chore(reset): v2.0.0 — storico certificato Deribit mainnet, ripartenza pulita

Reset del progetto su fondamenta verificate dopo la scoperta che l'intera libreria "validata OOS" era artefatto di feed contaminato (print fantasma del feed Cerbero TESTNET + storico Binance/USDT). - Storico ricostruito da Deribit MAINNET (ccxt pubblico, tokenless) e CERTIFICATO (certify_feed.py): BTC/ETH puliti su TUTTA la storia (mediana 2-6 bps vs Coinbase USD), integrita' OHLC + coerenza resample (maxΔ 0.00) + cross-venue OK. Alt esclusi (illiquidi/divergenti: LTC/DOGE 50-82% barre flat; XRP/BNB non certificabili). - Verdetto sul feed pulito: FADE / PAIRS / XS01 / TSM01 morti (ogni portafoglio Sharpe -2.3..-3.0, DD ~40%); solo SH01 e frammenti HONEST con segnale residuo, da ri-validare in isolamento. - Cleanup "restart pulito": strategie, stack live (src/live, src/portfolio, runner/executor, yml, docker), ~100 script ricerca/gate, waste/games/ portfolios, dati non certificati + cache e 60+ diari -> archiviati in Old/ (preservati, non cancellati). Diario consolidato in un unico documento. - Skeleton ricerca tenuto: Strategy ABC + indicatori + src/fractal + src/backtest/engine + load_data; tool dati certificati (rebuild_history, certify_feed, audit_feed, multi_source_check). - Universo dati ATTIVO: solo BTC/ETH (5m/15m/1h); guardrail fisico (load_data su alt -> FileNotFoundError). Esecuzione DISABILITATA, conto flat. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-19 15:16:03 +00:00
parent 8401a280b9
commit 14522262e6
383 changed files with 1971 additions and 779 deletions
@@ -0,0 +1,86 @@
+import sys; sys.path.insert(0,".")
+import numpy as np, pandas as pd
+from scripts.analysis.regime_lab import load_features
+from scripts.analysis.explore_lab import atr
+
+FEE=0.001; LEV=3
+
+def build(df, gate, k=2.5, sl_atr=2.0, mb=24, bb=50):
+    c=df['close'].values; a=atr(df,14)
+    ma=pd.Series(c).rolling(bb).mean().values; sd=pd.Series(c).rolling(bb).std().values
+    ent=[]
+    for i in range(bb+14,len(c)-1):
+        if np.isnan(sd[i]) or sd[i]==0 or np.isnan(a[i]): continue
+        if not gate(df,i): continue
+        if c[i]<ma[i]-k*sd[i]: d,sl=1,c[i]-sl_atr*a[i]
+        elif c[i]>ma[i]+k*sd[i]: d,sl=-1,c[i]+sl_atr*a[i]
+        else: continue
+        ent.append({'i':i,'d':d,'tp':ma[i],'sl':sl,'mb':mb})
+    return ent
+
+def per_year(df, ent):
+    """replay intrabar fedele (sl-first, tp, poi max_bars@close) -> per anno {n,ret%,win%}."""
+    h=df['high'].values; l=df['low'].values; c=df['close'].values
+    ts=pd.to_datetime(df['timestamp'],unit='ms',utc=True)
+    Y={}
+    last=-1
+    for e in ent:
+        i=e['i']
+        if i<=last: continue
+        d=e['d']; tp=e['tp']; sl=e['sl']; j=min(i+e['mb'],len(c)-1)
+        exit_p=c[j]
+        for t in range(i+1,j+1):
+            if d==1:
+                if l[t]<=sl: exit_p=sl; j=t; break
+                if h[t]>=tp: exit_p=tp; j=t; break
+            else:
+                if h[t]>=sl: exit_p=sl; j=t; break
+                if l[t]<=tp: exit_p=tp; j=t; break
+        ret=(exit_p-c[i])/c[i]*d*LEV - FEE*LEV
+        last=j; yr=ts.iloc[i].year
+        if yr not in Y: Y[yr]=[0,0.0,0]
+        Y[yr][0]+=1; Y[yr][1]+=ret*100; Y[yr][2]+= (ret>0)
+    return Y
+
+# gate functions
+def g_hurst_calm(df,i): return df['hurst'].iloc[i]<0.55 and not np.isnan(df['dvol_pct'].iloc[i]) and df['dvol_pct'].iloc[i]<0.40
+def g_vrp_neg(df,i):    return not np.isnan(df['vrp'].iloc[i]) and df['vrp'].iloc[i]<0
+def g_hig_vrp(df,i):
+    hi=df['higuchi'].iloc[i]; return (not np.isnan(hi)) and hi>1.5 and (not np.isnan(df['vrp'].iloc[i])) and df['vrp'].iloc[i]<0
+def g_none(df,i): return True
+
+STRATS=[("HurstCalmFade (hurst<.55 & DVOL<p40)",g_hurst_calm),
+        ("VRP<0 Fade (core driver)",g_vrp_neg),
+        ("HigVRP Fade (Higuchi>1.5 & VRP<0)",g_hig_vrp),
+        ("Fade NUDA (no gate, baseline)",g_none)]
+
+# regime mercato BTC per anno (da BTC close annuale)
+btc=load_features("BTC","1d")
+bts=pd.to_datetime(btc['timestamp'],unit='ms',utc=True); bc=btc['close'].values
+mkt={}
+for yr in range(2021,2027):
+    m=(bts.dt.year==yr).values
+    if m.sum()>5:
+        r=(bc[m][-1]/bc[m][0]-1)*100
+        mkt[yr]=("BULL" if r>40 else "BEAR" if r<-30 else "RANGE", r)
+
+for asset in ("BTC","ETH"):
+    df=load_features(asset,"1h")
+    print(f"\n{'='*78}\n  {asset} 1h — performance per anno (somma ret% per-trade, netto leva3x+fee0.10%)\n{'='*78}")
+    print(f"  {'Strategia':<40} " + " ".join(f"{y}" for y in range(2021,2027)))
+    for name,g in STRATS:
+        ent=build(df,g); Y=per_year(df,ent)
+        cells=[]
+        for y in range(2021,2027):
+            if y in Y and Y[y][0]>0:
+                cells.append(f"{Y[y][1]:+5.0f}")
+            else: cells.append("   . ")
+        print(f"  {name:<40} " + " ".join(cells))
+    # riga trades/anno per la strategia principale
+    ent=build(df,g_hurst_calm); Y=per_year(df,ent)
+    tr=" ".join(f"{Y.get(y,[0])[0]:>5}" for y in range(2021,2027))
+    print(f"  {'(HurstCalmFade trades/anno)':<40} {tr}")
+
+print(f"\n  REGIME MERCATO BTC per anno (ret% annuale prezzo):")
+for y in range(2021,2027):
+    if y in mkt: print(f"    {y}: {mkt[y][0]:6} ({mkt[y][1]:+.0f}%)")