyhooklib as sk
+from src.strategies.skyhook import SkyhookParams
+
+ASSETS = ("BTC", "ETH")
+
+# V1 base (everything except the regime bands stays fixed)
+BASE = dict(ptn_n=55, sl_atr=2.5, tp_atr=6.0)
+
+VOLA_LO = [20, 30, 35, 45]
+VOLA_HI = [88, 95, 100]
+VOL_LO  = [0, 30, 45, 55]
+VOL_HI  = [80, 100]
+
+
+def mk(vlo, vhi, vol_lo, vol_hi):
+    return SkyhookParams(vola_lo=vlo, vola_hi=vhi, vol_lo=vol_lo, vol_hi=vol_hi, **BASE)
+
+
+def eval_combo(vlo, vhi, vol_lo, vol_hi):
+    p = mk(vlo, vhi, vol_lo, vol_hi)
+    res = {}
+    for a in ASSETS:
+        r = sk.run_asset(a, p, sk.FEE_RT)
+        res[a] = r
+    min_full = min(res[a]["full"]["sharpe"] for a in ASSETS)
+    min_hold = min(res[a]["holdout"]["sharpe"] for a in ASSETS)
+    min_tr   = min(res[a]["full"]["n_trades"] for a in ASSETS)
+    max_dd   = max(res[a]["full"]["maxdd"] for a in ASSETS)
+    return dict(vlo=vlo, vhi=vhi, vol_lo=vol_lo, vol_hi=vol_hi,
+                min_full=min_full, min_hold=min_hold, min_tr=min_tr, max_dd=max_dd,
+                btc_hold=res["BTC"]["holdout"]["sharpe"], eth_hold=res["ETH"]["holdout"]["sharpe"],
+                btc_full=res["BTC"]["full"]["sharpe"], eth_full=res["ETH"]["full"]["sharpe"])
+
+
+def main():
+    rows = []
+    combos = list(itertools.product(VOLA_LO, VOLA_HI, VOL_LO, VOL_HI))
+    print(f"Sweeping {len(combos)} band combos x 2 assets = {len(combos)*2} run_asset calls")
+    for vlo, vhi, vol_lo, vol_hi in combos:
+        rows.append(eval_combo(vlo, vhi, vol_lo, vol_hi))
+
+    # rank by min HOLD-OUT (subject to minFull>=0.5 and >=20 trades both assets)
+    valid = [r for r in rows if r["min_full"] >= 0.5 and r["min_tr"] >= 20]
+    valid.sort(key=lambda r: r["min_hold"], reverse=True)
+
+    print("\n=== TOP 15 by min HOLD-OUT (minFull>=0.5, minTrades>=20) ===")
+    print(f"{'vlo':>4}{'vhi':>5}{'vol_lo':>7}{'vol_hi':>7}  {'minF':>6}{'minH':>6}{'minTr':>6}{'maxDD':>7}  {'btcH':>6}{'ethH':>6}")
+    for r in valid[:15]:
+        print(f"{r['vlo']:>4}{r['vhi']:>5}{r['vol_lo']:>7}{r['vol_hi']:>7}  "
+              f"{r['min_full']:>6.2f}{r['min_hold']:>6.2f}{r['min_tr']:>6}{r['max_dd']*100:>6.0f}%  "
+              f"{r['btc_hold']:>6.2f}{r['eth_hold']:>6.2f}")
+
+    # full table sorted for plateau inspection (group by lo/hi neighbors)
+    rows_sorted = sorted(rows, key=lambda r: r["min_hold"], reverse=True)
+    print("\n=== ALL combos by min HOLD-OUT ===")
+    for r in rows_sorted:
+        flag = "" if (r["min_full"] >= 0.5 and r["min_tr"] >= 20) else "  (low-full/trades)"
+        print(f"  vlo={r['vlo']:>3} vhi={r['vhi']:>3} vol_lo={r['vol_lo']:>3} vol_hi={r['vol_hi']:>3} | "
+              f"minF={r['min_full']:+.2f} minH={r['min_hold']:+.2f} tr={r['min_tr']:>3} DD={r['max_dd']*100:.0f}%{flag}")
+
+    if not valid:
+        print("\nNo valid combo (minFull>=0.5 & >=20 trades). Best raw:")
+        print(rows_sorted[0])
+        return
+
+    # plateau check: for the top combo, look at neighbors in the grid
+    top = valid[0]
+    print(f"\n=== WINNER: vlo={top['vlo']} vhi={top['vhi']} vol_lo={top['vol_lo']} vol_hi={top['vol_hi']} ===")
+    print(f"  minFull={top['min_full']:+.2f} minHold={top['min_hold']:+.2f} minTr={top['min_tr']} maxDD={top['max_dd']*100:.0f}%")
+
+    # neighbor plateau: same vol_lo/vol_hi, vary vola_lo/vola_hi to adjacent grid values
+    def find(vlo, vhi, vol_lo, vol_hi):
+        for r in rows:
+            if r["vlo"]==vlo and r["vhi"]==vhi and r["vol_lo"]==vol_lo and r["vol_hi"]==vol_hi:
+                return r
+        return None
+    print("\n  Plateau neighbors (min HOLD-OUT):")
+    for vlo in VOLA_LO:
+        for vhi in VOLA_HI:
+            r = find(vlo, vhi, top['vol_lo'], top['vol_hi'])
+            if r:
+                mark = " <-- WIN" if (vlo==top['vlo'] and vhi==top['vhi']) else ""
+                print(f"    vola_lo={vlo:>3} vola_hi={vhi:>3}: minH={r['min_hold']:+.2f} minF={r['min_full']:+.2f}{mark}")
+
+    # final study + causality + marginal on the winner
+    p = mk(top['vlo'], top['vhi'], top['vol_lo'], top['vol_hi'])
+    print("\n=== STUDY (winner) ===")
+    rep = sk.study(f"SKH_P_REGIME_vlo{top['vlo']}_vhi{top['vhi']}_vollo{top['vol_lo']}_volhi{top['vol_hi']}", p)
+    print(sk.fmt(rep))
+    print("\ncausality:", sk.causality(p))
+    print("\nmarginal:", sk.marginal(p))
+    print("\nas_json:", sk.as_json(rep))
+
+
+if __name__ == "__main__":
+    main()

scripts/research/skyhook/runs/SKH_P_REGIME_plateau.py

@@ -0,0 +1,46 @@
+"""SKH_P_REGIME_plateau — tight plateau probe around the sweep winner
+vola_lo=20, vola_hi=88, vol_lo=55, vol_hi=80 (V1 base: ptn_n=55, sl_atr=2.5, tp_atr=6.0).
+
+Confirm neighbors in ALL 4 band dims are also good (no knife-edge).
+"""
+from __future__ import annotations
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+import skyhooklib as sk
+from src.strategies.skyhook import SkyhookParams
+
+ASSETS = ("BTC", "ETH")
+BASE = dict(ptn_n=55, sl_atr=2.5, tp_atr=6.0)
+WIN = dict(vola_lo=20, vola_hi=88, vol_lo=55, vol_hi=80)
+
+
+def ev(**bands):
+    p = SkyhookParams(**{**BASE, **bands})
+    res = {a: sk.run_asset(a, p, sk.FEE_RT) for a in ASSETS}
+    mf = min(res[a]["full"]["sharpe"] for a in ASSETS)
+    mh = min(res[a]["holdout"]["sharpe"] for a in ASSETS)
+    tr = min(res[a]["full"]["n_trades"] for a in ASSETS)
+    dd = max(res[a]["full"]["maxdd"] for a in ASSETS)
+    return mf, mh, tr, dd
+
+
+def probe(dim, values):
+    print(f"\n-- perturb {dim} (others at winner) --")
+    for v in values:
+        bands = dict(WIN); bands[dim] = v
+        mf, mh, tr, dd = ev(**bands)
+        mark = " <-- WIN" if v == WIN[dim] else ""
+        print(f"   {dim}={v:>4}: minF={mf:+.2f} minH={mh:+.2f} tr={tr:>3} DD={dd*100:.0f}%{mark}")
+
+
+def main():
+    mf, mh, tr, dd = ev(**WIN)
+    print(f"WINNER {WIN}: minF={mf:+.2f} minH={mh:+.2f} tr={tr} DD={dd*100:.0f}%")
+    probe("vola_lo", [15, 20, 25, 30])
+    probe("vola_hi", [83, 85, 88, 90, 92])
+    probe("vol_lo",  [45, 50, 55, 60, 65])
+    probe("vol_hi",  [75, 78, 80, 82, 85])
+
+
+if __name__ == "__main__":
+    main()

scripts/research/skyhook/runs/SKH_P_RR.py

@@ -0,0 +1,87 @@
+"""SKH_P_RR — fine-sweep reward:risk on the ptn_n=55 V1 base.
+
+V1 base: SkyhookParams(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35, vola_hi=95, vol_lo=0.0)
+  -> minFull +0.69, HOLD +0.64 (BTC 0.64 / ETH 0.64), DD ~40-49% (HIGH).
+
+Sweep: sl_atr in {2.0,2.25,2.5,2.75,3.0,3.5} x tp_atr in {5,6,7,8,9,10}.
+Objective: maximize min-asset HOLD-OUT subject to minFull>=0.5, cut DD. Report best + plateau.
+"""
+from __future__ import annotations
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+import skyhooklib as sk
+from src.strategies.skyhook import SkyhookParams
+
+BASE = dict(ptn_n=55, vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+
+SL_GRID = [2.0, 2.25, 2.5, 2.75, 3.0, 3.5]
+TP_GRID = [5.0, 6.0, 7.0, 8.0, 9.0, 10.0]
+
+def cell(sl, tp):
+    p = SkyhookParams(sl_atr=sl, tp_atr=tp, **BASE)
+    out = {}
+    for a in ("BTC", "ETH"):
+        r = sk.run_asset(a, p, fee_rt=sk.FEE_RT)
+        out[a] = r
+    min_full = min(out[a]["full"]["sharpe"] for a in out)
+    min_hold = min(out[a]["holdout"]["sharpe"] for a in out)
+    min_tr = min(out[a]["full"]["n_trades"] for a in out)
+    max_dd = max(out[a]["full"]["maxdd"] for a in out)
+    return dict(sl=sl, tp=tp, min_full=min_full, min_hold=min_hold,
+                min_tr=min_tr, max_dd=max_dd,
+                btc_full=out["BTC"]["full"]["sharpe"], eth_full=out["ETH"]["full"]["sharpe"],
+                btc_hold=out["BTC"]["holdout"]["sharpe"], eth_hold=out["ETH"]["holdout"]["sharpe"],
+                btc_dd=out["BTC"]["full"]["maxdd"], eth_dd=out["ETH"]["full"]["maxdd"])
+
+print("=== SKH_P_RR sweep (ptn_n=55 base) — fee=0.10%RT ===")
+print(f"{'sl':>5} {'tp':>5} | {'minFull':>7} {'minHold':>7} {'minTr':>5} {'maxDD':>6} | "
+      f"{'btcF':>5} {'ethF':>5} {'btcH':>5} {'ethH':>5} {'btcDD':>5} {'ethDD':>5}")
+results = []
+for sl in SL_GRID:
+    for tp in TP_GRID:
+        if tp <= sl:  # tp must exceed sl for a sensible R:R; skip degenerate
+            continue
+        c = cell(sl, tp)
+        results.append(c)
+        flag = ""
+        if c["min_full"] >= 0.5 and c["min_tr"] >= 20:
+            flag = " *"  # eligible
+        print(f"{sl:>5} {tp:>5} | {c['min_full']:>+7.2f} {c['min_hold']:>+7.2f} "
+              f"{c['min_tr']:>5} {c['max_dd']*100:>5.0f}% | "
+              f"{c['btc_full']:>+5.2f} {c['eth_full']:>+5.2f} "
+              f"{c['btc_hold']:>+5.2f} {c['eth_hold']:>+5.2f} "
+              f"{c['btc_dd']*100:>4.0f}% {c['eth_dd']*100:>4.0f}%{flag}")
+
+# Eligible = minFull>=0.5, minTrades>=20. Rank by min_hold, tie-break lower maxDD.
+elig = [c for c in results if c["min_full"] >= 0.5 and c["min_tr"] >= 20]
+print(f"\nEligible cells (minFull>=0.5, minTr>=20): {len(elig)}")
+if elig:
+    elig_sorted = sorted(elig, key=lambda c: (-round(c["min_hold"], 3), c["max_dd"]))
+    print("Top by minHold (tie-break lower maxDD):")
+    for c in elig_sorted[:6]:
+        print(f"  sl={c['sl']} tp={c['tp']}: minHold={c['min_hold']:+.2f} "
+              f"minFull={c['min_full']:+.2f} maxDD={c['max_dd']*100:.0f}% minTr={c['min_tr']}")
+    best = elig_sorted[0]
+    # DD-cutting candidate: best minHold among cells with maxDD < V1-ish (lower DD priority)
+    dd_cands = sorted(elig, key=lambda c: (c["max_dd"], -round(c["min_hold"], 3)))
+    print("\nTop by lowest maxDD (DD-cut objective):")
+    for c in dd_cands[:6]:
+        print(f"  sl={c['sl']} tp={c['tp']}: maxDD={c['max_dd']*100:.0f}% "
+              f"minHold={c['min_hold']:+.2f} minFull={c['min_full']:+.2f} minTr={c['min_tr']}")
+
+    print("\n=== STUDY on best-by-minHold ===")
+    pbest = SkyhookParams(sl_atr=best["sl"], tp_atr=best["tp"], **BASE)
+    rep = sk.study(f"P_RR_sl{best['sl']}_tp{best['tp']}", pbest)
+    print(sk.fmt(rep))
+    print("causality:", sk.causality(pbest))
+    print("marginal:", {k: v for k, v in sk.marginal(pbest).items()
+                        if k in ("corr_full","marginal_verdict","has_insample_edge","is_hedge","robust_oos")})
+    try:
+        mg = sk.marginal(pbest)
+        print("marginal-full-keys:", list(mg.keys()))
+        print("blend w25 uplift_hold:", mg.get("blends",{}).get("w25",{}).get("uplift_hold"))
+    except Exception as e:
+        print("marginal err:", e)
+    print("\nAS_JSON_STUDY:", sk.as_json(rep))
+else:
+    print("No eligible cell — V1 base may already be at the frontier.")

scripts/research/skyhook/runs/SKH_R_EXPAND.py

@@ -0,0 +1,306 @@
+"""SKH_R_EXPAND — REGIME variant: VOLATILITY-EXPANSION gate.
+
+Hypothesis (the brief: "enter when vol+volume regime AND breakout coincide"):
+Instead of the Chande01 *cycle* band on ATR, define the regime as a genuine VOLATILITY
+EXPANSION: trade only when ATR is RISING vs its own moving average (a vol breakout) AND
+volume is elevated vs its own moving average. The intuition is that a Donchian breakout that
+fires WHILE volatility is expanding on rising participation (volume) is more likely to be a
+real move than one that fires inside a quiet/contracting regime (chop, mean-reversion).
+
+Regime definition (HTF, causal):
+  vol_expansion = ATR[i] >= k_atr * MA(ATR, w_atr)         (ATR above its own MA -> rising)
+  volume_elev   = volume[i] >= k_vol * MA(volume, w_vol)   (participation elevated)
+  regime_ok     = vol_expansion AND volume_elev
+MA is a CAUSAL rolling mean (uses x[i-w+1..i] inclusive of the current, already-closed bar).
+k_atr / k_vol are tunable multipliers (1.0 = "above MA"; >1 = "well above MA").
+
+Pattern/composer/entry/exit reuse the V1 Skyhook building blocks unchanged:
+  ptn_n=55 Donchian, sl_atr=2.5, tp_atr=6.0, asymmetric time exits, max 1/day.
+
+Causality: every regime feature uses only x[0..i] (rolling MA, ATR ewm, donchian shift(1)),
+INCLUSIVE of the current HTF bar — legit because at HTF close[i] the bar is fully known. The
+HTF feature is merged BACKWARD onto LTF on the HTF-close timestamp (<= LTF close). We verify
+the truncated-prefix guard ourselves on BOTH assets.
+"""
+from __future__ import annotations
+
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+
+import numpy as np
+import pandas as pd
+
+import skyhooklib as sk
+from src.backtest.harness import backtest_signals
+from src.strategies import skyhook as S
+from src.strategies.skyhook import SkyhookParams, HTF_MIN
+
+HOLDOUT = pd.Timestamp("2025-01-01", tz="UTC")
+FEE = sk.FEE_RT
+
+
+# ---------------------------------------------------------------------------
+# Causal rolling MA (inclusive of current, already-closed bar). min_periods enforced.
+# ---------------------------------------------------------------------------
+def causal_ma(x: np.ndarray, win: int, min_periods: int | None = None) -> np.ndarray:
+    mp = win if min_periods is None else min_periods
+    return pd.Series(np.asarray(x, float)).rolling(win, min_periods=mp).mean().values
+
+
+# ---------------------------------------------------------------------------
+# HTF feature df: volatility-EXPANSION regime gate + Donchian pattern (V1 pattern reused).
+#   regime_ok = (ATR >= k_atr*MA(ATR,w_atr)) AND (volume >= k_vol*MA(volume,w_vol))
+# ---------------------------------------------------------------------------
+def expand_htf_features(htf: pd.DataFrame, p: SkyhookParams,
+                        w_atr: int, k_atr: float,
+                        w_vol: int, k_vol: float) -> pd.DataFrame:
+    atr_htf = S.atr(htf, p.atr_win)
+    vol_htf = htf["volume"].values.astype(float)
+
+    atr_ma = causal_ma(atr_htf, w_atr)
+    vol_ma = causal_ma(vol_htf, w_vol)
+
+    # rising-vol = current ATR above k_atr * its own MA ; same for volume.
+    # NaN during warmup -> False (no trade until the regime is computable).
+    vol_expansion = np.where(np.isfinite(atr_ma) & (atr_ma > 0), atr_htf >= k_atr * atr_ma, False)
+    volume_elev = np.where(np.isfinite(vol_ma) & (vol_ma > 0), vol_htf >= k_vol * vol_ma, False)
+    regime_ok = vol_expansion & volume_elev
+
+    ptn_long, ptn_short = S.donchian_breakout(htf, p.ptn_n)
+    comp_long = regime_ok & ptn_long
+    comp_short = regime_ok & ptn_short
+    if p.long_only:
+        comp_short = np.zeros_like(comp_short, dtype=bool)
+
+    close_ts = htf["timestamp"].astype("int64").values + HTF_MIN * 60 * 1000
+    return pd.DataFrame({
+        "close_ts": close_ts,
+        # store the ratios for diagnostics (not used downstream)
+        "buz_vola": np.where(np.isfinite(atr_ma) & (atr_ma > 0), atr_htf / atr_ma, np.nan),
+        "buz_volume": np.where(np.isfinite(vol_ma) & (vol_ma > 0), vol_htf / vol_ma, np.nan),
+        "comp_long": comp_long.astype(bool), "comp_short": comp_short.astype(bool),
+    })
+
+
+def expand_entries(ltf: pd.DataFrame, htf: pd.DataFrame, p: SkyhookParams,
+                   w_atr, k_atr, w_vol, k_vol) -> list:
+    """Same entry/exit machinery as S.skyhook_entries, regime from expansion features."""
+    feat = expand_htf_features(htf, p, w_atr, k_atr, w_vol, k_vol)
+    m = S.merge_htf_to_ltf(ltf, feat)
+
+    c = m["close"].values.astype(float)
+    a = S.atr(m, p.ltf_atr_win)
+    comp_long = np.nan_to_num(m["comp_long"].values).astype(bool)
+    comp_short = np.nan_to_num(m["comp_short"].values).astype(bool)
+    days = pd.to_datetime(m["datetime"]).dt.floor("D").values
+
+    entries: list = [None] * len(m)
+    count_today: dict = {}
+    for i in range(len(m)):
+        if not np.isfinite(a[i]) or a[i] <= 0:
+            continue
+        day = days[i]
+        if count_today.get(day, 0) >= p.max_per_day:
+            continue
+        if comp_long[i]:
+            direction, mb = 1, p.uscitalong
+        elif comp_short[i]:
+            direction, mb = -1, p.uscitashort
+        else:
+            continue
+        if p.exit_mode == "atr":
+            sl_off, tp_off = p.sl_atr * a[i], p.tp_atr * a[i]
+        else:
+            sl_off, tp_off = p.sl_pct * c[i], p.tp_pct * c[i]
+        if direction == 1:
+            sl, tp = c[i] - sl_off, c[i] + tp_off
+        else:
+            sl, tp = c[i] + sl_off, c[i] - tp_off
+        entries[i] = {"dir": direction, "tp": float(tp), "sl": float(sl), "max_bars": int(mb)}
+        count_today[day] = count_today.get(day, 0) + 1
+    return entries
+
+
+# ---------------------------------------------------------------------------
+# Eval helpers
+# ---------------------------------------------------------------------------
+def _split(eq, idx, mask):
+    e = eq[mask]
+    if len(e) < 5:
+        return dict(sharpe=0.0, ret=0.0, maxdd=0.0, n=int(len(e)))
+    r = np.diff(e) / e[:-1]; r = r[np.isfinite(r)]
+    ix = idx[mask]
+    dt = pd.Series(ix).diff().dt.total_seconds().median() or 86400
+    bpy = 86400 * 365.25 / dt
+    sh = float(np.mean(r) / np.std(r) * np.sqrt(bpy)) if len(r) and np.std(r) > 0 else 0.0
+    pk = np.maximum.accumulate(e); dd = float(np.max((pk - e) / pk))
+    return dict(sharpe=round(sh, 3), ret=round(float(e[-1] / e[0] - 1), 4), maxdd=round(dd, 4), n=int(len(e)))
+
+
+def eval_cfg(cfg, p):
+    out = {}
+    for a in ("BTC", "ETH"):
+        ltf, htf = sk.frames(a)
+        ent = expand_entries(ltf, htf, p, **cfg)
+        m = backtest_signals(ltf, ent, fee_rt=FEE, leverage=1.0, asset=a, tf="230m")
+        eq = m.equity
+        idx = pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True))
+        hmask = np.asarray(idx >= HOLDOUT)
+        full = dict(sharpe=round(m.sharpe, 3), ret=round(m.net_return, 4), maxdd=round(m.max_dd, 4),
+                    n_trades=int(m.n_trades), win_rate=round(m.win_rate, 1))
+        hold = _split(eq, idx, hmask)
+        out[a] = dict(full=full, hold=hold, _eq=eq, _idx=idx)
+    return out
+
+
+def summarize(tag, res):
+    mf = min(res[a]["full"]["sharpe"] for a in res)
+    mh = min(res[a]["hold"]["sharpe"] for a in res)
+    mt = min(res[a]["full"]["n_trades"] for a in res)
+    mdd = max(res[a]["full"]["maxdd"] for a in res)
+    print(f"  [{tag}] minFull={mf:+.2f} minHold={mh:+.2f} minTr={mt} maxDD={mdd*100:.0f}%"
+          f" | BTC F{res['BTC']['full']['sharpe']:+.2f}/H{res['BTC']['hold']['sharpe']:+.2f} n{res['BTC']['full']['n_trades']}"
+          f" ETH F{res['ETH']['full']['sharpe']:+.2f}/H{res['ETH']['hold']['sharpe']:+.2f} n{res['ETH']['full']['n_trades']}")
+    return dict(minFull=mf, minHold=mh, minTr=mt, maxDD=mdd, res=res)
+
+
+def v1_like_params(**kw):
+    base = dict(ptn_n=55, sl_atr=2.5, tp_atr=6.0)  # V1 pattern/exit
+    base.update(kw)
+    return SkyhookParams(**base)
+
+
+# ---------------------------------------------------------------------------
+# Causality self-check on the structural variant (truncated-prefix guard)
+# ---------------------------------------------------------------------------
+def check_causality(cfg, p, asset="BTC", tail=150):
+    ltf, htf = sk.frames(asset)
+    full = expand_entries(ltf, htf, p, **cfg)
+    n = len(ltf); bad = 0; checked = 0
+    for frac in (0.80, 0.92):
+        cut = int(n * frac)
+        cut_ts = int(ltf["timestamp"].iloc[cut - 1])
+        htf_cut = htf[htf["timestamp"] <= cut_ts].reset_index(drop=True)
+        sub = expand_entries(ltf.iloc[:cut].reset_index(drop=True), htf_cut, p, **cfg)
+        for i in range(max(0, cut - tail), cut):
+            checked += 1
+            a, b = full[i], sub[i]
+            if (a is None) != (b is None):
+                bad += 1
+            elif a is not None and (a["dir"] != b["dir"]
+                                    or abs(a["sl"] - b["sl"]) > 1e-6
+                                    or abs(a["tp"] - b["tp"]) > 1e-6):
+                bad += 1
+    return dict(ok=bool(bad == 0), mismatches=int(bad), checked=int(checked))
+
+
+if __name__ == "__main__":
+    print("=== SKH_R_EXPAND: volatility-EXPANSION regime (ATR rising vs its MA + volume elevated) ===\n")
+
+    # --- V1 reference (Chande01 regime) ---
+    print("--- V1 reference (Chande01 regime, ptn_n=55 sl2.5 tp6 vola35-95 vol0) ---")
+    v1 = SkyhookParams(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35, vola_hi=95, vol_lo=0.0)
+    v1res = {}
+    for a in ("BTC", "ETH"):
+        r = sk.run_asset(a, v1, FEE)
+        v1res[a] = dict(full=dict(sharpe=r["full"]["sharpe"], n_trades=r["full"]["n_trades"],
+                                  maxdd=r["full"]["maxdd"]), hold=dict(sharpe=r["holdout"]["sharpe"]))
+    v1_minFull = min(v1res[a]['full']['sharpe'] for a in v1res)
+    v1_minHold = min(v1res[a]['hold']['sharpe'] for a in v1res)
+    v1_maxDD = max(v1res[a]['full']['maxdd'] for a in v1res)
+    print(f"  V1 minFull={v1_minFull:+.2f} minHold={v1_minHold:+.2f} maxDD={v1_maxDD*100:.0f}%"
+          f" | BTC F{v1res['BTC']['full']['sharpe']:+.2f}/H{v1res['BTC']['hold']['sharpe']:+.2f}"
+          f" ETH F{v1res['ETH']['full']['sharpe']:+.2f}/H{v1res['ETH']['hold']['sharpe']:+.2f}\n")
+
+    p = v1_like_params()  # same pattern/exit as V1; only regime changes
+
+    # --- Sweep: expansion-MA windows + multipliers ---
+    # k=1.0 -> "ATR above its own MA" (mild rising). k>1 -> stronger expansion (fewer trades).
+    # w_atr/w_vol: lookback for the MA (HTF bars; 690min each). vol elevated mirrored on volume.
+    print("--- volatility-EXPANSION sweep ---")
+    cfgs = {
+        # (w_atr,k_atr) , (w_vol,k_vol)
+        "atr20k1.0_vol20k1.0": dict(w_atr=20, k_atr=1.00, w_vol=20, k_vol=1.00),
+        "atr20k1.0_vol20k1.2": dict(w_atr=20, k_atr=1.00, w_vol=20, k_vol=1.20),
+        "atr20k1.1_vol20k1.0": dict(w_atr=20, k_atr=1.10, w_vol=20, k_vol=1.00),
+        "atr20k1.1_vol20k1.2": dict(w_atr=20, k_atr=1.10, w_vol=20, k_vol=1.20),
+        "atr10k1.0_vol10k1.0": dict(w_atr=10, k_atr=1.00, w_vol=10, k_vol=1.00),
+        "atr10k1.1_vol10k1.2": dict(w_atr=10, k_atr=1.10, w_vol=10, k_vol=1.20),
+        "atr30k1.0_vol30k1.0": dict(w_atr=30, k_atr=1.00, w_vol=30, k_vol=1.00),
+        "atr20k1.0_volOFF":    dict(w_atr=20, k_atr=1.00, w_vol=20, k_vol=0.00),  # vol gate off (k=0 always true)
+        "atr20k1.2_vol20k1.0": dict(w_atr=20, k_atr=1.20, w_vol=20, k_vol=1.00),
+    }
+    results = {}
+    for tag, cfg in cfgs.items():
+        results[tag] = (cfg, summarize(tag, eval_cfg(cfg, p)))
+
+    # --- Pick winner by minHold (subject to minFull>=0.5, minTr>=20) ---
+    elig = {t: v for t, (c, v) in results.items() if v["minFull"] >= 0.5 and v["minTr"] >= 20}
+    print(f"\n--- eligible (minFull>=0.5, minTr>=20): {list(elig.keys())} ---")
+    if elig:
+        win_tag = max(elig, key=lambda t: elig[t]["minHold"])
+    else:
+        win_tag = max(results, key=lambda t: results[t][1]["minHold"])
+    win_cfg, win_v = results[win_tag][0], results[win_tag][1]
+    print(f"\n*** WINNER = {win_tag}  minFull={win_v['minFull']:+.2f} minHold={win_v['minHold']:+.2f}"
+          f" minTr={win_v['minTr']} maxDD={win_v['maxDD']*100:.0f}% ***")
+    print(f"    cfg = {win_cfg}")
+
+    # --- Causality on winner (BOTH assets) ---
+    caus = check_causality(win_cfg, p, "BTC")
+    caus_eth = check_causality(win_cfg, p, "ETH")
+    print(f"\ncausality(BTC) = {caus}")
+    print(f"causality(ETH) = {caus_eth}")
+
+    # --- Fee sweep on winner (both assets, FULL) ---
+    print("\n--- fee sweep on winner (FULL sharpe) ---")
+    fee_ok_all = True
+    for a in ("BTC", "ETH"):
+        ltf, htf = sk.frames(a)
+        ent_cache = expand_entries(ltf, htf, p, **win_cfg)
+        row = []
+        for f in (0.0, 0.001, 0.002, 0.003):
+            m = backtest_signals(ltf, ent_cache, fee_rt=f, leverage=1.0, asset=a, tf="230m")
+            row.append((f, round(m.sharpe, 3)))
+        sh030 = dict(row)[0.003]
+        fee_ok_all = fee_ok_all and (sh030 > 0)
+        print(f"  {a}: " + " ".join(f"{f*100:.2f}%={s:+.2f}" for f, s in row))
+    print(f"  fee_survives 0.30%RT (both): {fee_ok_all}")
+
+    # --- Per-year on winner ---
+    print("\n--- per-year (winner) ---")
+    for a in ("BTC", "ETH"):
+        ltf, htf = sk.frames(a)
+        ent = expand_entries(ltf, htf, p, **win_cfg)
+        m = backtest_signals(ltf, ent, fee_rt=FEE, leverage=1.0, asset=a, tf="230m")
+        yr = " ".join(f"{y}:{v*100:+.0f}%" for y, v in m.yearly.items())
+        print(f"  {a}: {yr}")
+
+    # --- Marginal vs TP01 on winner ---
+    def daily_series(a, p, cfg):
+        ltf, htf = sk.frames(a)
+        ent = expand_entries(ltf, htf, p, **cfg)
+        m = backtest_signals(ltf, ent, fee_rt=FEE, leverage=1.0, asset=a, tf="230m")
+        s = pd.Series(m.equity, index=pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True)))
+        return s.resample("1D").last().ffill().pct_change().dropna()
+
+    import altlib as al
+    sb = daily_series("BTC", p, win_cfg); se = daily_series("ETH", p, win_cfg)
+    J = pd.concat({"BTC": sb, "ETH": se}, axis=1, join="inner").fillna(0.0)
+    cand_daily = 0.5 * J["BTC"] + 0.5 * J["ETH"]
+    marg = al.marginal_vs_tp01(cand_daily)
+    print("\n--- marginal vs TP01 (winner) ---")
+    print(f"  corr_full={marg.get('corr_full')} corr_hold={marg.get('corr_hold')}")
+    print(f"  blend w25 uplift_hold={marg.get('blends',{}).get('w25',{}).get('uplift_hold')}")
+    print(f"  marginal_verdict={marg.get('marginal_verdict')} robust_oos={marg.get('robust_oos')}")
+    print(f"  has_insample_edge={marg.get('has_insample_edge')} is_hedge={marg.get('is_hedge')}")
+    print(f"  clean_year_uplift={marg.get('clean_year_uplift')} jackknife_min_uplift={marg.get('jackknife_min_uplift')}")
+    print(f"  multicut_persistent={marg.get('multicut_persistent')}")
+
+    # --- Final verdict echo ---
+    print("\n=== SUMMARY ===")
+    print(f"V1:            minFull={v1_minFull:+.2f} minHold={v1_minHold:+.2f} maxDD={v1_maxDD*100:.0f}%")
+    print(f"EXPAND {win_tag}: minFull={win_v['minFull']:+.2f} minHold={win_v['minHold']:+.2f}"
+          f" maxDD={win_v['maxDD']*100:.0f}% causalityOK={caus['ok'] and caus_eth['ok']} feeOK={fee_ok_all}")
+    beats = (win_v['minHold'] > v1_minHold) and win_v['minFull'] >= 0.5 and fee_ok_all
+    print(f"BEATS V1 HOLD-OUT: {beats}")

scripts/research/skyhook/runs/SKH_R_EXPAND_study.py

@@ -0,0 +1,72 @@
+"""SKH_R_EXPAND_study — final study() on the two best volatility-expansion configs.
+
+We use the project's HONEST study() harness. Because the expansion regime is a STRUCTURAL
+change (not expressible via SkyhookParams bands), we monkeypatch htf_features INSIDE
+skyhooklib's namespace to our expansion-features for the duration of each study, so study()
+runs the exact same leak-free FULL+HOLDOUT+fee-sweep+per-year machinery on our entries.
+
+This is safe: we only swap the feature builder (regime def); pattern/composer/entry/exit and
+all the eval code are unchanged. We restore the original after each study.
+"""
+from __future__ import annotations
+
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+
+import numpy as np
+import pandas as pd
+
+import skyhooklib as sk
+from src.strategies import skyhook as S
+from src.strategies.skyhook import SkyhookParams
+
+# import our feature builder
+from SKH_R_EXPAND import expand_htf_features
+
+ORIG_FEAT = S.htf_features
+
+
+def make_patched_features(cfg):
+    def _feat(htf, p):
+        return expand_htf_features(htf, p, **cfg)
+    return _feat
+
+
+def study_expand(name, p, cfg):
+    """Run sk.study with htf_features patched to the expansion regime defined by cfg."""
+    patched = make_patched_features(cfg)
+    # skyhook_entries calls skyhook.htf_features via the module-level name S.htf_features.
+    S.htf_features = patched
+    try:
+        rep = sk.study(name, p)
+        caus = sk.causality(p, "BTC")
+        caus_eth = sk.causality(p, "ETH")
+        marg = sk.marginal(p)
+    finally:
+        S.htf_features = ORIG_FEAT
+    return rep, (caus, caus_eth), marg
+
+
+if __name__ == "__main__":
+    p = SkyhookParams(ptn_n=55, sl_atr=2.5, tp_atr=6.0)
+
+    configs = {
+        # best DD-cut + above-V1 full, balanced volume gate (maxDD 34%)
+        "EXP_atr20vol20": dict(w_atr=20, k_atr=1.0, w_vol=20, k_vol=1.0),
+        # best hold-out / no volume gate (minFull 0.81, DD 40%)
+        "EXP_atr20volOFF": dict(w_atr=20, k_atr=1.0, w_vol=20, k_vol=0.0),
+    }
+
+    for name, cfg in configs.items():
+        print(f"\n########## {name}  cfg={cfg} ##########")
+        rep, (caus, caus_eth), marg = study_expand(name, p, cfg)
+        print(sk.fmt(rep))
+        print(f"causality BTC={caus} ETH={caus_eth}")
+        print(f"marginal: verdict={marg.get('marginal_verdict')} corr_full={marg.get('corr_full')}"
+              f" blend_w25_uplift_hold={marg.get('blends',{}).get('w25',{}).get('uplift_hold')}")
+        print(f"          has_insample_edge={marg.get('has_insample_edge')} is_hedge={marg.get('is_hedge')}"
+              f" robust_oos={marg.get('robust_oos')} clean_year_uplift={marg.get('clean_year_uplift')}"
+              f" multicut_persistent={marg.get('multicut_persistent')}")
+        v = rep["verdict"]
+        print(f"VERDICT[{name}]: grade={v['grade']} minFull={v['min_asset_full_sharpe']}"
+              f" minHold={v['min_asset_holdout_sharpe']} minTrades={v['min_trades']} feeOK={v['fee_survives']}")

scripts/research/skyhook/runs/SKH_R_PCTL.py

@@ -0,0 +1,308 @@
+"""SKH_R_PCTL — REGIME variant: replace Chande01 regime with CAUSAL expanding/rolling
+PERCENTILE-RANK of ATR and volume (0-1), gate on rank bands.
+
+Hypothesis: Chande01 measures the *direction/momentum* of the vol/volume cycle (rising vs
+falling), mapped to 0-100. A percentile-RANK instead measures *where the current level sits*
+within its own history (is ATR/volume HIGH or LOW relative to the past). This is a more
+natural "regime" definition: trade only when vol/volume is in a chosen part of its own
+distribution. We test expanding (full history) and rolling-window percentile ranks.
+
+Causality: rank[i] uses only x[0..i] (expanding) or x[i-w+1..i] (rolling), INCLUSIVE of the
+current bar — this is legitimate because at HTF close[i] the bar's ATR/volume is known. The
+HTF feature is then merged backward to LTF on HTF-close timestamp (<= LTF close). We verify
+the truncated-prefix guard ourselves.
+
+Pattern/composer/entry/exit reuse the V1 Skyhook building blocks unchanged.
+"""
+from __future__ import annotations
+
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+
+import numpy as np
+import pandas as pd
+
+import skyhooklib as sk
+from src.backtest.harness import backtest_signals
+from src.strategies import skyhook as S
+from src.strategies.skyhook import SkyhookParams, HTF_MIN, LTF_MIN
+
+HOLDOUT = pd.Timestamp("2025-01-01", tz="UTC")
+FEE = sk.FEE_RT
+
+
+# ---------------------------------------------------------------------------
+# Causal percentile-rank (0-1). Fraction of the window strictly < current value,
+# computed inclusive of the current bar (legit: bar is closed). min_periods enforced.
+# ---------------------------------------------------------------------------
+def pctl_rank(x: np.ndarray, win: int | None, min_periods: int = 30) -> np.ndarray:
+    """Causal percentile rank in [0,1]. win=None -> EXPANDING; else rolling window.
+    rank[i] = (#{x[j] < x[i]} + 0.5*#{x[j]==x[i]}) / count  over the (expanding/rolling) window."""
+    x = np.asarray(x, float)
+    s = pd.Series(x)
+    if win is None:
+        # expanding rank: pandas .expanding().rank(pct=True) gives rank/count INCLUSIVE of i,
+        # which counts <= (so the current bar's own value is included). Use 'average' to break ties.
+        r = s.expanding(min_periods=min_periods).rank(pct=True)
+    else:
+        r = s.rolling(win, min_periods=min(min_periods, win)).rank(pct=True)
+    return r.values  # NaN until min_periods reached
+
+
+# ---------------------------------------------------------------------------
+# HTF feature df with percentile-rank regime gate + Donchian pattern (V1 pattern reused).
+# ---------------------------------------------------------------------------
+def pctl_htf_features(htf: pd.DataFrame, p: SkyhookParams,
+                      vola_win: int | None, vol_win: int | None,
+                      vola_lo: float, vola_hi: float,
+                      vol_lo: float, vol_hi: float) -> pd.DataFrame:
+    """Regime via CAUSAL percentile-rank (0-1) of ATR and volume; pattern via Donchian.
+    Bands here are in [0,1] (percentile space), NOT 0-100 like Chande01."""
+    atr_htf = S.atr(htf, p.atr_win)
+    vola_rank = pctl_rank(atr_htf, vola_win)
+    vol_rank = pctl_rank(htf["volume"].values, vol_win)
+    ptn_long, ptn_short = S.donchian_breakout(htf, p.ptn_n)
+
+    # regime_ok requires a valid (non-NaN) rank in band; NaN (warmup) -> False
+    vr_ok = np.where(np.isfinite(vola_rank), (vola_rank >= vola_lo) & (vola_rank <= vola_hi), False)
+    vol_ok = np.where(np.isfinite(vol_rank), (vol_rank >= vol_lo) & (vol_rank <= vol_hi), False)
+    regime_ok = vr_ok & vol_ok
+
+    comp_long = regime_ok & ptn_long
+    comp_short = regime_ok & ptn_short
+    if p.long_only:
+        comp_short = np.zeros_like(comp_short, dtype=bool)
+
+    close_ts = htf["timestamp"].astype("int64").values + HTF_MIN * 60 * 1000
+    return pd.DataFrame({
+        "close_ts": close_ts,
+        "buz_vola": vola_rank, "buz_volume": vol_rank,
+        "comp_long": comp_long.astype(bool), "comp_short": comp_short.astype(bool),
+    })
+
+
+def pctl_entries(ltf: pd.DataFrame, htf: pd.DataFrame, p: SkyhookParams,
+                 vola_win, vol_win, vola_lo, vola_hi, vol_lo, vol_hi) -> list:
+    """Same entry/exit machinery as S.skyhook_entries, but regime from pctl features."""
+    feat = pctl_htf_features(htf, p, vola_win, vol_win, vola_lo, vola_hi, vol_lo, vol_hi)
+    m = S.merge_htf_to_ltf(ltf, feat)
+
+    c = m["close"].values.astype(float)
+    a = S.atr(m, p.ltf_atr_win)
+    comp_long = np.nan_to_num(m["comp_long"].values).astype(bool)
+    comp_short = np.nan_to_num(m["comp_short"].values).astype(bool)
+    days = pd.to_datetime(m["datetime"]).dt.floor("D").values
+
+    entries: list = [None] * len(m)
+    count_today: dict = {}
+    for i in range(len(m)):
+        if not np.isfinite(a[i]) or a[i] <= 0:
+            continue
+        day = days[i]
+        if count_today.get(day, 0) >= p.max_per_day:
+            continue
+        if comp_long[i]:
+            direction, mb = 1, p.uscitalong
+        elif comp_short[i]:
+            direction, mb = -1, p.uscitashort
+        else:
+            continue
+        if p.exit_mode == "atr":
+            sl_off, tp_off = p.sl_atr * a[i], p.tp_atr * a[i]
+        else:
+            sl_off, tp_off = p.sl_pct * c[i], p.tp_pct * c[i]
+        if direction == 1:
+            sl, tp = c[i] - sl_off, c[i] + tp_off
+        else:
+            sl, tp = c[i] + sl_off, c[i] - tp_off
+        entries[i] = {"dir": direction, "tp": float(tp), "sl": float(sl), "max_bars": int(mb)}
+        count_today[day] = count_today.get(day, 0) + 1
+    return entries
+
+
+# ---------------------------------------------------------------------------
+# Eval helpers
+# ---------------------------------------------------------------------------
+def _split(eq, idx, mask):
+    e = eq[mask]
+    if len(e) < 5:
+        return dict(sharpe=0.0, ret=0.0, maxdd=0.0, n=int(len(e)))
+    r = np.diff(e) / e[:-1]; r = r[np.isfinite(r)]
+    ix = idx[mask]
+    dt = pd.Series(ix).diff().dt.total_seconds().median() or 86400
+    bpy = 86400 * 365.25 / dt
+    sh = float(np.mean(r) / np.std(r) * np.sqrt(bpy)) if len(r) and np.std(r) > 0 else 0.0
+    pk = np.maximum.accumulate(e); dd = float(np.max((pk - e) / pk))
+    return dict(sharpe=round(sh, 3), ret=round(float(e[-1] / e[0] - 1), 4), maxdd=round(dd, 4), n=int(len(e)))
+
+
+def eval_cfg(cfg, p):
+    """Run both assets; return dict per asset with full+holdout."""
+    out = {}
+    for a in ("BTC", "ETH"):
+        ltf, htf = sk.frames(a)
+        ent = pctl_entries(ltf, htf, p, **cfg)
+        m = backtest_signals(ltf, ent, fee_rt=FEE, leverage=1.0, asset=a, tf="230m")
+        eq = m.equity
+        idx = pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True))
+        hmask = np.asarray(idx >= HOLDOUT)
+        full = dict(sharpe=round(m.sharpe, 3), ret=round(m.net_return, 4), maxdd=round(m.max_dd, 4),
+                    n_trades=int(m.n_trades), win_rate=round(m.win_rate, 1))
+        hold = _split(eq, idx, hmask)
+        out[a] = dict(full=full, hold=hold, _eq=eq, _idx=idx)
+    return out
+
+
+def summarize(tag, res):
+    mf = min(res[a]["full"]["sharpe"] for a in res)
+    mh = min(res[a]["hold"]["sharpe"] for a in res)
+    mt = min(res[a]["full"]["n_trades"] for a in res)
+    mdd = max(res[a]["full"]["maxdd"] for a in res)
+    print(f"  [{tag}] minFull={mf:+.2f} minHold={mh:+.2f} minTr={mt} maxDD={mdd*100:.0f}%"
+          f" | BTC F{res['BTC']['full']['sharpe']:+.2f}/H{res['BTC']['hold']['sharpe']:+.2f}"
+          f" ETH F{res['ETH']['full']['sharpe']:+.2f}/H{res['ETH']['hold']['sharpe']:+.2f}")
+    return dict(minFull=mf, minHold=mh, minTr=mt, maxDD=mdd, res=res)
+
+
+# Build a SkyhookParams matching V1 non-regime knobs (pattern + exits)
+def v1_like_params(**kw):
+    base = dict(ptn_n=55, sl_atr=2.5, tp_atr=6.0)  # V1 pattern/exit
+    base.update(kw)
+    return SkyhookParams(**base)
+
+
+# ---------------------------------------------------------------------------
+# Causality self-check on the structural variant
+# ---------------------------------------------------------------------------
+def check_causality(cfg, p, asset="BTC", tail=150):
+    ltf, htf = sk.frames(asset)
+    full = pctl_entries(ltf, htf, p, **cfg)
+    n = len(ltf); bad = 0; checked = 0
+    for frac in (0.80, 0.92):
+        cut = int(n * frac)
+        cut_ts = int(ltf["timestamp"].iloc[cut - 1])
+        htf_cut = htf[htf["timestamp"] <= cut_ts].reset_index(drop=True)
+        sub = pctl_entries(ltf.iloc[:cut].reset_index(drop=True), htf_cut, p, **cfg)
+        for i in range(max(0, cut - tail), cut):
+            checked += 1
+            a, b = full[i], sub[i]
+            if (a is None) != (b is None):
+                bad += 1
+            elif a is not None and (a["dir"] != b["dir"]
+                                    or abs(a["sl"] - b["sl"]) > 1e-6
+                                    or abs(a["tp"] - b["tp"]) > 1e-6):
+                bad += 1
+    return dict(ok=bool(bad == 0), mismatches=int(bad), checked=int(checked))
+
+
+if __name__ == "__main__":
+    print("=== SKH_R_PCTL: percentile-rank regime ===\n")
+
+    # --- V1 reference for comparison (Chande01 regime) ---
+    print("--- V1 reference (Chande01 regime, ptn_n=55 sl2.5 tp6 vola35-95 vol0) ---")
+    v1 = SkyhookParams(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35, vola_hi=95, vol_lo=0.0)
+    v1res = {}
+    for a in ("BTC", "ETH"):
+        r = sk.run_asset(a, v1, FEE)
+        v1res[a] = dict(full=dict(sharpe=r["full"]["sharpe"], n_trades=r["full"]["n_trades"],
+                                  maxdd=r["full"]["maxdd"]),
+                        hold=dict(sharpe=r["holdout"]["sharpe"]))
+    print(f"  V1 minFull={min(v1res[a]['full']['sharpe'] for a in v1res):+.2f}"
+          f" minHold={min(v1res[a]['hold']['sharpe'] for a in v1res):+.2f}"
+          f" maxDD={max(v1res[a]['full']['maxdd'] for a in v1res)*100:.0f}%"
+          f" | BTC F{v1res['BTC']['full']['sharpe']:+.2f}/H{v1res['BTC']['hold']['sharpe']:+.2f}"
+          f" ETH F{v1res['ETH']['full']['sharpe']:+.2f}/H{v1res['ETH']['hold']['sharpe']:+.2f}\n")
+
+    p = v1_like_params()  # same pattern/exit as V1; only regime changes
+
+    # --- Sweep: percentile-rank regime bands ---
+    # Two regime intuitions to test:
+    #   (A) HIGH-vol/HIGH-volume regime (breakout-friendly): ranks in upper band.
+    #   (B) MID regime (avoid blow-off + dead): ranks in a middle band.
+    # vol_lo=0 means "no lower bound on volume" (mirror V1's vol_lo=0).
+    print("--- EXPANDING percentile-rank sweep ---")
+    cfgs = {
+        # vola band, vol band, both expanding (win=None)
+        "exp_volaHi_vol0":   dict(vola_win=None, vol_win=None, vola_lo=0.35, vola_hi=0.95, vol_lo=0.0, vol_hi=1.0),
+        "exp_volaMid_vol0":  dict(vola_win=None, vol_win=None, vola_lo=0.30, vola_hi=0.80, vol_lo=0.0, vol_hi=1.0),
+        "exp_volaHi_volHi":  dict(vola_win=None, vol_win=None, vola_lo=0.35, vola_hi=0.95, vol_lo=0.40, vol_hi=1.0),
+        "exp_volaLo_vol0":   dict(vola_win=None, vol_win=None, vola_lo=0.0,  vola_hi=0.70, vol_lo=0.0, vol_hi=1.0),
+        "exp_volaWide_vol0": dict(vola_win=None, vol_win=None, vola_lo=0.20, vola_hi=1.0,  vol_lo=0.0, vol_hi=1.0),
+    }
+    results = {}
+    for tag, cfg in cfgs.items():
+        results[tag] = (cfg, summarize(tag, eval_cfg(cfg, p)))
+
+    print("\n--- ROLLING percentile-rank sweep (win=60 HTF bars ~ recent regime) ---")
+    cfgs_roll = {
+        "roll60_volaHi_vol0":  dict(vola_win=60, vol_win=60, vola_lo=0.35, vola_hi=0.95, vol_lo=0.0, vol_hi=1.0),
+        "roll60_volaMid_vol0": dict(vola_win=60, vol_win=60, vola_lo=0.30, vola_hi=0.80, vol_lo=0.0, vol_hi=1.0),
+        "roll120_volaHi_vol0": dict(vola_win=120, vol_win=120, vola_lo=0.35, vola_hi=0.95, vol_lo=0.0, vol_hi=1.0),
+        "roll60_volaHi_volHi": dict(vola_win=60, vol_win=60, vola_lo=0.35, vola_hi=0.95, vol_lo=0.40, vol_hi=1.0),
+    }
+    for tag, cfg in cfgs_roll.items():
+        results[tag] = (cfg, summarize(tag, eval_cfg(cfg, p)))
+
+    # --- Pick winner by minHold (subject to minFull>=0.5, minTr>=20) ---
+    elig = {t: v for t, (c, v) in results.items()
+            if v["minFull"] >= 0.5 and v["minTr"] >= 20}
+    print(f"\n--- eligible (minFull>=0.5, minTr>=20): {list(elig.keys())} ---")
+    if elig:
+        win_tag = max(elig, key=lambda t: elig[t]["minHold"])
+    else:
+        # fall back to best minHold overall to report honestly
+        win_tag = max(results, key=lambda t: results[t][1]["minHold"])
+    win_cfg, win_v = results[win_tag][0], results[win_tag][1]
+    print(f"\n*** WINNER = {win_tag}  minFull={win_v['minFull']:+.2f} minHold={win_v['minHold']:+.2f}"
+          f" minTr={win_v['minTr']} maxDD={win_v['maxDD']*100:.0f}% ***")
+    print(f"    cfg = {win_cfg}")
+
+    # --- Causality on winner ---
+    caus = check_causality(win_cfg, p, "BTC")
+    print(f"\ncausality(BTC) = {caus}")
+    caus_eth = check_causality(win_cfg, p, "ETH")
+    print(f"causality(ETH) = {caus_eth}")
+
+    # --- Fee sweep on winner (both assets, FULL) ---
+    print("\n--- fee sweep on winner (FULL sharpe) ---")
+    fee_ok_all = True
+    for a in ("BTC", "ETH"):
+        ltf, htf = sk.frames(a)
+        ent_cache = pctl_entries(ltf, htf, p, **win_cfg)
+        row = []
+        for f in (0.0, 0.001, 0.002, 0.003):
+            m = backtest_signals(ltf, ent_cache, fee_rt=f, leverage=1.0, asset=a, tf="230m")
+            row.append((f, round(m.sharpe, 3)))
+        sh030 = dict(row)[0.003]
+        fee_ok_all = fee_ok_all and (sh030 > 0)
+        print(f"  {a}: " + " ".join(f"{f*100:.2f}%={s:+.2f}" for f, s in row))
+    print(f"  fee_survives 0.30%RT (both): {fee_ok_all}")
+
+    # --- Marginal vs TP01 on winner ---
+    # Build a daily 50/50 series the same way skyhooklib does, but with our entries.
+    def daily_series(a, p, cfg):
+        ltf, htf = sk.frames(a)
+        ent = pctl_entries(ltf, htf, p, **cfg)
+        m = backtest_signals(ltf, ent, fee_rt=FEE, leverage=1.0, asset=a, tf="230m")
+        s = pd.Series(m.equity, index=pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True)))
+        return s.resample("1D").last().ffill().pct_change().dropna()
+
+    import altlib as al
+    sb = daily_series("BTC", p, win_cfg); se = daily_series("ETH", p, win_cfg)
+    J = pd.concat({"BTC": sb, "ETH": se}, axis=1, join="inner").fillna(0.0)
+    cand_daily = 0.5 * J["BTC"] + 0.5 * J["ETH"]
+    marg = al.marginal_vs_tp01(cand_daily)
+    print("\n--- marginal vs TP01 (winner) ---")
+    print(f"  corr_full={marg.get('corr_full')} corr_hold={marg.get('corr_hold')}")
+    print(f"  blend w25 uplift_hold={marg.get('blends',{}).get('w25',{}).get('uplift_hold')}")
+    print(f"  marginal_verdict={marg.get('marginal_verdict')} robust_oos={marg.get('robust_oos')}")
+    print(f"  has_insample_edge={marg.get('has_insample_edge')} is_hedge={marg.get('is_hedge')}")
+    print(f"  clean_year_uplift={marg.get('clean_year_uplift')} jackknife_min_uplift={marg.get('jackknife_min_uplift')}")
+
+    # --- Final verdict echo ---
+    print("\n=== SUMMARY ===")
+    print(f"V1:     minFull={min(v1res[a]['full']['sharpe'] for a in v1res):+.2f}"
+          f" minHold={min(v1res[a]['hold']['sharpe'] for a in v1res):+.2f}"
+          f" maxDD={max(v1res[a]['full']['maxdd'] for a in v1res)*100:.0f}%")
+    print(f"PCTL {win_tag}: minFull={win_v['minFull']:+.2f} minHold={win_v['minHold']:+.2f}"
+          f" maxDD={win_v['maxDD']*100:.0f}% causalityOK={caus['ok'] and caus_eth['ok']} feeOK={fee_ok_all}")

scripts/research/skyhook/runs/SKH_R_PCTL_final.py

@@ -0,0 +1,102 @@
+"""SKH_R_PCTL final: verify top configs with sk.study + marginal, refine for minFull/DD."""
+from __future__ import annotations
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+import numpy as np, pandas as pd
+import skyhooklib as sk
+from src.backtest.harness import backtest_signals
+from src.strategies import skyhook as S
+from src.strategies.skyhook import SkyhookParams, HTF_MIN
+
+# import the structural builder from the sweep script
+import importlib.util
+spec = importlib.util.spec_from_file_location(
+    "skr", "/opt/docker/PythagorasGoal/scripts/research/skyhook/runs/SKH_R_PCTL.py")
+skr = importlib.util.module_from_spec(spec)
+# avoid running its __main__
+import builtins
+_orig_name = "__main__"
+spec.loader.exec_module(skr)  # defines functions; __main__ guard prevents the sweep
+
+HOLDOUT = sk.HOLDOUT
+FEE = sk.FEE_RT
+
+
+def study_struct(name, cfg, p):
+    """sk.study-equivalent for our structural variant: FULL+HOLD+fee-sweep+per-year BOTH assets."""
+    per_asset = {}
+    fee_ok_all = True
+    for a in ("BTC", "ETH"):
+        ltf, htf = sk.frames(a)
+        ent = skr.pctl_entries(ltf, htf, p, **cfg)
+        m = backtest_signals(ltf, ent, fee_rt=FEE, leverage=1.0, asset=a, tf="230m")
+        eq = m.equity
+        idx = pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True))
+        hmask = np.asarray(idx >= HOLDOUT)
+        full = dict(sharpe=round(m.sharpe, 3), ret=round(m.net_return, 4), maxdd=round(m.max_dd, 4),
+                    n_trades=int(m.n_trades), win_rate=round(m.win_rate, 1))
+        hold = skr._split(eq, idx, hmask)
+        sweep = {}
+        for f in (0.0, 0.001, 0.002, 0.003):
+            mf = backtest_signals(ltf, ent, fee_rt=f, leverage=1.0, asset=a, tf="230m")
+            sweep[f"{f*100:.2f}%"] = round(mf.sharpe, 3)
+        fee_ok_all = fee_ok_all and (sweep["0.30%"] > 0)
+        per_asset[a] = dict(full=full, hold=hold, yearly={int(y): round(v, 4) for y, v in m.yearly.items()},
+                            fee_sweep=sweep)
+    mf = min(per_asset[a]["full"]["sharpe"] for a in per_asset)
+    mh = min(per_asset[a]["hold"]["sharpe"] for a in per_asset)
+    mt = min(per_asset[a]["full"]["n_trades"] for a in per_asset)
+    mdd = max(per_asset[a]["full"]["maxdd"] for a in per_asset)
+    grade = "PASS" if (mt >= 20 and mf >= 0.5 and mh >= 0.2 and fee_ok_all) else \
+            ("WEAK" if (mt >= 20 and mf >= 0.3 and mh >= 0.0) else "FAIL")
+    print(f"\n=== {name} -> {grade}  (minFull={mf:+.2f} minHold={mh:+.2f} minTr={mt} maxDD={mdd*100:.0f}% feeOK={fee_ok_all})")
+    for a in per_asset:
+        pa = per_asset[a]
+        yr = " ".join(f"{y}:{r*100:+.0f}%" for y, r in pa["yearly"].items())
+        print(f"  {a}: FULL Sh={pa['full']['sharpe']:+.2f} ret={pa['full']['ret']*100:+.0f}% DD={pa['full']['maxdd']*100:.0f}%"
+              f" n={pa['full']['n_trades']} wr={pa['full']['win_rate']:.0f}% | HOLD Sh={pa['hold']['sharpe']:+.2f} ret={pa['hold']['ret']*100:+.0f}%")
+        print(f"      fee: " + " ".join(f"{k}={v:+.2f}" for k, v in pa["fee_sweep"].items()))
+        print(f"      yr: {yr}")
+    return dict(grade=grade, minFull=mf, minHold=mh, minTr=mt, maxDD=mdd, fee_ok=fee_ok_all, per_asset=per_asset)
+
+
+def marginal_struct(cfg, p):
+    import altlib as al
+    def daily(a):
+        ltf, htf = sk.frames(a)
+        ent = skr.pctl_entries(ltf, htf, p, **cfg)
+        m = backtest_signals(ltf, ent, fee_rt=FEE, leverage=1.0, asset=a, tf="230m")
+        s = pd.Series(m.equity, index=pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True)))
+        return s.resample("1D").last().ffill().pct_change().dropna()
+    sb, se = daily("BTC"), daily("ETH")
+    J = pd.concat({"BTC": sb, "ETH": se}, axis=1, join="inner").fillna(0.0)
+    cand = 0.5 * J["BTC"] + 0.5 * J["ETH"]
+    return al.marginal_vs_tp01(cand)
+
+
+if __name__ == "__main__":
+    p = skr.v1_like_params()
+
+    # Candidate A: best minHold (exp_volaHi_volHi) -- minFull 0.53
+    cfgA = dict(vola_win=None, vol_win=None, vola_lo=0.35, vola_hi=0.95, vol_lo=0.40, vol_hi=1.0)
+    # Candidate B: best minFull + lower DD (exp_volaLo_vol0) -- minFull 0.70, DD 39%
+    cfgB = dict(vola_win=None, vol_win=None, vola_lo=0.0, vola_hi=0.70, vol_lo=0.0, vol_hi=1.0)
+
+    # Refinements to lift minFull on A while keeping hold-out: tighten vola band / add small vol floor
+    cfgC = dict(vola_win=None, vol_win=None, vola_lo=0.10, vola_hi=0.80, vol_lo=0.30, vol_hi=1.0)
+    # B + modest vol floor to keep DD low but lift hold
+    cfgD = dict(vola_win=None, vol_win=None, vola_lo=0.0, vola_hi=0.70, vol_lo=0.30, vol_hi=1.0)
+
+    rA = study_struct("PCTL-A exp_volaHi_volHi", cfgA, p)
+    rB = study_struct("PCTL-B exp_volaLo_vol0", cfgB, p)
+    rC = study_struct("PCTL-C exp_volaLoMid_volFloor", cfgC, p)
+    rD = study_struct("PCTL-D exp_volaLo_volFloor", cfgD, p)
+
+    print("\n\n##### MARGINAL vs TP01 #####")
+    for tag, cfg, r in [("A", cfgA, rA), ("B", cfgB, rB), ("C", cfgC, rC), ("D", cfgD, rD)]:
+        mg = marginal_struct(cfg, p)
+        print(f"[{tag}] grade={r['grade']} minFull={r['minFull']:+.2f} minHold={r['minHold']:+.2f} DD={r['maxDD']*100:.0f}%"
+              f" | corr_full={mg.get('corr_full')} upliftHold={mg.get('blends',{}).get('w25',{}).get('uplift_hold')}"
+              f" verdict={mg.get('marginal_verdict')} robust_oos={mg.get('robust_oos')}"
+              f" insample_edge={mg.get('has_insample_edge')} hedge={mg.get('is_hedge')}"
+              f" cleanYr={mg.get('clean_year_uplift')}")

scripts/research/skyhook/runs/SKH_R_RV.py

@@ -0,0 +1,281 @@
+"""SKH_R_RV — REGIME family: define BuzVola from REALIZED VOL (rolling std of HTF log-returns,
+annualized) instead of ATR, then Chande-normalize. Question: does a returns-based vol regime
+gate better OOS than the ATR-based one?
+
+Structural variant: we rebuild htf_features ourselves, swapping ONLY the BuzVola source from
+chande01(atr) to chande01(realized_vol). Everything else (BuzVolume, Donchian pattern, composer,
+entries, exits) is IDENTICAL to the engine so the comparison is clean. Causal-only: realized vol
+uses log-returns up to the HTF close; chande01 is causal rolling; donchian uses shift(1); the
+HTF->LTF merge is backward on HTF close. We verify causality with a truncated-prefix guard.
+
+V1 reference to beat: SkyhookParams(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35, vola_hi=95,
+vol_lo=0.0) -> minFull +0.69, minHold +0.64 (BTC .64/ETH .64), fee-safe to 0.30%RT, DD ~40-49%.
+"""
+from __future__ import annotations
+
+import sys
+sys.path.insert(0, "/opt/docker/PythagorasGoal")
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/alt")
+
+import numpy as np
+import pandas as pd
+
+import skyhooklib as sk
+from src.strategies import skyhook as S
+from src.strategies.skyhook import SkyhookParams, HTF_MIN, LTF_MIN
+from src.backtest.harness import backtest_signals
+
+HOLDOUT = pd.Timestamp("2025-01-01", tz="UTC")
+FEE = sk.FEE_RT
+
+
+# ---------------------------------------------------------------------------
+# Realized-vol BuzVola: rolling std of HTF LOG-returns, annualized, Chande-normalized.
+# rv_win = lookback bars for the realized-vol estimate.
+# ---------------------------------------------------------------------------
+def realized_vol(htf: pd.DataFrame, rv_win: int) -> np.ndarray:
+    c = htf["close"].values.astype(float)
+    logret = np.zeros_like(c)
+    logret[1:] = np.log(c[1:] / c[:-1])
+    # annualization factor: bars per year at 690 min
+    bars_per_year = 365.25 * 24 * 60 / HTF_MIN
+    rv = pd.Series(logret).rolling(rv_win, min_periods=rv_win).std().values * np.sqrt(bars_per_year)
+    return rv
+
+
+def htf_features_rv(htf: pd.DataFrame, p: SkyhookParams, rv_win: int) -> pd.DataFrame:
+    """Same as S.htf_features but BuzVola = chande01(realized_vol) instead of chande01(atr)."""
+    rv = realized_vol(htf, rv_win)
+    buz_vola = S.chande01(rv, p.n_vola)
+    buz_volume = S.chande01(htf["volume"].values, p.n_volume)
+    ptn_long, ptn_short = S.donchian_breakout(htf, p.ptn_n)
+    regime_ok = ((buz_vola >= p.vola_lo) & (buz_vola <= p.vola_hi)
+                 & (buz_volume >= p.vol_lo) & (buz_volume <= p.vol_hi))
+    comp_long = regime_ok & ptn_long
+    comp_short = regime_ok & ptn_short
+    if p.long_only:
+        comp_short = np.zeros_like(comp_short, dtype=bool)
+    close_ts = htf["timestamp"].astype("int64").values + HTF_MIN * 60 * 1000
+    return pd.DataFrame({
+        "close_ts": close_ts,
+        "buz_vola": buz_vola, "buz_volume": buz_volume,
+        "comp_long": comp_long.astype(bool), "comp_short": comp_short.astype(bool),
+    })
+
+
+def rv_entries(ltf: pd.DataFrame, htf: pd.DataFrame, p: SkyhookParams, rv_win: int) -> list:
+    """Identical to S.skyhook_entries but using the RV-based htf_features."""
+    feat = htf_features_rv(htf, p, rv_win)
+    m = S.merge_htf_to_ltf(ltf, feat)
+    c = m["close"].values.astype(float)
+    a = S.atr(m, p.ltf_atr_win)
+    comp_long = np.nan_to_num(m["comp_long"].values).astype(bool)
+    comp_short = np.nan_to_num(m["comp_short"].values).astype(bool)
+    days = pd.to_datetime(m["datetime"]).dt.floor("D").values
+    entries: list = [None] * len(m)
+    count_today: dict = {}
+    for i in range(len(m)):
+        if not np.isfinite(a[i]) or a[i] <= 0:
+            continue
+        day = days[i]
+        if count_today.get(day, 0) >= p.max_per_day:
+            continue
+        if comp_long[i]:
+            direction, mb = 1, p.uscitalong
+        elif comp_short[i]:
+            direction, mb = -1, p.uscitashort
+        else:
+            continue
+        if p.exit_mode == "atr":
+            sl_off, tp_off = p.sl_atr * a[i], p.tp_atr * a[i]
+        else:
+            sl_off, tp_off = p.sl_pct * c[i], p.tp_pct * c[i]
+        if direction == 1:
+            sl, tp = c[i] - sl_off, c[i] + tp_off
+        else:
+            sl, tp = c[i] + sl_off, c[i] - tp_off
+        entries[i] = {"dir": direction, "tp": float(tp), "sl": float(sl), "max_bars": int(mb)}
+        count_today[day] = count_today.get(day, 0) + 1
+    return entries
+
+
+# ---------------------------------------------------------------------------
+# Backtest one asset with RV entries -> FULL + HOLDOUT metrics
+# ---------------------------------------------------------------------------
+def _split(eq, idx, mask):
+    e = eq[mask]
+    if len(e) < 5:
+        return dict(sharpe=0.0, ret=0.0, maxdd=0.0, n=int(len(e)))
+    r = np.diff(e) / e[:-1]; r = r[np.isfinite(r)]
+    ix = idx[mask]
+    dt = pd.Series(ix).diff().dt.total_seconds().median() or 86400
+    bpy = 86400 * 365.25 / dt
+    sh = float(np.mean(r) / np.std(r) * np.sqrt(bpy)) if len(r) and np.std(r) > 0 else 0.0
+    pk = np.maximum.accumulate(e); dd = float(np.max((pk - e) / pk))
+    return dict(sharpe=round(sh, 3), ret=round(float(e[-1] / e[0] - 1), 4), maxdd=round(dd, 4), n=int(len(e)))
+
+
+def run_rv(asset: str, p: SkyhookParams, rv_win: int, fee=FEE) -> dict:
+    ltf, htf = sk.frames(asset)
+    ent = rv_entries(ltf, htf, p, rv_win)
+    m = backtest_signals(ltf, ent, fee_rt=fee, leverage=1.0, asset=asset, tf="230m")
+    eq = m.equity
+    idx = pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True))
+    hmask = np.asarray(idx >= HOLDOUT)
+    full = dict(sharpe=round(m.sharpe, 3), maxdd=round(m.max_dd, 4), ret=round(m.net_return, 4),
+                n_trades=int(m.n_trades), win_rate=round(m.win_rate, 1))
+    hold = _split(eq, idx, hmask)
+    n_ent = int(sum(e is not None for e in ent))
+    return dict(asset=asset, full=full, holdout=hold, n_entries=n_ent, _eq=eq, _idx=idx)
+
+
+def causality_rv(p: SkyhookParams, rv_win: int, asset="BTC", tail=200) -> dict:
+    ltf, htf = sk.frames(asset)
+    full = rv_entries(ltf, htf, p, rv_win)
+    n = len(ltf); bad = 0; checked = 0
+    for frac in (0.80, 0.92):
+        cut = int(n * frac)
+        cut_ts = int(ltf["timestamp"].iloc[cut - 1])
+        htf_cut = htf[htf["timestamp"] <= cut_ts].reset_index(drop=True)
+        sub = rv_entries(ltf.iloc[:cut].reset_index(drop=True), htf_cut, p, rv_win)
+        for i in range(max(0, cut - tail), cut):
+            checked += 1
+            a, b = full[i], sub[i]
+            if (a is None) != (b is None):
+                bad += 1
+            elif a is not None and (a["dir"] != b["dir"]
+                                    or abs(a["sl"] - b["sl"]) > 1e-6 or abs(a["tp"] - b["tp"]) > 1e-6):
+                bad += 1
+    return dict(ok=bool(bad == 0), mismatches=int(bad), checked=int(checked))
+
+
+def minrep(label, p, rv_win, fee=FEE):
+    rb = run_rv("BTC", p, rv_win, fee); re = run_rv("ETH", p, rv_win, fee)
+    mnf = min(rb["full"]["sharpe"], re["full"]["sharpe"])
+    mnh = min(rb["holdout"]["sharpe"], re["holdout"]["sharpe"])
+    mnt = min(rb["full"]["n_trades"], re["full"]["n_trades"])
+    print(f"  [{label} rv_win={rv_win}] minFull={mnf:+.2f} minHold={mnh:+.2f} minTr={mnt}  "
+          f"BTC(F{rb['full']['sharpe']:+.2f}/H{rb['holdout']['sharpe']:+.2f}/DD{rb['full']['maxdd']*100:.0f}%/n{rb['full']['n_trades']}) "
+          f"ETH(F{re['full']['sharpe']:+.2f}/H{re['holdout']['sharpe']:+.2f}/DD{re['full']['maxdd']*100:.0f}%/n{re['full']['n_trades']})")
+    return mnf, mnh, mnt, rb, re
+
+
+if __name__ == "__main__" and "--marginal" not in sys.argv:
+    # V1 geometry as the base (best known config). Sweep rv_win and the vola band.
+    base = dict(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+
+    print("=== SKH_R_RV: realized-vol BuzVola gate (V1 geometry) ===")
+    print("-- sweep rv_win (Chande lookback on annualized realized vol), V1 bands --")
+    grid = []
+    for rv_win in (8, 13, 20, 34):
+        p = SkyhookParams(**base)
+        mnf, mnh, mnt, rb, re = minrep("rvband", p, rv_win)
+        grid.append((rv_win, mnf, mnh, mnt))
+
+    # pick best holdout among feasible (minFull>=0.5, minTr>=20)
+    feas = [g for g in grid if g[1] >= 0.5 and g[3] >= 20]
+    pool = feas if feas else grid
+    best = max(pool, key=lambda g: g[2])
+    best_rv = best[0]
+    print(f"\n-- best rv_win by minHold (feasible): rv_win={best_rv} (minFull={best[1]:+.2f} minHold={best[2]:+.2f}) --")
+
+    # Around best rv_win, try widening the vola band (RV distribution may differ from ATR's)
+    print("\n-- band variations at best rv_win --")
+    bandvars = [
+        ("V1band", dict(vola_lo=35.0, vola_hi=95.0)),
+        ("wide",   dict(vola_lo=25.0, vola_hi=98.0)),
+        ("midhi",  dict(vola_lo=45.0, vola_hi=95.0)),
+        ("nogate", dict(vola_lo=0.0,  vola_hi=100.0)),
+    ]
+    cand = []
+    for nm, bd in bandvars:
+        pp = SkyhookParams(**{**base, **bd})
+        mnf, mnh, mnt, rb, re = minrep(nm, pp, best_rv)
+        cand.append((nm, bd, mnf, mnh, mnt))
+
+    feas2 = [c for c in cand if c[2] >= 0.5 and c[4] >= 20]
+    pool2 = feas2 if feas2 else cand
+    win = max(pool2, key=lambda c: c[3])
+    win_p = SkyhookParams(**{**base, **win[1]})
+    print(f"\n=== WINNER: band={win[0]} rv_win={best_rv}  minFull={win[2]:+.2f} minHold={win[3]:+.2f} ===")
+
+    # Causality on winner (both assets)
+    cb = causality_rv(win_p, best_rv, "BTC")
+    ce = causality_rv(win_p, best_rv, "ETH")
+    causal_ok = cb["ok"] and ce["ok"]
+    print(f"causality: BTC={cb} ETH={ce} -> ok={causal_ok}")
+
+    # Fee sweep on winner (min-asset full sharpe at each fee)
+    print("\n-- fee sweep (min-asset FULL sharpe) --")
+    fee_row = {}
+    for f in (0.0, 0.001, 0.002, 0.003):
+        rb = run_rv("BTC", win_p, best_rv, f); re = run_rv("ETH", win_p, best_rv, f)
+        fee_row[f"{f*100:.2f}%RT"] = round(min(rb["full"]["sharpe"], re["full"]["sharpe"]), 3)
+    print("  ", fee_row)
+    fee_survives = fee_row.get("0.30%RT", -9) > 0
+
+    # Marginal vs TP01 on winner. Build daily 50/50 series the same way skyhooklib does.
+    def daily_returns_rv(asset):
+        r = run_rv(asset, win_p, best_rv, FEE)
+        s = pd.Series(r["_eq"], index=r["_idx"])
+        return s.resample("1D").last().ffill().pct_change().dropna()
+    import altlib as al
+    sb = daily_returns_rv("BTC"); se = daily_returns_rv("ETH")
+    J = pd.concat({"BTC": sb, "ETH": se}, axis=1, join="inner").fillna(0.0)
+    cand_daily = 0.5 * J["BTC"] + 0.5 * J["ETH"]
+    mg = al.marginal_vs_tp01(cand_daily)
+    print("\n-- marginal vs TP01 --")
+    print(f"  corr_full={mg.get('corr_full')} verdict={mg.get('marginal_verdict')} "
+          f"w25_uplift_hold={mg.get('blends',{}).get('w25',{}).get('uplift_hold')} "
+          f"clean_year_uplift={mg.get('clean_year_uplift')} has_insample_edge={mg.get('has_insample_edge')} "
+          f"is_hedge={mg.get('is_hedge')} robust_oos={mg.get('robust_oos')}")
+
+    # Final per-asset detail at winner
+    rb = run_rv("BTC", win_p, best_rv); re = run_rv("ETH", win_p, best_rv)
+    print("\n=== FINAL (winner) ===")
+    for a, r in (("BTC", rb), ("ETH", re)):
+        f, h = r["full"], r["holdout"]
+        print(f"  {a}: FULL Sh={f['sharpe']:+.2f} ret={f['ret']*100:+.0f}% DD={f['maxdd']*100:.0f}% "
+              f"n={f['n_trades']} wr={f['win_rate']:.0f}% | HOLD Sh={h['sharpe']:+.2f} ret={h['ret']*100:+.0f}% "
+              f"DD={h['maxdd']*100:.0f}% n_entries={r['n_entries']}")
+    import json
+    summary = dict(label="R_RV", rv_win=best_rv, band=win[0], band_params=win[1],
+                   min_full=round(win[2], 3), min_hold=round(win[3], 3), min_trades=int(win[4]),
+                   btc_full=rb["full"]["sharpe"], eth_full=re["full"]["sharpe"],
+                   btc_hold=rb["holdout"]["sharpe"], eth_hold=re["holdout"]["sharpe"],
+                   avg_dd=round((rb["full"]["maxdd"] + re["full"]["maxdd"]) / 2, 4),
+                   causality_ok=causal_ok, fee_survives=fee_survives,
+                   corr_to_tp01=mg.get("corr_full"),
+                   blend_w25_uplift_hold=mg.get("blends", {}).get("w25", {}).get("uplift_hold"),
+                   marginal_verdict=mg.get("marginal_verdict"))
+    print("\nJSON " + json.dumps(summary, default=str))
+
+
+def _full_marginal():
+    """Re-run winner and dump the COMPLETE marginal dict + per-year, for the final report."""
+    import json, altlib as al
+    base = dict(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=45.0, vola_hi=95.0, vol_lo=0.0)
+    p = SkyhookParams(**base); rv_win = 34
+    def daily(asset):
+        r = run_rv(asset, p, rv_win, FEE)
+        s = pd.Series(r["_eq"], index=r["_idx"])
+        return s.resample("1D").last().ffill().pct_change().dropna()
+    J = pd.concat({"BTC": daily("BTC"), "ETH": daily("ETH")}, axis=1, join="inner").fillna(0.0)
+    cd = 0.5 * J["BTC"] + 0.5 * J["ETH"]
+    mg = al.marginal_vs_tp01(cd)
+    print("FULL MARGINAL:", json.dumps({k: mg.get(k) for k in
+          ("corr_full","corr_hold","marginal_verdict","has_insample_edge","is_hedge",
+           "robust_oos","clean_year_uplift","jackknife_min_uplift","multicut_persistent",
+           "multicut_uplift","cand_full_sharpe","cand_hold_sharpe","alpha_ann","resid_sharpe_full",
+           "null_pctl_insample")}, default=str))
+    print("BLENDS:", json.dumps(mg.get("blends"), default=str))
+    # per-year via backtest yearly on each asset
+    for a in ("BTC","ETH"):
+        ltf, htf = sk.frames(a); ent = rv_entries(ltf, htf, p, rv_win)
+        m = backtest_signals(ltf, ent, fee_rt=FEE, leverage=1.0, asset=a, tf="230m")
+        yr = " ".join(f"{int(y)}:{v*100:+.0f}%" for y, v in m.yearly.items())
+        print(f"  {a} per-year: {yr}")
+
+if __name__ == "__main__" and "--marginal" in sys.argv:
+    _full_marginal()

scripts/research/skyhook/skyhooklib.py

@@ -0,0 +1,217 @@
+"""skyhooklib — SHARED HONEST EVAL for the Skyhook (SKH01) multi-agent improvement wave.
+
+Every agent imports THIS so results are comparable and leak-free:
+  * data builders: certified 5m BTC/ETH -> 230m (exec) + 690m (signal), cached.
+  * study(): FULL + HOLD-OUT (2025-01-01+) + fee sweep + per-year, on BOTH assets, via the
+    project's honest intrabar engine (backtest_signals: TP/SL/max_bars, non-overlap).
+  * causality(): truncated-prefix guard (a Skyhook entry on a prefix must match the full run).
+  * marginal(): does Skyhook ADD to the existing TP01 portfolio? (altlib.marginal_vs_tp01).
+  * verdict(): conservative PASS/WEAK/FAIL on min-asset FULL & HOLD-OUT + fee survival.
+
+Quick start (inside an agent script):
+    import sys; sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+    import skyhooklib as sk
+    from src.strategies.skyhook import SkyhookParams
+    rep = sk.study("MY-VARIANT", SkyhookParams(ptn_n=20, sl_atr=2.5))
+    print(sk.fmt(rep)); print(sk.as_json(rep))
+"""
+from __future__ import annotations
+
+import json
+import sys
+from functools import lru_cache
+from pathlib import Path
+
+import numpy as np
+import pandas as pd
+
+_ROOT = Path(__file__).resolve().parents[3]
+if str(_ROOT) not in sys.path:
+    sys.path.insert(0, str(_ROOT))
+sys.path.insert(0, str(_ROOT / "scripts" / "research" / "alt"))
+
+from src.backtest.harness import backtest_signals  # noqa: E402
+from src.data.downloader import load_data  # noqa: E402
+from src.strategies.skyhook import (  # noqa: E402
+    SkyhookParams, build_frames, skyhook_entries, signal_counts)
+
+HOLDOUT = pd.Timestamp("2025-01-01", tz="UTC")
+FEE_RT = 0.001                       # 0.10% round-trip (Deribit taker)
+FEE_SWEEP = (0.0, 0.001, 0.002, 0.003)   # round-trip fee grid
+CERTIFIED = ("BTC", "ETH")
+
+
+@lru_cache(maxsize=4)
+def _frames(asset: str):
+    return build_frames(load_data(asset, "5m"))
+
+
+def frames(asset: str):
+    """(ltf 230m, htf 690m) certificati e cached."""
+    return _frames(asset.upper())
+
+
+def _split_metrics(eq: np.ndarray, idx: pd.DatetimeIndex, mask: np.ndarray) -> dict:
+    e = eq[mask]
+    if len(e) < 5:
+        return dict(sharpe=0.0, ret=0.0, maxdd=0.0, n=int(len(e)))
+    r = np.diff(e) / e[:-1]
+    r = r[np.isfinite(r)]
+    ix = idx[mask]
+    dt = pd.Series(ix).diff().dt.total_seconds().median() or 86400
+    bpy = 86400 * 365.25 / dt
+    sharpe = float(np.mean(r) / np.std(r) * np.sqrt(bpy)) if len(r) and np.std(r) > 0 else 0.0
+    pk = np.maximum.accumulate(e)
+    dd = float(np.max((pk - e) / pk))
+    return dict(sharpe=round(sharpe, 3), ret=round(float(e[-1] / e[0] - 1), 4),
+                maxdd=round(dd, 4), n=int(len(e)))
+
+
+def run_asset(asset: str, p: SkyhookParams, fee_rt: float = FEE_RT) -> dict:
+    """Backtest Skyhook su un asset (230m exec). Ritorna FULL+HOLDOUT+per-anno+diagnostica."""
+    ltf, htf = frames(asset)
+    entries = skyhook_entries(ltf, htf, p)
+    m = backtest_signals(ltf, entries, fee_rt=fee_rt, leverage=1.0, asset=asset, tf="230m")
+    eq = m.equity
+    idx = pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True))
+    hmask = np.asarray(idx >= HOLDOUT)
+    full = dict(sharpe=round(m.sharpe, 3), cagr=round(m.cagr, 4), maxdd=round(m.max_dd, 4),
+                ret=round(m.net_return, 4), n_trades=int(m.n_trades), win_rate=round(m.win_rate, 1))
+    hold = _split_metrics(eq, idx, hmask)
+    counts = signal_counts(ltf, htf, p)
+    return dict(asset=asset, full=full, holdout=hold,
+                yearly={int(y): round(v, 4) for y, v in m.yearly.items()},
+                counts=counts, _eq=eq, _idx=idx)
+
+
+def daily_returns(asset: str, p: SkyhookParams, fee_rt: float = FEE_RT) -> pd.Series:
+    """Rendimenti GIORNALIERI dell'equity Skyhook (per il lens marginal-vs-TP01).
+    NB approssimazione: l'equity di backtest_signals e' marcata a fine-trade (a gradini),
+    quindi i daily sono grezzi -> usalo SOLO per corr/uplift, non come headline Sharpe."""
+    r = run_asset(asset, p, fee_rt)
+    s = pd.Series(r["_eq"], index=r["_idx"])
+    return (s.resample("1D").last().ffill().pct_change().dropna())
+
+
+def skyhook_daily_5050(p: SkyhookParams, fee_rt: float = FEE_RT) -> pd.Series:
+    """Serie giornaliera 50/50 BTC+ETH (stessa convenzione di altlib.tp01_baseline_daily)."""
+    series = {a: daily_returns(a, p, fee_rt) for a in CERTIFIED}
+    J = pd.concat(series, axis=1, join="inner").fillna(0.0)
+    return 0.5 * J[CERTIFIED[0]] + 0.5 * J[CERTIFIED[1]]
+
+
+def marginal(p: SkyhookParams, fee_rt: float = FEE_RT) -> dict:
+    """Skyhook MIGLIORA il portafoglio TP01 esistente? (altlib.marginal_vs_tp01)."""
+    import altlib as al
+    return al.marginal_vs_tp01(skyhook_daily_5050(p, fee_rt))
+
+
+# ---------------------------------------------------------------------------
+# Causality guard (truncated-prefix): un ingresso emesso su un prefisso deve coincidere
+# con lo stesso indice della run completa (nessuna feature guarda il futuro).
+# ---------------------------------------------------------------------------
+def causality(p: SkyhookParams, asset: str = "BTC", tail: int = 200) -> dict:
+    ltf, htf = frames(asset)
+    full = skyhook_entries(ltf, htf, p)
+    n = len(ltf)
+    bad = 0; checked = 0
+    for frac in (0.80, 0.92):
+        cut = int(n * frac)
+        # taglia anche l'HTF alla stessa data di chiusura del prefisso LTF
+        cut_ts = int(ltf["timestamp"].iloc[cut - 1])
+        htf_cut = htf[htf["timestamp"] <= cut_ts].reset_index(drop=True)
+        sub = skyhook_entries(ltf.iloc[:cut].reset_index(drop=True), htf_cut, p)
+        for i in range(max(0, cut - tail), cut):
+            checked += 1
+            a, b = full[i], sub[i]
+            if (a is None) != (b is None):
+                bad += 1
+            elif a is not None and (a["dir"] != b["dir"]
+                                    or abs(a["sl"] - b["sl"]) > 1e-6
+                                    or abs(a["tp"] - b["tp"]) > 1e-6):
+                bad += 1
+    return dict(ok=bool(bad == 0), mismatches=int(bad), checked=int(checked))
+
+
+# ---------------------------------------------------------------------------
+# Verdict + drivers
+# ---------------------------------------------------------------------------
+def _verdict(per_asset: dict, fee_survives: bool) -> dict:
+    min_full = min(per_asset[a]["full"]["sharpe"] for a in per_asset)
+    min_hold = min(per_asset[a]["holdout"]["sharpe"] for a in per_asset)
+    min_trades = min(per_asset[a]["full"]["n_trades"] for a in per_asset)
+    enough = min_trades >= 20
+    pass_ = enough and min_full >= 0.5 and min_hold >= 0.2 and fee_survives
+    weak = enough and min_full >= 0.3 and min_hold >= 0.0
+    grade = "PASS" if pass_ else ("WEAK" if weak else "FAIL")
+    return dict(grade=grade, min_asset_full_sharpe=round(min_full, 3),
+                min_asset_holdout_sharpe=round(min_hold, 3),
+                min_trades=int(min_trades), fee_survives=bool(fee_survives))
+
+
+def study(name: str, p: SkyhookParams | None = None, assets=CERTIFIED,
+          fee_sweep=FEE_SWEEP) -> dict:
+    """Run completo: FULL+HOLDOUT+fee-sweep+per-anno su BTC&ETH + verdict conservativo."""
+    p = p or SkyhookParams()
+    per_asset = {}
+    fee_ok_all = True
+    for a in assets:
+        r = run_asset(a, p, FEE_RT)
+        sweep = {}
+        for f in fee_sweep:
+            rf = run_asset(a, p, f)
+            sweep[f"{f*100:.2f}%RT"] = rf["full"]["sharpe"]
+        fee_ok = sweep.get("0.30%RT", -9) > 0
+        fee_ok_all = fee_ok_all and fee_ok
+        per_asset[a] = dict(full=r["full"], holdout=r["holdout"], yearly=r["yearly"],
+                            counts=r["counts"], fee_sweep=sweep)
+    return dict(name=name, params=_params_dict(p), per_asset=per_asset,
+                verdict=_verdict(per_asset, fee_ok_all))
+
+
+def _params_dict(p: SkyhookParams) -> dict:
+    return {k: getattr(p, k) for k in p.__dataclass_fields__}
+
+
+# ---------------------------------------------------------------------------
+# Output
+# ---------------------------------------------------------------------------
+def _clean(o):
+    if isinstance(o, dict):
+        return {k: _clean(v) for k, v in o.items() if not k.startswith("_")}
+    if isinstance(o, (list, tuple)):
+        return [_clean(x) for x in o]
+    if isinstance(o, (np.floating,)):
+        return round(float(o), 4)
+    if isinstance(o, (np.integer,)):
+        return int(o)
+    if isinstance(o, (np.bool_,)):
+        return bool(o)
+    return o
+
+
+def as_json(rep: dict) -> str:
+    return json.dumps(_clean(rep), default=str)
+
+
+def fmt(rep: dict) -> str:
+    v = rep["verdict"]
+    lines = [f"=== {rep['name']} -> {v['grade']}  "
+             f"(minFull={v['min_asset_full_sharpe']:+.2f} minHold={v['min_asset_holdout_sharpe']:+.2f} "
+             f"minTrades={v['min_trades']} feeOK={v['fee_survives']})"]
+    for a, pa in rep["per_asset"].items():
+        f, h, c = pa["full"], pa["holdout"], pa["counts"]
+        yr = " ".join(f"{y}:{r*100:+.0f}%" for y, r in pa["yearly"].items())
+        lines.append(f"  {a}: FULL Sh={f['sharpe']:+.2f} ret={f['ret']*100:+.0f}% DD={f['maxdd']*100:.0f}% "
+                     f"n={f['n_trades']} wr={f['win_rate']:.0f}%  HOLD Sh={h['sharpe']:+.2f} ret={h['ret']*100:+.0f}% "
+                     f"| entries={c['entries']} (L{c['comp_long']}/S{c['comp_short']})")
+        lines.append(f"      fee sweep: " + " ".join(f"{k}={val:+.2f}" for k, val in pa["fee_sweep"].items()))
+        lines.append(f"      per-anno: {yr}")
+    return "\n".join(lines)
+
+
+if __name__ == "__main__":
+    print("--- SMOKE skyhooklib: baseline SkyhookParams() ---")
+    rep = study("SKH01-BASELINE", SkyhookParams())
+    print(fmt(rep))
+    print("\ncausality:", causality(SkyhookParams()))

scripts/research/skyhook/sweep.py

@@ -0,0 +1,46 @@
+"""Fast inline lever-scout for Skyhook before the agent wave. One fee (0.10% RT), both assets,
+min-asset FULL & HOLD-OUT Sharpe. Maps which knobs move the honest hold-out."""
+import sys
+from dataclasses import replace
+
+sys.path.insert(0, "/opt/docker/PythagorasGoal/scripts/research/skyhook")
+import skyhooklib as sk
+from src.strategies.skyhook import SkyhookParams
+
+
+def quick(p: SkyhookParams) -> dict:
+    rs = {a: sk.run_asset(a, p, sk.FEE_RT) for a in sk.CERTIFIED}
+    mf = min(rs[a]["full"]["sharpe"] for a in rs)
+    mh = min(rs[a]["holdout"]["sharpe"] for a in rs)
+    mt = min(rs[a]["full"]["n_trades"] for a in rs)
+    avg_dd = sum(rs[a]["full"]["maxdd"] for a in rs) / 2
+    return dict(minFull=mf, minHold=mh, minTr=mt, dd=round(avg_dd, 3),
+                btc=rs["BTC"]["full"]["sharpe"], eth=rs["ETH"]["full"]["sharpe"],
+                btcH=rs["BTC"]["holdout"]["sharpe"], ethH=rs["ETH"]["holdout"]["sharpe"])
+
+
+base = SkyhookParams()
+GRID = {
+    "long_only": [dict(long_only=True), dict(long_only=False)],
+    "ptn_n": [dict(ptn_n=n) for n in (8, 13, 20, 34, 55)],
+    "RR(sl,tp)": [dict(sl_atr=s, tp_atr=t) for s, t in
+                  ((1.5, 4.0), (2.0, 5.0), (2.5, 6.0), (3.0, 4.0), (2.0, 8.0), (3.0, 9.0))],
+    "exitbars": [dict(uscitalong=l, uscitashort=s) for l, s in
+                 ((12, 9), (24, 18), (36, 24), (48, 36))],
+    "vola_band": [dict(vola_lo=lo, vola_hi=hi) for lo, hi in
+                  ((0, 100), (35, 95), (50, 100), (50, 90), (20, 80))],
+    "vol_band": [dict(vol_lo=lo, vol_hi=hi) for lo, hi in
+                 ((0, 100), (50, 100), (60, 100), (40, 100), (50, 80))],
+}
+
+print(f"{'param':<14s} {'value':<28s} {'minF':>6s} {'minH':>6s} {'btc/eth F':>12s} {'btc/eth H':>12s} {'tr':>5s} {'dd':>5s}")
+print(f"  BASELINE: {quick(base)}")
+print("-" * 100)
+for fam, variants in GRID.items():
+    for v in variants:
+        p = replace(base, **v)
+        q = quick(p)
+        tag = "PASS" if (q["minFull"] >= 0.5 and q["minHold"] >= 0.2) else ""
+        print(f"{fam:<14s} {str(v):<28s} {q['minFull']:>+6.2f} {q['minHold']:>+6.2f} "
+              f"{q['btc']:>+5.2f}/{q['eth']:>+5.2f} {q['btcH']:>+5.2f}/{q['ethH']:>+5.2f} "
+              f"{q['minTr']:>5d} {q['dd']*100:>4.0f}% {tag}")

scripts/research/stops_lab.py

@@ -0,0 +1,101 @@
+"""STOPS LAB — protezioni CLASSICHE (stop-loss) sul combo vs la guardia-DD morbida.
+
+Confronta su Sharpe/MaxDD/2022/CAGR/NumTrades/time-in-market:
+  - baseline
+  - guardia-DD morbida (de-risk a 0.4x, gia' scelta)
+  - TRAILING STOP duro (uscita TOTALE) a -4/-6/-8% dal picco, re-entry su nuovo massimo
+  - TRAILING STOP con re-entry su RECUPERO (DD < meta' soglia)
+  - STOP MENSILE (flat per il resto del mese se perdita mensile > X%)
+  - VOL STOP (de-risk se vol realizzata 30g > 90 pctl espandente)
+Tesi da verificare: lo stop DURO taglia il DD ma fa whipsaw nel grind 2022 (Sharpe/CAGR peggiori,
+piu' trade). Dati: combo_daily (cache). sqrt(252).
+"""
+import sys
+from pathlib import Path
+import numpy as np, pandas as pd
+ROOT = Path(__file__).resolve().parents[2]
+sys.path.insert(0, str(ROOT)); sys.path.insert(0, str(ROOT / "scripts" / "research"))
+from combo_yearly_report import combo_daily
+ANN = np.sqrt(252.0)
+
+
+def _sh(r): r = np.asarray(pd.Series(r).dropna(), float); return float(np.mean(r)/np.std(r)*ANN) if len(r)>5 and np.std(r)>0 else 0.0
+def _dd(r): eq=np.cumprod(1+np.asarray(r,float)); pk=np.maximum.accumulate(eq); return float(np.max((pk-eq)/pk)) if len(eq) else 0.0
+def _yr(r,y): return float(np.prod(1+r[r.index.year==y].values)-1) if (r.index.year==y).any() else 0.0
+def _cagr(r): r=r.dropna(); return (np.prod(1+r.values))**(252/len(r))-1
+def _ntr(expo): return int((np.abs(np.diff(expo, prepend=expo[0]))>1e-9).sum())   # cambi di esposizione = trade
+
+
+def expo_softguard(r, trig=0.04, lvl=0.4):
+    rv=r.values; eq=np.cumprod(1+rv); pk=np.maximum.accumulate(eq); e=np.ones(len(rv)); on=True
+    for i in range(1,len(rv)):
+        dd=(pk[i-1]-eq[i-1])/pk[i-1] if pk[i-1]>0 else 0
+        if dd>trig: on=False
+        if dd<trig*0.4: on=True
+        e[i]=1.0 if on else lvl
+    return e
+
+
+def expo_trailstop(r, trig=0.06, reentry="newhigh"):
+    """Stop DURO trailing (uscita TOTALE) — test EQUO: trigger/re-entry sul MERCATO (NAV sempre-
+    investito di riferimento, non sull'equity stoppata che si congela). Esci se DD del NAV > trig;
+    re-entry 'newhigh' = NAV torna al picco; 'recover' = DD del NAV < trig/2."""
+    rv=r.values; n=len(rv)
+    nav=np.cumprod(1+rv); pk=np.maximum.accumulate(nav); dd=(pk-nav)/pk
+    e=np.ones(n); on=True
+    for i in range(1,n):
+        d=dd[i-1]
+        if on and d>trig:
+            on=False
+        elif not on:
+            if reentry=="newhigh" and nav[i-1]>=pk[i-1]-1e-12: on=True
+            elif reentry=="recover" and d<trig*0.5: on=True
+        e[i]=1.0 if on else 0.0
+    return e
+
+
+def expo_monthly(r, trig=0.05):
+    e=np.ones(len(r)); idx=r.index; cur=None; mret=1.0; stopped=False
+    for i in range(len(r)):
+        ym=(idx[i].year,idx[i].month)
+        if ym!=cur: cur=ym; mret=1.0; stopped=False
+        if stopped: e[i]=0.0; continue
+        mret*=(1+r.values[i])
+        if mret-1 < -trig: stopped=True
+    return e
+
+
+def expo_volstop(r, win=30, pctl=0.90):
+    rv=pd.Series(r.values,index=r.index); v=rv.rolling(win,min_periods=15).std()*ANN
+    thr=v.expanding(min_periods=60).quantile(pctl).shift(1)
+    e=np.where((v.shift(1)>thr).values, 0.4, 1.0); e=np.nan_to_num(e,nan=1.0)
+    return e
+
+
+def show(name, r, expo):
+    g=pd.Series(expo*r.values,index=r.index)
+    tim=float((expo>1e-9).mean())*100
+    print(f"  {name:30} Sh {_sh(g):>5.2f}  MaxDD {_dd(g.values)*100:>4.1f}%  2022 {_yr(g,2022)*100:>+5.1f}%  "
+          f"CAGR {_cagr(g)*100:>+5.1f}%  trades {_ntr(expo):>4}  inMkt {tim:>3.0f}%")
+
+
+def main():
+    print("="*104); print("  STOPS LAB — protezioni classiche (SL) vs guardia-DD morbida (combo TP01+GTAA, 2019-26)"); print("="*104)
+    r=combo_daily()
+    print(f"\n  {'(esposizione media applicata ai rendimenti del combo)':<30}")
+    show("baseline (nessuna)", r, np.ones(len(r)))
+    print("  --- guardia-DD MORBIDA (de-risk a 0.4x) ---")
+    show("soft-guard -4%", r, expo_softguard(r,0.04))
+    print("  --- STOP-LOSS DURO (uscita totale, trailing dal picco) ---")
+    for t in (0.04,0.06,0.08):
+        show(f"trail-stop -{t*100:.0f}% (re:newhigh)", r, expo_trailstop(r,t,"newhigh"))
+    show("trail-stop -6% (re:recover)", r, expo_trailstop(r,0.06,"recover"))
+    print("  --- altri classici ---")
+    show("stop mensile -5%", r, expo_monthly(r,0.05))
+    show("vol-stop (30g, >90pctl)", r, expo_volstop(r))
+    print("\n  NB: lo stop DURO che taglia molto il DD di solito paga in Sharpe/CAGR e in n.trade (whipsaw")
+    print("      nel grind). Confronta col soft-guard: stessa protezione, meno whipsaw?")
+
+
+if __name__ == "__main__":
+    main()

scripts/research/tail_hedge_lab.py

@@ -0,0 +1,131 @@
+"""TAIL-HEDGE LAB — proteggere il DD del combo (TP01+GTAA) e gli anni tipo 2022. Valuta OPZIONI.
+
+Il DD del combo (8.4%) e' grind-lento (2022), non crash. Le PUT proteggono i CRASH (rischio latente
+gap/overnight + leva), non il grind. Quindi confronto:
+  (A) OPZIONI: put-spread LONG su indice 50/50 BTC/ETH (mirror di VRP01), always-on vs GATED (hold
+      l'hedge solo quando esposti al trend / quando IV economica). Premio BS su DVOL reale, payoff sul
+      path. Misura il BLEED nei calmi + il payoff nei crash (+ stress sintetico -30% overnight).
+  (B) GUARDIA DRAWDOWN: de-risk il combo quando il DD da picco supera X% (o vol spike).
+  (C) VOL-TARGET: cappa la vol del combo a un livello piu' basso.
+Metriche: Sharpe, MaxDD, anno 2022, drag medio nei calmi, e payoff a uno shock -30%.
+"""
+import sys
+from pathlib import Path
+import numpy as np, pandas as pd
+ROOT = Path(__file__).resolve().parents[2]
+sys.path.insert(0, str(ROOT)); sys.path.insert(0, str(ROOT / "scripts" / "research"))
+from combo_yearly_report import combo_daily
+from src.data.downloader import load_data
+from src.strategies.trend_portfolio import resample_1d
+from src.portfolio.sleeves import _bs_put, _strike_from_delta, _HL_DIR, _tp01_returns
+from scipy.stats import norm
+
+ANN = np.sqrt(252.0)
+
+
+def _sh(r): r = np.asarray(pd.Series(r).dropna(), float); return float(np.mean(r)/np.std(r)*ANN) if len(r)>5 and np.std(r)>0 else 0.0
+def _dd(r): eq=np.cumprod(1+np.asarray(r,float)); pk=np.maximum.accumulate(eq); return float(np.max((pk-eq)/pk)) if len(eq) else 0.0
+def _yr(r,y): return float(np.prod(1+r[r.index.year==y].values)-1) if (r.index.year==y).any() else 0.0
+
+
+def crypto_index():
+    """Indice 50/50 BTC/ETH (prezzo daily) + DVOL medio, per il pricing dell'hedge."""
+    px={}; dv={}
+    for a in ("BTC","ETH"):
+        df=resample_1d(load_data(a,"1h")); s=pd.Series(df["close"].values.astype(float),index=pd.to_datetime(df["datetime"]))
+        if s.index.tz is None: s.index=s.index.tz_localize("UTC")
+        px[a]=s
+        d=pd.read_parquet(_HL_DIR/f"dvol_{a.lower()}.parquet")
+        dv[a]=pd.Series(d["close"].values.astype(float),index=pd.to_datetime(d["timestamp"],unit="ms",utc=True))
+    P=pd.concat(px,axis=1,join="inner").dropna(); D=pd.concat(dv,axis=1,join="inner").dropna()
+    idx_ret=0.5*P["BTC"].pct_change()+0.5*P["ETH"].pct_change()
+    dvol=(0.5*D["BTC"]+0.5*D["ETH"]).reindex(P.index).ffill()/100.0
+    return idx_ret.dropna(), dvol
+
+
+def hedge_overlay(idx_ret, dvol, kind="spread", buy_delta=-0.30, sell_delta=-0.10, tenor_d=7, gate_ivr=None):
+    """P&L settimanale per unita' di nozionale hedge di una protezione LONG comprata ogni settimana
+    sull'indice. kind='spread' = put-spread debit (compra strike vicino buy_delta, vende deep-OTM
+    sell_delta); kind='put' = long put nuda (buy_delta). gate_ivr: compra solo se IV-rank < soglia.
+    Ritorna (pnl_daily, premio_annuo_per_unita)."""
+    idx=idx_ret.index; r=idx_ret.values; sig=dvol.reindex(idx).ffill().values
+    S=np.cumprod(1+r); n=len(S)
+    ivr=pd.Series(sig,index=idx).rolling(252,min_periods=60).apply(lambda x:(x[-1]>=x).mean(),raw=True).values
+    out=np.zeros(n); prem_tot=0.0; i=30; T=tenor_d/365.25
+    while i+tenor_d<n:
+        if gate_ivr is not None and (not np.isfinite(ivr[i]) or ivr[i]>gate_ivr):
+            i+=tenor_d; continue
+        S0=S[i]; vol=sig[i]; ST=S[i+tenor_d]
+        Kb=_strike_from_delta(S0,T,vol,buy_delta)
+        if kind=="put":
+            prem=_bs_put(S0,Kb,T,vol); payoff=max(Kb-ST,0)
+        else:
+            Ks=_strike_from_delta(S0,T,vol,sell_delta)               # Kb>Ks (compra vicino, vende lontano)
+            prem=_bs_put(S0,Kb,T,vol)-_bs_put(S0,Ks,T,vol); payoff=max(Kb-ST,0)-max(Ks-ST,0)
+        fee=0.0005*S0
+        out[i+tenor_d]=(payoff-prem-fee)/S0; prem_tot+=(prem+fee)/S0
+        i+=tenor_d
+    yrs=(idx[-1]-idx[0]).days/365.25
+    return pd.Series(out,index=idx), (prem_tot/yrs if yrs>0 else 0.0)
+
+
+def dd_guard(combo, dd_trigger=0.05, look=20):
+    """De-risk: se il DD da picco supera dd_trigger -> esposizione 0.5 finche' non recupera meta'."""
+    r=combo.values; n=len(r); eq=np.cumprod(1+r); pk=np.maximum.accumulate(eq)
+    expo=np.ones(n); on=True
+    for i in range(1,n):
+        ddi=(pk[i-1]-eq[i-1])/pk[i-1]
+        if ddi>dd_trigger: on=False
+        if ddi<dd_trigger*0.4: on=True
+        expo[i]=1.0 if on else 0.4
+    return pd.Series(expo*r,index=combo.index)
+
+
+def voltarget(combo, tv=0.07):
+    rv=combo.rolling(30,min_periods=15).std().shift(1)*ANN
+    sc=np.clip(np.nan_to_num(tv/rv.replace(0,np.nan).values,nan=0.0),0,1.5)
+    return pd.Series(combo.values*sc,index=combo.index)
+
+
+def line(name, r, base=None):
+    r=r.dropna()
+    extra=""
+    if base is not None:
+        b=base.reindex(r.index).dropna(); r2=r.reindex(b.index)
+        extra=f"  Δ2022 {(_yr(r2,2022)-_yr(b,2022))*100:+.1f}pp"
+    print(f"  {name:34} Sh {_sh(r):>5.2f}  MaxDD {_dd(r.values)*100:>4.1f}%  2022 {_yr(r,2022)*100:>+5.1f}%  "
+          f"CAGR {((np.prod(1+r.values))**(252/len(r))-1)*100:>+5.1f}%{extra}")
+
+
+def main():
+    print("="*100); print("  TAIL-HEDGE LAB — proteggere DD/2022 del combo (TP01+GTAA)"); print("="*100)
+    combo=combo_daily()
+    idx_ret,dvol=crypto_index()
+    print("\n  BASELINE")
+    line("combo (1x)", combo)
+
+    print("\n  (A) OPZIONI — protezione LONG su crypto, sovrapposta al combo (size = spendi ~3%/anno)")
+    for kind,gate,lbl in [("spread",None,"put-spread sempre"),("spread",0.4,"put-spread gate IVR<0.4"),
+                          ("put",None,"long put -0.30 sempre"),("put",0.4,"long put gate IVR<0.4")]:
+        ov,annprem=hedge_overlay(idx_ret,dvol,kind=kind,gate_ivr=gate)
+        ov=ov.reindex(combo.index).fillna(0.0)
+        size=0.03/annprem if annprem>0 else 0.0                   # nozionale hedge per ~3%/anno di premio
+        line(f"+{lbl} (~3%/y, size {size:.2f}x)", combo+size*ov, base=combo)
+    print("  (stress -30% overnight dell'indice crypto, per unita' di nozionale hedge):")
+    S0=1.0; vol=float(dvol.iloc[-1]); T=7/365.25
+    for kind in ("spread","put"):
+        Kb=_strike_from_delta(S0,T,vol,-0.30)
+        if kind=="put": prem=_bs_put(S0,Kb,T,vol); pay=max(Kb-0.7,0)
+        else: Ks=_strike_from_delta(S0,T,vol,-0.10); prem=_bs_put(S0,Kb,T,vol)-_bs_put(S0,Ks,T,vol); pay=max(Kb-0.7,0)-max(Ks-0.7,0)
+        print(f"     {kind:7}: premio {prem*100:.2f}% -> payoff a -30% {pay*100:.2f}% (netto {(pay-prem)*100:+.2f}%)")
+
+    print("\n  (B) GUARDIA DRAWDOWN (de-risk a -X% dal picco)")
+    for t in (0.04,0.06):
+        line(f"+dd-guard {t*100:.0f}%", dd_guard(combo,t), base=combo)
+    print("\n  (C) VOL-TARGET del combo")
+    for tv in (0.05,0.07):
+        line(f"+vol-target {tv*100:.0f}%", voltarget(combo,tv), base=combo)
+
+
+if __name__ == "__main__":
+    main()

scripts/research/wf_crypto_lead.js

@@ -0,0 +1,145 @@
+export const meta = {
+  name: 'crypto-lead-sweep',
+  description: 'Sweep multi-agente dell\'anticipazione crypto->mercati IB (mercati x timing x lead), verifica avversariale multi-anno, sintesi della soluzione migliore',
+  phases: [
+    { title: 'Sweep', detail: '~52 agenti: grid (lead x mercato x giorno x predict x finestra) via harness onesto' },
+    { title: 'Verify', detail: 'top candidati: stress costi + OOS recente + multi-anno' },
+    { title: 'Synthesize', detail: 'migliore soluzione robusta + caveat di tradabilita' },
+  ],
+}
+
+// ---- universo mercati certificati (cache su disco) ----
+const MARKETS = ["SPY","QQQ","IWM","DIA","XLK","XLF","XLE","XLV","XLI","XLP","XLY","XLU","XLB","XLRE","XLC",
+                 "HYG","TLT","IEF","GLD","SLV","USO","DBC","VNQ","EEM","FXI","EWJ"]
+const LEADS = ["BTC","ETH"]
+const DAYS = ["all","mon"]
+const PREDICTS = ["gap","intraday"]
+const HOURS = ["overnight","6"]
+
+// ---- genera la grid completa ----
+const grid = []
+for (const lead of LEADS) for (const target of MARKETS) for (const day of DAYS)
+  for (const predict of PREDICTS) for (const hours of HOURS)
+    grid.push({ lead, target, day, predict, hours })
+log(`grid totale: ${grid.length} configurazioni (mercati ${MARKETS.length} x lead 2 x giorno 2 x predict 2 x finestra 2)`)
+
+// ---- chunk in batch da 8 -> ~52 agenti ----
+const BATCH = 8
+const batches = []
+for (let i = 0; i < grid.length; i += BATCH) batches.push(grid.slice(i, i + BATCH))
+log(`sweep: ${batches.length} agenti, ${BATCH} config ciascuno`)
+
+const RAW_SCHEMA = {
+  type: "object",
+  properties: { raw: { type: "string", description: "stdout JSON ESATTO dell'harness, senza modifiche" } },
+  required: ["raw"],
+}
+
+// ---- PHASE 1: SWEEP ----
+phase('Sweep')
+const sweepResults = await parallel(batches.map((batch, bi) => () =>
+  agent(
+    `Sei un esecutore deterministico. Esegui ESATTAMENTE questo comando dalla root del repo e restituisci il suo stdout.\n\n` +
+    `uv run python scripts/research/crypto_lead_harness.py --configs '${JSON.stringify(batch)}'\n\n` +
+    `Il comando stampa un array JSON di risultati. Mettilo VERBATIM nel campo "raw" (nessun commento, nessuna modifica ai numeri). ` +
+    `Se il comando fallisce, metti il messaggio d'errore in "raw".`,
+    { label: `sweep:${bi}`, phase: 'Sweep', schema: RAW_SCHEMA, effort: 'low' }
+  )
+))
+
+// ---- parse + flatten ----
+const all = []
+for (const r of sweepResults) {
+  if (!r || !r.raw) continue
+  try {
+    const arr = JSON.parse(r.raw.trim())
+    for (const x of arr) if (x && !x.err) all.push(x)
+  } catch (e) { /* skip batch non parsabile */ }
+}
+log(`risultati validi raccolti: ${all.length}/${grid.length}`)
+
+// ---- ranking robustezza: tutti gli anni positivi + t-stat alto + OOS positivo ----
+function score(x) {
+  const yr = x.years_tot ? x.years_pos / x.years_tot : 0
+  const oosOK = (x.sh_ls_oos > 0 || x.sh_lf_oos > 0) ? 1 : 0
+  return yr * Math.abs(x.t_incremental || 0) * (1 + Math.max(x.sh_ls_oos || 0, x.sh_lf_oos || 0)) * oosOK
+}
+all.sort((a, b) => score(b) - score(a))
+const top = all.slice(0, 12)
+// separa per tradabilita': intraday = ETF-tradabile; gap = fenomeno (serve future)
+const topIntraday = all.filter(x => x.predict === 'intraday').slice(0, 6)
+const topGap = all.filter(x => x.predict === 'gap').slice(0, 6)
+log(`TOP overall: ${top.map(x => `${x.lead}->${x.target}/${x.day}/${x.predict} t=${x.t_incremental} oos=${Math.max(x.sh_ls_oos,x.sh_lf_oos)} ${x.years_pos}/${x.years_tot}y`).join(' | ')}`)
+
+// ---- PHASE 2: VERIFY (stress costi + OOS recente, multi-anno) ----
+phase('Verify')
+const VERDICT_SCHEMA = {
+  type: "object",
+  properties: {
+    config: { type: "string" },
+    robust: { type: "boolean", description: "regge stress (costi alti + OOS recente) e multi-anno?" },
+    tradeable_via: { type: "string", description: "etf-intraday | futures-gap | none" },
+    sh_oos_4bps: { type: "number" }, sh_oos_10bps: { type: "number" }, sh_oos_recent: { type: "number" },
+    years_pos: { type: "number" }, years_tot: { type: "number" },
+    note: { type: "string", description: "1-2 frasi: cosa regge, cosa no, spiegazione alternativa" },
+  },
+  required: ["config", "robust", "tradeable_via", "note"],
+}
+const cands = [...new Map([...topIntraday, ...topGap, ...top].map(x => [`${x.lead}|${x.target}|${x.day}|${x.predict}|${x.hours}`, x])).values()].slice(0, 12)
+const verified = await parallel(cands.map((c) => () => {
+  const cfg = JSON.stringify([{ lead: c.lead, target: c.target, day: c.day, predict: c.predict, hours: c.hours }])
+  return agent(
+    `Verifica avversariale di UN candidato lead-lag crypto->mercato. Config: ${JSON.stringify(c)}.\n` +
+    `Esegui questi 3 comandi dalla root e leggi i campi t_incremental, sh_ls_oos, sh_lf_oos, years_pos, years_tot:\n` +
+    `1) base 4bps:   uv run python scripts/research/crypto_lead_harness.py --cost 0.0004 --oos 2022-01-01 --configs '${cfg}'\n` +
+    `2) stress 10bps: uv run python scripts/research/crypto_lead_harness.py --cost 0.0010 --oos 2022-01-01 --configs '${cfg}'\n` +
+    `3) OOS recente:  uv run python scripts/research/crypto_lead_harness.py --cost 0.0004 --oos 2024-01-01 --configs '${cfg}'\n\n` +
+    `Giudica: robust=true SOLO se l'edge resta positivo a 10bps E nell'OOS recente (2024+) E years_pos/years_tot>=0.6. ` +
+    `tradeable_via: "etf-intraday" se predict=intraday e regge (eseguibile comprando l'ETF al Monday/giorno open); ` +
+    `"futures-gap" se predict=gap e regge (il gap si cattura solo con i futures indice overnight, NON con l'ETF); "none" se non regge. ` +
+    `note: spiega anche un'alternativa plausibile (e' solo risk-beta? autocorrelazione? multiple-testing su ${grid.length} test?).`,
+    { label: `verify:${c.lead}->${c.target}/${c.predict}`, phase: 'Verify', schema: VERDICT_SCHEMA }
+  )
+}))
+const robust = verified.filter(v => v && v.robust)
+log(`candidati robusti: ${robust.length}/${cands.length}`)
+
+// ---- PHASE 3: SYNTHESIZE ----
+phase('Synthesize')
+const SYNTH_SCHEMA = {
+  type: "object",
+  properties: {
+    best_solution: { type: "string", description: "la migliore soluzione: lead+mercato+timing+predict+come si tradea" },
+    why: { type: "string" },
+    expected_edge: { type: "string", description: "Sharpe OOS onesto (post-stress), hit, anni positivi" },
+    tradeability: { type: "string", description: "ETF intraday vs futures overnight; eseguibile a $0.5-2k?" },
+    multi_year: { type: "string", description: "evidenza su piu' anni (per-anno)" },
+    caveats: { type: "string" },
+    runner_ups: { type: "string" },
+  },
+  required: ["best_solution", "why", "expected_edge", "tradeability", "multi_year", "caveats"],
+}
+const synthesis = await agent(
+  `Sei l'analista quant capo, disciplina ONESTA (questo progetto uccide i falsi positivi: multiple-testing, hold-out-luck, ` +
+  `tradabilita' reale a basso capitale). Obiettivo: dalla ricerca multi-agente sull'anticipazione crypto->mercati IB, ` +
+  `determina la SOLUZIONE MIGLIORE (mercato + timing + lead) verificata su PIU' ANNI.\n\n` +
+  `TOP candidati (sweep, ${grid.length} config testate): ${JSON.stringify(top)}\n\n` +
+  `VERDETTI di verifica avversariale (stress costi 10bps + OOS recente 2024+): ${JSON.stringify(verified)}\n\n` +
+  `Candidati robusti: ${JSON.stringify(robust)}\n\n` +
+  `Fatti noti: (a) il GAP di apertura (predict=gap) ha t-stat altissimi ma NON e' catturabile con l'ETF (serve un future ` +
+  `indice tradato overnight, es. MNQ/MES su IB); (b) l'INTRADAY (predict=intraday) e' debole ma eseguibile comprando l'ETF; ` +
+  `(c) abbiamo testato ${grid.length} configurazioni -> correggi mentalmente per multiple-testing (un t-stat ~2 non basta qui).\n\n` +
+  `Produci la sintesi: la soluzione migliore REALMENTE utile (distingui fenomeno-forte-non-tradabile da edge-tradabile), ` +
+  `il suo edge atteso onesto post-stress, come si tradea a $0.5-2k, l'evidenza multi-anno, i caveat, e i runner-up.`,
+  { label: 'synthesize', phase: 'Synthesize', schema: SYNTH_SCHEMA, effort: 'high' }
+)
+
+return {
+  grid_size: grid.length,
+  sweep_agents: batches.length,
+  results_collected: all.length,
+  top12: top,
+  verified,
+  robust_count: robust.length,
+  synthesis,
+}

src/live/dashboard.py

@@ -1,4 +1,4 @@
-"""DASHBOARD web del portafoglio attivo (TP01 + XS01) — monitoraggio PAPER, stdlib only.
+"""DASHBOARD web del portafoglio attivo (TP01 + XS01 + VRP01 + SKH01) — monitoraggio PAPER, stdlib only.
 
 Mostra: metriche (FULL/HOLD Sharpe, DD, CAGR), per-sleeve, posizioni correnti, equity (backtest +
 paper forward da scripts/live/paper_portfolio.py), ultima data dato. Nessuna auth -> solo rete
@@ -15,11 +15,13 @@ sys.path.insert(0, str(PROJECT_ROOT))
 import numpy as np, pandas as pd
 from src.portfolio.portfolio import StrategyPortfolio, metrics, HOLDOUT
 from src.portfolio.sleeves import active_sleeves
+from src.portfolio.gtaa import gtaa_weights
 from src.live.shadow import shadow_report, tp01_trades
 from src.version import APP_VERSION
 
 PAPER = PROJECT_ROOT / "data" / "paper_portfolio" / "state.json"
 PREVDAY = PROJECT_ROOT / "data" / "paper_prevday" / "state.json"
+COMBO = PROJECT_ROOT / "data" / "paper_combo" / "state.json"
 _CACHE = {"t": 0.0, "data": None}
 _TTL = 120.0
 
@@ -35,6 +37,11 @@ def build():
     spark = [(str(idx[i].date()), float(eq[i])) for i in range(0, len(eq), step)]
     paper = json.loads(PAPER.read_text()) if PAPER.exists() else None
     prevday = json.loads(PREVDAY.read_text()) if PREVDAY.exists() else None
+    combo = json.loads(COMBO.read_text()) if COMBO.exists() else None
+    try:
+        gtaa_w = gtaa_weights()                               # pesi ETF correnti azionabili (cache eq_*)
+    except Exception:
+        gtaa_w = None
     try:
         shadow = shadow_report()                              # mainnet sola lettura, best-effort
     except Exception as e:
@@ -49,6 +56,7 @@ def build():
         full=bt["full"], holdout=bt["holdout"], weights=bt["weights"],
         per_sleeve=bt["per_sleeve"], yearly=bt["yearly"],
         positions=pf.current_positions(), spark=spark, paper=paper, prevday=prevday,
+        combo=combo, gtaa_weights=gtaa_w,
         shadow=shadow, trades=trades, bh=None,
     )
     _CACHE.update(t=time.time(), data=data)
@@ -79,7 +87,19 @@ def html():
     yrs = "".join(f"<span class=y>{y}: {v['ret']*100:+.0f}%</span>" for y, v in sorted(d["yearly"].items()))
     pos = ""
     for sl, p in d["positions"].items():
-        pos += f"<tr><td>{sl}</td><td>{'flat (in cash)' if p == {'BTC': 0.0, 'ETH': 0.0} else (p if p is not None else 'stat-mode (book 19 gambe)')}</td></tr>"
+        if p == {'BTC': 0.0, 'ETH': 0.0}:
+            ptxt = 'flat (in cash)'
+        elif p is not None:
+            ptxt = str(p)
+        elif 'XS01' in sl:
+            ptxt = 'stat-mode (book 19 gambe)'
+        elif 'SKH' in sl:
+            ptxt = 'forward-monitor (segnale dual-TF, no pos-fn)'
+        elif 'VRP' in sl:
+            ptxt = 'stat-mode (book opzioni settimanale)'
+        else:
+            ptxt = 'n/d'
+        pos += f"<tr><td>{sl}</td><td>{ptxt}</td></tr>"
     pp = d["paper"]
     if pp:
         days = (pd.Timestamp(pp["last"]) - pd.Timestamp(pp["start"])).days
@@ -104,6 +124,28 @@ def html():
             f"forward da {pd.Timestamp(pv['start_ts'], unit='ms').date()}")
     else:
         prevday_html = "non inizializzato (gira <code>scripts/live/paper_prevday.py</code>)"
+    cb = d.get("combo")
+    if cb:
+        cdays = (pd.Timestamp(cb["last"]) - pd.Timestamp(cb["start"])).days
+        cret = cb["equity"] / cb["initial"] - 1
+        wc = cb.get("w_crypto", 0.5)
+        head = (f"start {cb['initial']:.0f} · {cb['start'][:10]} → {cb['last'][:10]} ({cdays}g, "
+                f"{cb['n_days']} barre) · blend {wc*100:.0f}/{(1-wc)*100:.0f} TP01/GTAA")
+        combo_html = f"{head}<br><b>NUDO</b> &nbsp; eq <b>{cb['equity']:.2f}</b> &nbsp; ret <b>{cret*100:+.2f}%</b> &nbsp; maxDD {cb['max_dd']*100:.1f}%"
+        if "equity_g" in cb:
+            crg = cb["equity_g"] / cb["initial"] - 1
+            combo_html += (f"<br><b>PROTETTO</b> (guardia-DD −{cb.get('dd_trigger',0.04)*100:.0f}%) &nbsp; "
+                           f"eq <b>{cb['equity_g']:.2f}</b> &nbsp; ret <b>{crg*100:+.2f}%</b> &nbsp; maxDD {cb['max_dd_g']*100:.1f}%")
+    else:
+        combo_html = "non inizializzato (gira <code>scripts/live/paper_combo.py</code>)"
+    gw = d.get("gtaa_weights")
+    if gw:
+        asof = gw.get("_asof", "?"); cash = gw.get("_cash")
+        gw_html = (", ".join(f"{k} <b>{v:.0%}</b>" for k, v in gw.items() if not k.startswith("_") and v)
+                   + (f" · cash <b>{cash:.0%}</b>" if cash is not None else "")
+                   + f" <span style='color:#8a93a0'>(asof {asof})</span>")
+    else:
+        gw_html = "n/d (cache ETF assente — gira fetch_ib_equities.py)"
     sh = d.get("shadow")
     if sh and "error" not in sh:
         bits = " &nbsp;·&nbsp; ".join(
@@ -156,7 +198,7 @@ th{{color:#8a93a0;font-weight:500}}.y{{display:inline-block;background:#161b22;b
 .warn{{color:#f1c40f;font-size:12px}}
 .section{{font-size:15px;font-weight:700;letter-spacing:.06em;text-transform:uppercase;margin:34px 0 14px;padding:10px 14px;border-radius:9px;background:#12181f;border-left:5px solid #2ecc71;color:#d7dee6}}
 .section.live{{border-left-color:#e74c3c;background:#1c1316;color:#f0c4c4}}</style></head><body>
-<h1>PythagorasGoal — Portafoglio attivo (TP01 + XS01 + VRP01)</h1>
+<h1>PythagorasGoal — Portafoglio attivo (TP01 + XS01 + VRP01 + SKH01)</h1>
 <div class=sub>monitor · v{d['version']} · ultimo dato {d['last_data']} · esecuzione REALE non attiva (solo micro-test)</div>
 <div class="section">PAPER — simulato (backtest + forward virtuale)</div>
 <div class=cards>
@@ -175,6 +217,10 @@ th{{color:#8a93a0;font-weight:500}}.y{{display:inline-block;background:#161b22;b
 <h3 style="font-size:14px;color:#8a93a0">Trades TP01 — entry/exit (segnale causale, ultimi 15)</h3>
 <table><tr><th>data</th><th>asset</th><th>azione</th><th>posizione</th><th>prezzo</th></tr>{trows}</table>
 <div style="margin-top:10px">{yrs}</div>
+<div class="section">COMBO DEPLOYABLE — cross-venue (paper, forward-only)</div>
+<div class=box><b>TP01 (Deribit) + GTAA (IB)</b> — le DUE gambe ESEGUIBILI a basso capitale, scorrelate (corr ~0.21) → blend Sharpe ~1.5, drawdown dimezzato. <b>Nessuna esecuzione reale</b>:<br>{combo_html}<br>
+<span style="color:#8a93a0;font-size:13px">posizioni azionabili IB (GTAA, peso ETF):</span> {gw_html}</div>
+<p class=warn>⚠️ PAPER cross-venue: valida l'operativita' su due conti (Deribit + IB) a rischio zero. Lo Sharpe ~1.5 e' ottimistico (finestra crypto corta/favorevole); il dato robusto e' la diversificazione (corr 0.21, DD dimezzato), non il livello. XS01/VRP01 esclusi (STAT-MODE): qui solo TP01+GTAA.</p>
 <div class="section">FORWARD-MONITOR — lead paper (non deploy)</div>
 <div class=box><b>PREVDAY range-breakout</b> — lead ORTOGONALE a TP01 (corr ~0.15 full / ~0 hold; marginal ADDS, non-hedge, robusto allo shift del confine-giorno). Forward-only, <b>nessuna esecuzione reale</b>:<br>{prevday_html}</div>
 <p class=warn>⚠️ LEAD in osservazione, NON deployato. Sopravvissuto alla verifica avversariale dell'onda intraday; lo teniamo in paper per validarlo fuori-campione-vero. I due libri (modeled vs real-$600) mostrano l'haircut di fill che lo scettico aveva segnalato.</p>
@@ -182,7 +228,7 @@ th{{color:#8a93a0;font-weight:500}}.y{{display:inline-block;background:#161b22;b
 <div class=box><b>Shadow TP01</b> (cosa farebbe ORA sul conto reale, nessun ordine inviato):<br>{shadow_html}</div>
 <h3 style="font-size:14px;color:#8a93a0">Trades REALI eseguiti su Deribit</h3>
 <table><tr><th>data/ora UTC</th><th>strum.</th><th>dir</th><th>amount</th><th>prezzo</th><th>fee USDC</th></tr>{live_trows}</table>
-<p class=warn>⚠️ Paper/monitor. XS01 e' STAT-MODE (book a 19 gambe market-neutral, non eseguibile a €2k, storia ~2.5 anni). VRP01 = lead short-vol MODELLATO (non deploy pieno). TP01 e' l'unico deployable pieno: lo "Shadow live" mostra cosa farebbe sul mainnet, ma NON invia ordini.</p>
+<p class=warn>⚠️ Paper/monitor. XS01 e' STAT-MODE (book a 19 gambe market-neutral, non eseguibile a €2k, storia ~2.5 anni). VRP01 = lead short-vol MODELLATO (non deploy pieno). SKH01 (Skyhook dual-TF regime+breakout, BTC/ETH) = diversificatore quasi-ortogonale (corr ~0.09) aggiunto @25%: alza il FULL Sharpe del portafoglio 1.68→2.13 e dimezza il DD (14→8%) — RESEARCH/forward-monitor (book a 230m, causalita' verificata su harness ma costi reali e codice d'esecuzione da validare prima del deploy). TP01 e' l'unico deployable pieno: lo "Shadow live" mostra cosa farebbe sul mainnet, ma NON invia ordini.</p>
 </body></html>"""
 
 

src/portfolio/gtaa.py

@@ -0,0 +1,73 @@
+"""GTAA — sleeve EQUITY/ETF: trend difensivo multi-asset (analogo di TP01, su IB).
+
+Validato sul branch research/equities-ib (diari 2026-06-22-equities-*). Trend long-flat TSMOM
+multi-orizzonte su un paniere di classi (azioni US/tech/small + bond + oro + credito), equal-weight
+sugli asset in trend-up, vol-target 12%. Difensivo: Sharpe ~0.64 / maxDD ~15% standalone, e — il punto
+— corr ~0.21 col crypto (TP01) -> diversificatore reale (blend Sharpe ~1.5, DD dimezzato).
+
+Eseguibile su IB a basso capitale (ETF frazionabili, switch mensile/basso turnover). Legge la CACHE
+su disco data/raw/eq_*.parquet (ADJUSTED_LAST, scritta da scripts/research/fetch_ib_equities.py);
+in produzione va rinfrescata giornalmente (gateway IB). Espone rendimenti + PESI CORRENTI (posizioni).
+"""
+from __future__ import annotations
+from functools import lru_cache
+from pathlib import Path
+import numpy as np, pandas as pd
+
+RAW = Path(__file__).resolve().parents[2] / "data" / "raw"
+EQ_UNIVERSE = ("SPY", "QQQ", "IWM", "TLT", "GLD", "HYG")
+HORIZONS = (21, 63, 126, 252)
+TARGET_VOL = 0.12
+FEE_SIDE = 0.0002
+ANN = np.sqrt(252.0)
+
+
+@lru_cache(maxsize=16)
+def _close(sym: str) -> pd.Series:
+    p = RAW / f"eq_{sym.lower()}_1d.parquet"
+    if not p.exists():
+        raise FileNotFoundError(f"{p} assente — gira scripts/research/fetch_ib_equities.py")
+    d = pd.read_parquet(p)
+    return pd.Series(d["close"].astype(float).values,
+                     index=pd.to_datetime(d["timestamp"], unit="ms", utc=True)).sort_index()
+
+
+def _exposure(close: pd.Series) -> pd.Series:
+    """Esposizione long-flat [0,1] su un asset: frazione di orizzonti in trend-up, vol-targeted, cap 1.
+    Causale (solo dati <= i)."""
+    px = close.values; n = len(px); tgt = np.zeros(n); mh = max(HORIZONS)
+    for i in range(mh, n):
+        tgt[i] = np.mean([1.0 if px[i] > px[i - H] else 0.0 for H in HORIZONS])
+    s = pd.Series(tgt, index=close.index)
+    rv = close.pct_change().rolling(63, min_periods=20).std().shift(1) * ANN
+    scale = np.clip(np.nan_to_num(TARGET_VOL / rv.replace(0, np.nan).values, nan=0.0), 0, 10.0)
+    return (s * scale).clip(0, 1.0)
+
+
+def _gated_returns(sym: str) -> pd.Series:
+    close = _close(sym); ex = _exposure(close)
+    ret = close.pct_change().fillna(0.0).values
+    held = np.zeros(len(ex)); held[1:] = ex.values[:-1]      # causale: esposizione decisa a i-1, tenuta in i
+    net = held * ret - FEE_SIDE * np.abs(np.diff(held, prepend=0.0))
+    return pd.Series(net, index=close.index)
+
+
+def gtaa_returns(universe=EQ_UNIVERSE) -> pd.Series:
+    """Rendimenti netti daily del GTAA: EW dei rendimenti trend-gated sugli asset disponibili."""
+    cols = {a: _gated_returns(a) for a in universe}
+    return pd.concat(cols, axis=1).sort_index().mean(axis=1, skipna=True).dropna()
+
+
+def gtaa_weights(universe=EQ_UNIVERSE) -> dict:
+    """Pesi target CORRENTI (ultima barra): quanto allocare a ciascun ETF (e quanto in cash).
+    weight_i = esposizione_i / N_disponibili. Azionabile su IB."""
+    out = {}; n = len(universe)
+    for a in universe:
+        try:
+            ex = _exposure(_close(a))
+            out[a] = round(float(ex.iloc[-1]) / n, 4)
+        except FileNotFoundError:
+            continue
+    out["_cash"] = round(1.0 - sum(out.values()), 4)
+    out["_asof"] = str(_close("SPY").index[-1].date())
+    return out

src/portfolio/sleeves.py

@@ -207,12 +207,49 @@ def vrp_sleeve(weight: float = 0.20) -> Sleeve:
     return Sleeve("VRP01_shortvol", weight, _vrp_combo_returns)
 
 
+# ----------------------------- SKH01-V2-DD: Skyhook dual-TF regime+breakout (BTC/ETH) -----------------------------
+# Sistema dual-timeframe (segnale 690m, exec 230m): entra solo quando coincidono REGIME
+# (BuzVola/BuzVolume tipo-Chande) E PATTERN (Donchian breakout). NON e' un trend-follower.
+# Vincitrice dell'onda DD-reduction (famiglia ASYM_LS): exit a percentuale fissa ASIMMETRICA
+# (long sl4%/tp10%, short sl2%/tp8% piu' stretto) -> taglia il DD standalone (BTC 21% / ETH 27%)
+# alzando hold-out (minHold +1.26) e valore di portafoglio. Quasi-ortogonale a TP01 (corr ~0.09):
+# blend 0.75*TP01+0.25*SKH -> hold-out Sharpe 0.31->1.17 (+0.87), DD full 14%->9%. Marginal ADDS,
+# has_insample_edge, robust_oos (multicut 7/7 anni), is_hedge=False. Verificato leak-free (causalita'
+# 0/400) + 2 scettici avversariali. Diario 2026-06-23-skyhook.md.
+# CAVEAT ONESTI: equity marcata a fine-trade (daily lumpy); ETH DD 27% ha margine sottile vs 30%;
+# il book opera a 230m -> ribilanciamento piu' frequente del resto (verificare costi reali a deploy).
+from src.strategies.skyhook import SKH01_V2_DD, build_frames, skyhook_entries
+from src.backtest.harness import backtest_signals
+
+
+def _skyhook_returns() -> pd.Series:
+    """SKH01-V2-DD: book 50/50 BTC+ETH del sistema regime+breakout dual-TF, riportato su griglia
+    GIORNALIERA. Causale (decide a close[i], exit intrabar TP/SL/max_bars, non-overlap), netto 0.10% RT."""
+    series = {}
+    for a in ASSETS:
+        ltf, htf = build_frames(load_data(a, "5m"))
+        ent = skyhook_entries(ltf, htf, SKH01_V2_DD)
+        m = backtest_signals(ltf, ent, fee_rt=0.001, leverage=1.0, asset=a, tf="230m")
+        s = pd.Series(m.equity, index=pd.DatetimeIndex(pd.to_datetime(m.eq_index, utc=True)))
+        series[a] = s.resample("1D").last().ffill().pct_change().dropna()
+    J = pd.concat(series, axis=1, join="inner").fillna(0.0)
+    return pd.Series(0.5 * J["BTC"].values + 0.5 * J["ETH"].values, index=J.index)
+
+
+def skyhook_sleeve(weight: float = 0.25) -> Sleeve:
+    return Sleeve("SKH01_skyhook", weight, _skyhook_returns)
+
+
 # ----------------------------- REGISTRY -----------------------------
 def active_sleeves() -> list[Sleeve]:
     """Sleeve ATTIVI nel portafoglio (pesi rinormalizzati; sleeve a date diverse si attivano
-    quando parte la loro storia). Aggiungere qui SOLO strategie validate col gauntlet."""
+    quando parte la loro storia). Aggiungere qui SOLO strategie validate col gauntlet.
+
+    SKH01 cablato @0.25 effettivo: i tre sleeve preesistenti scalati nel restante 0.75 mantenendo
+    il loro rapporto 55:25:20 (-> 41.25/18.75/15), cosi' Skyhook pesa esattamente 25% del book."""
     return [
-        tp01_sleeve(weight=0.55),      # trend difensivo, BTC/ETH, dal 2019 (l'unico deployable pieno)
-        xsec_sleeve(weight=0.25),      # cross-sectional momentum Hyperliquid, dal 2024 (scorrelato, stat-mode)
-        vrp_sleeve(weight=0.20),       # options short-vol (put credit spread + gate IV-rank), dal 2021 (lead modellato, scorrelato)
+        tp01_sleeve(weight=0.4125),    # trend difensivo, BTC/ETH, dal 2019 (l'unico deployable pieno)
+        xsec_sleeve(weight=0.1875),    # cross-sectional momentum Hyperliquid, dal 2024 (scorrelato, stat-mode)
+        vrp_sleeve(weight=0.15),       # options short-vol (put credit spread + gate IV-rank), dal 2021 (lead modellato, scorrelato)
+        skyhook_sleeve(weight=0.25),   # dual-TF regime+breakout BTC/ETH, dal 2019 (quasi-ortogonale, exit %-asimmetrici, research)
     ]

src/strategies/skyhook.py

@@ -0,0 +1,266 @@
+"""SKYHOOK (SKH01) — dual-timeframe regime+breakout system, ported to BTC/ETH (2026-06-23).
+
+NON e' un trend-follower: entra SOLO quando coincidono (a) un REGIME di volatilita'/volume e
+(b) un PATTERN di breakout/momentum. Porting onesto su BTC/ETH certificati (Deribit mainnet)
+di un sistema ES (E-mini S&P) genetico a doppio timeframe.
+
+Architettura (dal brief):
+  * data2 = HTF 690 min (genera il SEGNALE: regime + pattern)
+  * data1 = LTF 230 min (ESEGUE: ingressi/uscite)        NB 690 = 3 x 230 (HTF = 3x LTF)
+  Entrambi resampled dal feed 5m certificato con origin='epoch' -> i confini 690 sono un
+  SOTTOINSIEME dei confini 230, quindi una barra HTF chiude esattamente su una chiusura LTF.
+
+Pipeline per barra (evaluate_bar):  barre -> indicatori -> fasce regime -> pattern -> composer
+                                    -> ingresso/uscita -> SkyhookDecision
+  1. INDICATORI (sul HTF, tipo-Chande, normalizzati 0-100):
+       BuzVola  = chande01(ATR)     -> dove sei nel CICLO di volatilita'   (flat -> 50)
+       BuzVolume= chande01(volume)  -> dove sei nel CICLO di volume        (rampa -> 100)
+     Ancore della demo del brief (trend lineare): ATR costante -> BuzVola=50 (neutro);
+     volume in rampa -> BuzVolume=100. Entrambe RICOSTRUITE esattamente da chande01.
+  2. FASCE REGIME (Vola, Volume): trade ammesso solo se BuzVola in [vola_lo,vola_hi] E
+     BuzVolume in [vol_lo,vol_hi]. (Le "fasce 4/3/2 - 4/2/2" del sistema originale sono
+     ricostruite come bande-soglia tunabili: i magici interi non sono nel brief.)
+  3. PATTERN (breakout su data2/HTF): Donchian leak-free a `ptn_n` barre (default 13, da 13/13/1).
+       ptn_long  = close_htf rompe il massimo delle ptn_n barre PRECEDENTI
+       ptn_short = close_htf rompe il minimo  delle ptn_n barre PRECEDENTI
+  4. COMPOSER: contenitore_long = regime_ok AND ptn_long ; contenitore_short = regime_ok AND ptn_short
+  5. INGRESSO (max 1 al giorno): se il composer e' attivo -> OPEN_LONG / OPEN_SHORT alla
+     chiusura LTF. (stop-and-reverse: non-overlap nell'engine -> il rovescio entra alla prima
+     barra utile dopo l'uscita se il segnale persiste.)
+  6. USCITE: time-based ASIMMETRICO (uscitalong=24, uscitashort=18 barre LTF) + hard stop/profit.
+     Lo "stop 2000 / profit 5000" in $ del sistema ES e' tradotto in CRYPTO come multipli di ATR
+     LTF (scale-free): sl = k_sl*ATR, tp = k_tp*ATR (default 2.0/5.0 ~ il rapporto 40:100 pt ES),
+     con modalita' 'pct' alternativa (stop/profit in percentuale).
+
+CAUSALITA': ogni feature usa dati <= close della barra (HTF: donchian con shift(1), chande01
+rolling causale). Il merge HTF->LTF e' merge_asof BACKWARD sulla CHIUSURA HTF (<= chiusura LTF):
+una barra HTF e' usata solo quando e' realmente chiusa. backtest_signals apre a close[i].
+
+API:
+    from src.strategies.skyhook import SkyhookParams, build_frames, skyhook_entries
+    ltf, htf = build_frames(load_data("BTC","5m"))         # resample 5m -> 230m + 690m
+    entries = skyhook_entries(ltf, htf, SkyhookParams())   # list[dict|None] len(ltf), per backtest_signals
+    from src.backtest.harness import backtest_signals
+    m = backtest_signals(ltf, entries, fee_rt=0.001); m.print_summary("SKH01 BTC")
+"""
+from __future__ import annotations
+
+from dataclasses import dataclass, field
+
+import numpy as np
+import pandas as pd
+
+# 690 = 3 x 230 ; entrambi multipli esatti di 5m (138 e 46 barre da 5m)
+HTF_MIN = 690    # data2 — segnale
+LTF_MIN = 230    # data1 — esecuzione
+
+
+# ---------------------------------------------------------------------------
+# Resample dal feed 5m certificato (origin='epoch' -> confini deterministici e allineati)
+# ---------------------------------------------------------------------------
+def resample_5m(df5: pd.DataFrame, minutes: int) -> pd.DataFrame:
+    """5m -> `minutes` barre (origin epoch). Schema con 'datetime' + 'timestamp' (open-labeled)."""
+    g = df5[["timestamp", "open", "high", "low", "close", "volume"]].copy()
+    g.index = pd.to_datetime(g["timestamp"], unit="ms", utc=True)
+    out = (g.resample(f"{minutes}min", label="left", closed="left", origin="epoch")
+             .agg({"open": "first", "high": "max", "low": "min", "close": "last", "volume": "sum"})
+             .dropna(subset=["open"]))
+    out["datetime"] = out.index
+    epoch = pd.Timestamp("1970-01-01", tz="UTC")
+    out["timestamp"] = ((out.index - epoch) // pd.Timedelta(milliseconds=1)).astype("int64")
+    return out.reset_index(drop=True)[["timestamp", "open", "high", "low", "close", "volume", "datetime"]]
+
+
+def build_frames(df5: pd.DataFrame) -> tuple[pd.DataFrame, pd.DataFrame]:
+    """Da un feed 5m certificato -> (ltf 230m exec, htf 690m signal)."""
+    return resample_5m(df5, LTF_MIN), resample_5m(df5, HTF_MIN)
+
+
+# ---------------------------------------------------------------------------
+# Indicatori causali
+# ---------------------------------------------------------------------------
+def atr(df: pd.DataFrame, win: int = 14) -> np.ndarray:
+    h, l, c = df["high"].values, df["low"].values, df["close"].values
+    pc = np.roll(c, 1); pc[0] = c[0]
+    tr = np.maximum(h - l, np.maximum(np.abs(h - pc), np.abs(l - pc)))
+    return pd.Series(tr).ewm(alpha=1.0 / win, adjust=False).mean().values
+
+
+def chande01(x: np.ndarray, n: int) -> np.ndarray:
+    """Chande Momentum Oscillator su `x`, normalizzato 0-100 (tipo-Chande).
+    CMO = (Su - Sd)/(Su + Sd) in [-1,1] sulle n variazioni; mappato (1+CMO)*50 -> [0,100].
+    Serie piatta (variazioni nulle) -> 50 (neutro). Causale (rolling fino a i)."""
+    x = np.asarray(x, float)
+    d = np.diff(x, prepend=x[0])
+    up = np.where(d > 0, d, 0.0)
+    dn = np.where(d < 0, -d, 0.0)
+    su = pd.Series(up).rolling(n, min_periods=n).sum().values
+    sd = pd.Series(dn).rolling(n, min_periods=n).sum().values
+    denom = su + sd
+    cmo = np.divide(su - sd, denom, out=np.zeros_like(denom), where=denom > 0)
+    out = 50.0 * (1.0 + cmo)
+    out[~np.isfinite(out)] = 50.0
+    return out
+
+
+def donchian_breakout(df: pd.DataFrame, n: int) -> tuple[np.ndarray, np.ndarray]:
+    """Breakout leak-free: close[i] rompe il max/min delle n barre STRETTAMENTE precedenti."""
+    hi = pd.Series(df["high"].values).rolling(n, min_periods=n).max().shift(1).values
+    lo = pd.Series(df["low"].values).rolling(n, min_periods=n).min().shift(1).values
+    c = df["close"].values.astype(float)
+    return (c > hi), (c < lo)
+
+
+# ---------------------------------------------------------------------------
+# Parametri
+# ---------------------------------------------------------------------------
+@dataclass
+class SkyhookParams:
+    # indicatori (HTF)
+    atr_win: int = 14
+    n_vola: int = 13           # finestra Chande su ATR   (da PtnL 13)
+    n_volume: int = 13         # finestra Chande su volume (da PtnL 13)
+    # fasce regime (bande-soglia su 0-100). Default = "regime di breakout":
+    #   volume vivo (BuzVolume alto) + volatilita' presente ma non da blow-off.
+    vola_lo: float = 35.0
+    vola_hi: float = 95.0
+    vol_lo: float = 50.0
+    vol_hi: float = 100.0
+    # pattern (HTF) — Donchian breakout
+    ptn_n: int = 13            # da PtnL 13/13/1
+    # composer / direzione
+    long_only: bool = False    # Skyhook e' L/S di natura; True = solo long (stile crypto difensivo)
+    # ingresso
+    max_per_day: int = 1
+    # uscite — time-based asimmetrico (barre LTF)
+    uscitalong: int = 24
+    uscitashort: int = 18
+    # uscite — hard stop/profit (LONG, e SHORT se gli override sotto sono None)
+    exit_mode: str = "atr"     # 'atr' = multipli di ATR LTF ; 'pct' = percentuale fissa
+    sl_atr: float = 2.0
+    tp_atr: float = 5.0
+    sl_pct: float = 0.03
+    tp_pct: float = 0.075
+    ltf_atr_win: int = 14
+    # uscite — OVERRIDE asimmetrico SHORT (None = usa i valori simmetrici sopra).
+    # In crypto lo short si fa steamrollare da uno spike vola: stop short piu' stretti
+    # tagliano il draw-down standalone senza toccare il segnale (vedi SKH01-V2-DD, diario).
+    exit_mode_short: str | None = None
+    sl_atr_short: float | None = None
+    tp_atr_short: float | None = None
+    sl_pct_short: float | None = None
+    tp_pct_short: float | None = None
+
+
+# ---------------------------------------------------------------------------
+# Feature HTF -> merge causale su LTF
+# ---------------------------------------------------------------------------
+def htf_features(htf: pd.DataFrame, p: SkyhookParams) -> pd.DataFrame:
+    """Calcola regime+pattern sull'HTF e li restituisce indicizzati per CHIUSURA HTF (timestamp
+    di chiusura = open + 690min). Cosi' il merge backward su LTF e' strettamente causale."""
+    buz_vola = chande01(atr(htf, p.atr_win), p.n_vola)
+    buz_volume = chande01(htf["volume"].values, p.n_volume)
+    ptn_long, ptn_short = donchian_breakout(htf, p.ptn_n)
+    regime_ok = ((buz_vola >= p.vola_lo) & (buz_vola <= p.vola_hi)
+                 & (buz_volume >= p.vol_lo) & (buz_volume <= p.vol_hi))
+    comp_long = regime_ok & ptn_long
+    comp_short = regime_ok & ptn_short
+    if p.long_only:
+        comp_short = np.zeros_like(comp_short, dtype=bool)
+    close_ts = htf["timestamp"].astype("int64").values + HTF_MIN * 60 * 1000
+    return pd.DataFrame({
+        "close_ts": close_ts,
+        "buz_vola": buz_vola, "buz_volume": buz_volume,
+        "comp_long": comp_long.astype(bool), "comp_short": comp_short.astype(bool),
+    })
+
+
+def merge_htf_to_ltf(ltf: pd.DataFrame, feat: pd.DataFrame) -> pd.DataFrame:
+    """Attacca a ogni barra LTF l'ultima feature HTF la cui CHIUSURA <= chiusura LTF (causale)."""
+    left = ltf.copy()
+    left["close_ts"] = left["timestamp"].astype("int64").values + LTF_MIN * 60 * 1000
+    m = pd.merge_asof(left.sort_values("close_ts"),
+                      feat.sort_values("close_ts"),
+                      on="close_ts", direction="backward")
+    return m.sort_index().reset_index(drop=True)
+
+
+# ---------------------------------------------------------------------------
+# Generatore di ingressi per backtest_signals  ({'dir','tp','sl','max_bars'})
+# ---------------------------------------------------------------------------
+def skyhook_entries(ltf: pd.DataFrame, htf: pd.DataFrame, p: SkyhookParams | None = None) -> list:
+    """Lista di entry-dict (uno per barra LTF, None = niente segnale), pronta per
+    backtest_signals. Max `max_per_day` ingressi/giorno (prima barra qualificante del giorno).
+    sl/tp e max_bars asimmetrici per direzione. Tutto causale (decide a close[i])."""
+    p = p or SkyhookParams()
+    feat = htf_features(htf, p)
+    m = merge_htf_to_ltf(ltf, feat)
+
+    c = m["close"].values.astype(float)
+    a = atr(m, p.ltf_atr_win)
+    comp_long = np.nan_to_num(m["comp_long"].values).astype(bool)
+    comp_short = np.nan_to_num(m["comp_short"].values).astype(bool)
+    days = pd.to_datetime(m["datetime"]).dt.floor("D").values
+
+    entries: list = [None] * len(m)
+    count_today: dict = {}
+    for i in range(len(m)):
+        if not np.isfinite(a[i]) or a[i] <= 0:
+            continue
+        day = days[i]
+        if count_today.get(day, 0) >= p.max_per_day:
+            continue
+        if comp_long[i]:
+            direction, mb = 1, p.uscitalong
+            mode, sl_a, tp_a, sl_p, tp_p = p.exit_mode, p.sl_atr, p.tp_atr, p.sl_pct, p.tp_pct
+        elif comp_short[i]:
+            direction, mb = -1, p.uscitashort
+            # SHORT: usa l'override asimmetrico dove presente, altrimenti i valori simmetrici.
+            mode = p.exit_mode_short if p.exit_mode_short is not None else p.exit_mode
+            sl_a = p.sl_atr_short if p.sl_atr_short is not None else p.sl_atr
+            tp_a = p.tp_atr_short if p.tp_atr_short is not None else p.tp_atr
+            sl_p = p.sl_pct_short if p.sl_pct_short is not None else p.sl_pct
+            tp_p = p.tp_pct_short if p.tp_pct_short is not None else p.tp_pct
+        else:
+            continue
+        if mode == "atr":
+            sl_off, tp_off = sl_a * a[i], tp_a * a[i]
+        else:
+            sl_off, tp_off = sl_p * c[i], tp_p * c[i]
+        if direction == 1:
+            sl, tp = c[i] - sl_off, c[i] + tp_off
+        else:
+            sl, tp = c[i] + sl_off, c[i] - tp_off
+        entries[i] = {"dir": direction, "tp": float(tp), "sl": float(sl), "max_bars": int(mb)}
+        count_today[day] = count_today.get(day, 0) + 1
+    return entries
+
+
+# ---------------------------------------------------------------------------
+# Config canoniche (vedi docs/diary/2026-06-23-skyhook.md)
+# ---------------------------------------------------------------------------
+# SKH01-V1: vincente del primo lever-scout/grid (regime gate + breakout lento + stop larghi).
+SKH01_V1 = SkyhookParams(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35.0, vola_hi=95.0, vol_lo=0.0)
+
+# SKH01-V2-DD: vincente dell'onda DD-reduction (famiglia ASYM_LS). Stesso SEGNALE del winner
+# intermedio (ptn_n=45, banda vola larga) ma EXIT a percentuale fissa ASIMMETRICA: short con SL
+# piu' stretto (2% vs 4% long) -> taglia il draw-down standalone (maxDD BTC 21% / ETH 27% <30%)
+# alzando hold-out e uplift di portafoglio. Verificato leak-free + 2 scettici avversariali.
+SKH01_V2_DD = SkyhookParams(
+    ptn_n=45, vola_lo=35.0, vola_hi=95.0, vol_lo=0.0,
+    uscitalong=24, uscitashort=16,
+    exit_mode="pct", sl_pct=0.04, tp_pct=0.10,          # LONG
+    exit_mode_short="pct", sl_pct_short=0.02, tp_pct_short=0.08,  # SHORT (SL piu' stretto)
+)
+
+
+def signal_counts(ltf: pd.DataFrame, htf: pd.DataFrame, p: SkyhookParams | None = None) -> dict:
+    """Diagnostica: quante barre passano regime/pattern/composer (prima del cap giornaliero)."""
+    p = p or SkyhookParams()
+    feat = htf_features(htf, p)
+    m = merge_htf_to_ltf(ltf, feat)
+    cl = np.nan_to_num(m["comp_long"].values).astype(bool)
+    cs = np.nan_to_num(m["comp_short"].values).astype(bool)
+    ent = skyhook_entries(ltf, htf, p)
+    return dict(ltf_bars=len(m), comp_long=int(cl.sum()), comp_short=int(cs.sum()),
+                entries=int(sum(e is not None for e in ent)))

tests/test_skyhook.py

@@ -0,0 +1,142 @@
+"""Test della strategia SKH01 (Skyhook) — dual-timeframe regime+breakout su BTC/ETH.
+
+Coprono: fedelta' al brief (ancore demo BuzVola/BuzVolume), allineamento dual-TF, assenza di
+look-ahead (causalita'), e robustezza onesta del config V1 su entrambi gli asset.
+"""
+import sys
+from pathlib import Path
+
+import numpy as np
+import pandas as pd
+
+PROJECT_ROOT = Path(__file__).resolve().parents[1]
+sys.path.insert(0, str(PROJECT_ROOT))
+sys.path.insert(0, str(PROJECT_ROOT / "scripts" / "research" / "skyhook"))
+
+from src.data.downloader import load_data
+from src.strategies.skyhook import (
+    HTF_MIN, LTF_MIN, SKH01_V2_DD, SkyhookParams, build_frames, chande01, skyhook_entries)
+
+# config V1 (vincente del lever-scout/grid; vedi diario 2026-06-23-skyhook)
+V1 = dict(ptn_n=55, sl_atr=2.5, tp_atr=6.0, vola_lo=35, vola_hi=95, vol_lo=0.0)
+
+
+# ---------------------------------------------------------------------------
+# Fedelta' al brief: indicatori tipo-Chande, normalizzati 0-100.
+# ---------------------------------------------------------------------------
+def test_chande01_anchors():
+    """Semantica del brief: volatilita'/volume STEADY -> 50 (neutro); in RAMPA -> 100; in CALO -> 0."""
+    n = 100
+    assert abs(chande01(np.full(n, 7.0), 13)[-1] - 50.0) < 1e-9      # costante -> neutro
+    assert abs(chande01(np.arange(n, dtype=float), 13)[-1] - 100.0) < 1e-9   # rampa su -> 100
+    assert abs(chande01(np.arange(n, 0, -1, dtype=float), 13)[-1] - 0.0) < 1e-9  # rampa giu' -> 0
+
+
+def test_demo_buzvola_buzvolume():
+    """Ancore della demo: ATR costante (vol steady) -> BuzVola 50; volume in rampa -> BuzVolume 100."""
+    n = 100
+    buz_vola = chande01(np.full(n, 2.0), 13)            # ATR steady
+    buz_volume = chande01(np.linspace(1000, 5000, n), 13)  # volume in rampa
+    assert abs(buz_vola[-1] - 50.0) < 1e-9
+    assert abs(buz_volume[-1] - 100.0) < 1e-9
+    # oscillatori sempre in [0,100]
+    assert chande01(np.random.default_rng(0).normal(size=500).cumsum() + 100, 13)[20:].min() >= -1e-9
+    assert chande01(np.random.default_rng(1).normal(size=500).cumsum() + 100, 13)[20:].max() <= 100 + 1e-9
+
+
+# ---------------------------------------------------------------------------
+# Allineamento dual-timeframe: 690 = 3 x 230, confini HTF subset dei confini LTF.
+# ---------------------------------------------------------------------------
+def test_dual_tf_alignment():
+    assert HTF_MIN == 3 * LTF_MIN
+    ltf, htf = build_frames(load_data("BTC", "5m"))
+    # ogni timestamp (open) HTF e' anche un open LTF (stessa griglia epoch)
+    ltf_opens = set(ltf["timestamp"].astype("int64").tolist())
+    htf_opens = htf["timestamp"].astype("int64").tolist()
+    inside = sum(t in ltf_opens for t in htf_opens)
+    assert inside / len(htf_opens) > 0.99, "i confini HTF devono essere un sottoinsieme dei confini LTF"
+
+
+# ---------------------------------------------------------------------------
+# Causalita': gli ingressi su un prefisso devono coincidere con la run completa.
+# ---------------------------------------------------------------------------
+def test_no_lookahead_entries():
+    p = SkyhookParams(**V1)
+    ltf, htf = build_frames(load_data("BTC", "5m"))
+    full = skyhook_entries(ltf, htf, p)
+    n = len(ltf)
+    cut = int(n * 0.85)
+    cut_ts = int(ltf["timestamp"].iloc[cut - 1])
+    htf_cut = htf[htf["timestamp"] <= cut_ts].reset_index(drop=True)
+    sub = skyhook_entries(ltf.iloc[:cut].reset_index(drop=True), htf_cut, p)
+    for i in range(cut - 200, cut):
+        a, b = full[i], sub[i]
+        assert (a is None) == (b is None)
+        if a is not None:
+            assert a["dir"] == b["dir"]
+            assert abs(a["sl"] - b["sl"]) < 1e-6 and abs(a["tp"] - b["tp"]) < 1e-6
+
+
+# ---------------------------------------------------------------------------
+# Robustezza onesta del config V1: PASS su BTC E ETH, netto fee, OOS.
+# ---------------------------------------------------------------------------
+def test_v1_robust_both_assets():
+    import skyhooklib as sk
+    p = SkyhookParams(**V1)
+    for a in ("BTC", "ETH"):
+        r = sk.run_asset(a, p, sk.FEE_RT)
+        assert r["full"]["sharpe"] >= 0.5, f"{a} FULL Sharpe basso: {r['full']['sharpe']}"
+        assert r["holdout"]["sharpe"] >= 0.2, f"{a} HOLD-OUT Sharpe basso: {r['holdout']['sharpe']}"
+        assert r["full"]["n_trades"] >= 20, f"{a} troppo pochi trade: {r['full']['n_trades']}"
+    assert sk.causality(p, "BTC")["ok"] and sk.causality(p, "ETH")["ok"]
+
+
+# ---------------------------------------------------------------------------
+# Exit asimmetrici SHORT (SKH01-V2-DD): l'override cambia SOLO gli short; i default
+# (None) preservano esattamente il comportamento simmetrico precedente.
+# ---------------------------------------------------------------------------
+def test_short_override_backward_compatible():
+    """Con gli override SHORT a None, gli ingressi sono identici alla versione simmetrica."""
+    ltf, htf = build_frames(load_data("BTC", "5m"))
+    base = SkyhookParams(**V1)
+    # stessi parametri ma con campi override esplicitamente None (= default)
+    same = SkyhookParams(**V1, exit_mode_short=None, sl_pct_short=None, tp_pct_short=None)
+    e0, e1 = skyhook_entries(ltf, htf, base), skyhook_entries(ltf, htf, same)
+    assert e0 == e1, "i campi override a None NON devono cambiare nulla (backward-compat)"
+
+
+def test_short_override_changes_only_shorts():
+    """Un SL short piu' stretto (pct) modifica gli stop SHORT ma lascia intatti i LONG."""
+    ltf, htf = build_frames(load_data("ETH", "5m"))
+    sym = SkyhookParams(ptn_n=45, vola_lo=35, vola_hi=95, vol_lo=0.0,
+                        exit_mode="pct", sl_pct=0.04, tp_pct=0.10)
+    asym = SkyhookParams(ptn_n=45, vola_lo=35, vola_hi=95, vol_lo=0.0,
+                         exit_mode="pct", sl_pct=0.04, tp_pct=0.10,
+                         sl_pct_short=0.02, tp_pct_short=0.08)
+    es, ea = skyhook_entries(ltf, htf, sym), skyhook_entries(ltf, htf, asym)
+    longs_same = shorts_diff = 0
+    for a, b in zip(es, ea):
+        if a is None or b is None:
+            assert (a is None) == (b is None)
+            continue
+        assert a["dir"] == b["dir"]
+        if a["dir"] == 1:                                  # LONG invariati
+            assert abs(a["sl"] - b["sl"]) < 1e-6 and abs(a["tp"] - b["tp"]) < 1e-6
+            longs_same += 1
+        else:                                              # SHORT con SL/TP diversi
+            assert abs(a["sl"] - b["sl"]) > 1e-6
+            shorts_diff += 1
+    assert longs_same > 0 and shorts_diff > 0
+
+
+def test_v2dd_robust_both_assets():
+    """SKH01-V2-DD: PASS netto fee su BTC&ETH, hold-out forte, e maxDD standalone <30%."""
+    import skyhooklib as sk
+    p = SKH01_V2_DD
+    for a in ("BTC", "ETH"):
+        r = sk.run_asset(a, p, sk.FEE_RT)
+        assert r["full"]["sharpe"] >= 0.5, f"{a} FULL Sharpe basso: {r['full']['sharpe']}"
+        assert r["holdout"]["sharpe"] >= 0.5, f"{a} HOLD-OUT Sharpe basso: {r['holdout']['sharpe']}"
+        assert r["full"]["maxdd"] < 0.30, f"{a} maxDD non sotto 30%: {r['full']['maxdd']}"
+        assert r["full"]["n_trades"] >= 20, f"{a} troppo pochi trade: {r['full']['n_trades']}"
+    assert sk.causality(p, "BTC")["ok"] and sk.causality(p, "ETH")["ok"]