Spaces:
Running
Running
File size: 13,293 Bytes
8e50444 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 | import sys, os
import numpy as np, pandas as pd
from scipy.stats import spearmanr
import warnings; warnings.filterwarnings('ignore')
try:
from v30_causal_engine import evaluate_slice, CAP
except ImportError:
from backtesting.v30_causal_engine import evaluate_slice, CAP
class QuantValidator:
"""
Generalized Quantitative Validation Framework.
Allows institutional-grade auditing of any causal trading strategy.
"""
def __init__(self, dc, spy, vf, daily_ret, strategy_fn, default_params,
signal_fn=None, audit_fn=None, txn_param_name='txn_bps',
rebal_param_name='rebal_days'):
self.dc = dc
self.spy = spy
self.vf = vf
self.daily_ret = daily_ret
self.strategy_fn = strategy_fn
self.default_params = default_params
self.signal_fn = signal_fn
self.audit_fn = audit_fn
self.txn_param_name = txn_param_name
self.rebal_param_name = rebal_param_name
def _run(self, dc_sub, spy_sub, vf_sub, ret_sub, params):
c = self.strategy_fn(dc_sub, spy_sub, vf_sub, ret_sub, **params)
if isinstance(c, dict) and 'curve' in c:
return c['curve']
return c
def _get_ew_baseline(self):
# Quick equal-weight causal baseline
nav = CAP; port_rets = []; hist = []
sma = self.spy.rolling(200).mean()
for i in range(1, len(self.dc)):
if len(port_rets) >= 21:
w = port_rets[-60:] if len(port_rets) >= 60 else port_rets[-21:]
vs = 0.18 / (np.std(w)*np.sqrt(252)+1e-8)
else: vs = 0.5
sp, sm = self.spy.values[i-1], sma.values[i-1]
if pd.isna(sm) or sp <= sm: vs *= 0.50
vs = float(np.clip(vs, 0.05, 1.0))
lr = self.daily_ret.iloc[i][[t for t in self.vf if t in self.daily_ret.columns]].dropna()
day_ret = lr.mean() * vs if len(lr) > 0 else 0.0
nav *= (1 + day_ret)
port_rets.append(day_ret)
hist.append(nav)
return pd.Series(hist, index=self.dc.index[1:len(hist)+1])
def run_all_phases(self, param_grid=None):
self.run_phase1()
self.run_phase2()
self.run_phase3()
if param_grid:
self.run_phase4(param_grid)
self.run_phase5()
def run_phase1(self):
print("\n========================================")
print(" PHASE 1: SIGNAL VALIDATION")
print("========================================")
if self.signal_fn:
print("--- Test 1.1 & 1.2: IC Analysis ---")
raw_signal = self.signal_fn(self.dc, self.spy, self.vf)
fwd = self.dc[self.vf].pct_change(60).shift(-60)
sma = self.spy.rolling(200).mean()
ic_vals, dates, regimes = [], [], []
# Calculate IC every 60 days
for i in range(200, len(self.dc)-60, 60):
sig = raw_signal.iloc[i].dropna()
common = sig.index.intersection(fwd.iloc[i].dropna().index)
if len(common) >= 15:
corr, _ = spearmanr(sig[common].values, fwd.iloc[i][common].values)
if not np.isnan(corr):
ic_vals.append(corr)
dates.append(self.dc.index[i])
regimes.append("ON" if self.spy.iloc[i] > sma.iloc[i] else "OFF")
ic_series = pd.Series(ic_vals, index=dates)
ic_mean = ic_series.mean()
t_stat = ic_mean / (ic_series.std() / np.sqrt(len(ic_series))) if len(ic_series)>0 else 0
print(f"Mean IC: {ic_mean:.4f} | t-stat: {t_stat:.2f}")
print(f"Result: {'PASS' if t_stat > 2.0 else 'FAIL'} (Threshold: t > 2.0)")
print("\nBy Regime:")
for reg in ["ON", "OFF"]:
mask = [r == reg for r in regimes]
sub = ic_series[mask]
if len(sub) > 0:
rt_stat = sub.mean() / (sub.std() / np.sqrt(len(sub)))
print(f"Risk-{reg} days: t-stat = {rt_stat:.2f} (n={len(sub)})")
else:
print("--- Test 1.1 & 1.2: Skipped (No signal_fn provided) ---")
print("\n--- Test 1.3: Baseline Comparison (Equal Weight) ---")
c_orig = self._run(self.dc, self.spy, self.vf, self.daily_ret, self.default_params)
m_orig = evaluate_slice(c_orig, "2008-01-01", "2025-12-31")
c_ew = self._get_ew_baseline()
m_ew = evaluate_slice(c_ew, "2008-01-01", "2025-12-31")
print(f"Strategy Sharpe: {m_orig['sharpe']:.4f}")
print(f"Eq-Weight Sharpe: {m_ew['sharpe']:.4f}")
diff = m_orig['sharpe'] - m_ew['sharpe']
print(f"Excess Sharpe: {diff:+.4f}")
print(f"Result: {'PASS' if diff > 0.05 else 'FAIL'} (Threshold: > +0.05)")
def run_phase2(self):
print("\n========================================")
print(" PHASE 2: BACKTEST INTEGRITY")
print("========================================")
print("--- Test 2.1: Strict Train/Test Split ---")
c = self._run(self.dc, self.spy, self.vf, self.daily_ret, self.default_params)
m_train = evaluate_slice(c, "2008-01-01", "2018-12-31")
m_test = evaluate_slice(c, "2019-01-01", "2025-12-31")
print(f"Train Sharpe (2008-2018): {m_train['sharpe']:.4f}")
print(f"Test Sharpe (2019-2025): {m_test['sharpe']:.4f}")
diff = m_test['sharpe'] - m_train['sharpe']
print(f"Difference: {diff:+.4f}")
print(f"Result: {'PASS' if diff > -0.20 else 'FAIL'} (Test decay must not exceed -0.20)")
print("\n--- Test 2.2: Start Date Sensitivity ---")
sharpes = []
rebal = self.default_params.get(self.rebal_param_name, 60)
step = 4 if rebal <= 60 else 6
offsets = list(range(0, rebal, step))
print(f"Running {len(offsets)} offsets...", end="", flush=True)
for off in offsets:
c_off = self._run(self.dc.iloc[off:], self.spy.iloc[off:], self.vf, self.daily_ret.iloc[off:], self.default_params)
m = evaluate_slice(c_off, "2008-01-01", "2025-12-31")
sharpes.append(m['sharpe'])
print(".", end="", flush=True)
print()
s_mean, s_range = np.mean(sharpes), max(sharpes) - min(sharpes)
print(f"Mean Sharpe: {s_mean:.4f} | Range: {s_range:.4f}")
print(f"Result: {'PASS' if s_range < 0.20 else 'FAIL'} (Path dependency range < 0.20)")
print("\n--- Test 2.3: Survivorship Bias (Poison Universe) ---")
c_base = self._run(self.dc, self.spy, self.vf, self.daily_ret, self.default_params)
base_cagr = evaluate_slice(c_base, "2008-01-01", "2025-12-31")['cagr']
POISON_TICKERS = ["PTON", "CLOV", "NKLA", "QS", "SPCE", "SKLZ", "GOEV", "FSR", "SOFI", "BYND", "ZM", "DOCU", "TDOC", "UPST", "AFRM", "CHPT"]
import yfinance as yf
poison_raw = yf.download(POISON_TICKERS, start="2006-01-01", end="2025-12-31", progress=False)
if isinstance(poison_raw.columns, pd.MultiIndex):
lvl0 = poison_raw.columns.get_level_values(0).unique().tolist()
poison_close = poison_raw["Close"] if "Close" in lvl0 else poison_raw
if isinstance(poison_close.columns, pd.MultiIndex): poison_close.columns = poison_close.columns.get_level_values(-1)
else: poison_close = poison_raw
valid_poison = [t for t in POISON_TICKERS if t in poison_close.columns and poison_close[t].notna().sum() > 100]
dc_poisoned = self.dc.copy()
for t in valid_poison:
if t not in dc_poisoned.columns:
dc_poisoned[t] = poison_close[t].reindex(dc_poisoned.index).ffill()
poison_vf = list(dict.fromkeys(list(self.vf) + valid_poison))
c_poison = self._run(dc_poisoned, self.spy, poison_vf, dc_poisoned.pct_change(), self.default_params)
poison_cagr = evaluate_slice(c_poison, "2008-01-01", "2025-12-31")['cagr']
print(f"Baseline CAGR: {base_cagr:.1f}% | Poison CAGR: {poison_cagr:.1f}%")
diff = base_cagr - poison_cagr
print(f"Result: {'PASS' if diff <= 3.0 else 'FAIL'} (Decay <= 3.0%)")
def run_phase3(self):
print("\n========================================")
print(" PHASE 3: ROBUSTNESS TESTING")
print("========================================")
print("--- Test 3.1: Transaction Cost Stress ---")
if self.txn_param_name not in self.default_params:
print(f"Error: Txn param '{self.txn_param_name}' not in default params. Cannot stress test.")
else:
for bps in [20, 40, 60]:
p = self.default_params.copy()
p[self.txn_param_name] = bps
c = self._run(self.dc, self.spy, self.vf, self.daily_ret, p)
m = evaluate_slice(c, "2008-01-01", "2025-12-31")
print(f"At {bps} bps: Sharpe = {m['sharpe']:.4f}")
if bps == 60:
print(f"Result: {'PASS' if m['sharpe'] >= 0.60 else 'FAIL'} (Threshold > 0.60 at 60bps)")
print("\n--- Test 3.5: Momentum Crash Autopsy ---")
c = self._run(self.dc, self.spy, self.vf, self.daily_ret, self.default_params)
crashes = {
"GFC (Lehman)": ("2008-09-01", "2009-03-31"),
"China Shock": ("2015-06-01", "2015-09-30"),
"Rate Shock": ("2022-01-01", "2022-12-31")
}
for name, (start, end) in crashes.items():
if start >= c.index[0].strftime('%Y-%m-%d') and end <= c.index[-1].strftime('%Y-%m-%d'):
c_slice = c.loc[start:end]
s_slice = self.spy.loc[start:end]
if len(c_slice) == 0: continue
c_ret = (c_slice.iloc[-1] / c_slice.iloc[0]) - 1
s_ret = (s_slice.iloc[-1] / s_slice.iloc[0]) - 1
print(f"{name:<15}: Strat {c_ret*100:>+6.1f}% | SPY {s_ret*100:>+6.1f}%")
print("Result: PASS (Quantified)")
def run_phase4(self, param_grid):
print("\n========================================")
print(" PHASE 4: PARAMETER ROBUSTNESS")
print("========================================")
print("Sweeping parameters...")
import itertools
keys = list(param_grid.keys())
combos = list(itertools.product(*(param_grid[k] for k in keys)))
res = {}
for combo in combos:
p = self.default_params.copy()
for k, v in zip(keys, combo):
p[k] = v
c = self._run(self.dc, self.spy, self.vf, self.daily_ret, p)
m = evaluate_slice(c, "2008-01-01", "2025-12-31")
res[combo] = m['sharpe']
for combo, sharpe in res.items():
print(f"Params {dict(zip(keys, combo))}: Sharpe {sharpe:.4f}")
best = max(res, key=res.get)
print(f"\nBest Params {dict(zip(keys, best))} -> Evaluating Overfit Risk...")
p = self.default_params.copy()
for k, v in zip(keys, best): p[k] = v
c = self._run(self.dc, self.spy, self.vf, self.daily_ret, p)
m_train = evaluate_slice(c, "2008-01-01", "2018-12-31")
m_test = evaluate_slice(c, "2019-01-01", "2025-12-31")
print(f"Train (2008-2018): {m_train['sharpe']:.4f}")
print(f"Test (2019-2025): {m_test['sharpe']:.4f}")
print(f"Result: {'PASS' if m_test['sharpe'] >= m_train['sharpe'] - 0.15 else 'FAIL'} (Test didn't collapse)")
def run_phase5(self):
print("\n========================================")
print(" PHASE 5: FORWARD VALIDITY")
print("========================================")
c = self._run(self.dc, self.spy, self.vf, self.daily_ret, self.default_params)
regimes = {
"2020 V-Crash": ("2020-02-01", "2020-12-31"),
"2022 Slow Bear": ("2022-01-01", "2022-12-31")
}
fails = 0
for name, (start, end) in regimes.items():
if start >= c.index[0].strftime('%Y-%m-%d') and end <= c.index[-1].strftime('%Y-%m-%d'):
c_slice, s_slice = c.loc[start:end], self.spy.loc[start:end]
c_ret = (c_slice.iloc[-1] / c_slice.iloc[0]) - 1
s_ret = (s_slice.iloc[-1] / s_slice.iloc[0]) - 1
print(f"{name}: Strat {c_ret*100:+.1f}% | SPY {s_ret*100:+.1f}%")
if name == "2022 Slow Bear" and c_ret < s_ret:
print(" -> FAIL: Lost more than SPY")
fails += 1
elif name == "2020 V-Crash" and c_ret < s_ret * 0.30:
print(" -> FAIL: Captured < 30% of SPY recovery")
fails += 1
print(f"\nResult: {'PASS' if fails == 0 else 'FAIL'} (Regime tests)")
|