Back
Regime-Conditional Strategy Analyzer
Split backtest results by market regime and compute conditional win rates, Sharpe ratios, and P(win | regime) probability tables.
Feb 23, 2026, Eric
The Core Idea
A strategy with a 55% overall win rate might be 65% in low-volatility regimes and 40% in high-volatility regimes. Treating it as a uniform 55% ignores the regime dependency — which is the most actionable piece of information a backtest can give you. This tool computes conditional probabilities: P(win | regime A) vs P(win | regime B).
Regime Indicator
CSV columns: return (required), vix, price/close (optional)
Conditional Probability Table
| Metric | VIX ≤ 25 (Low vol) | VIX > 25 (High vol) |
|---|---|---|
| N (trades) | 376 | 124 |
| P(win | regime) | 62.5% | 47.6% |
| Avg return / trade | 0.285% | -0.127% |
| Annualized Sharpe | 4.36 | -1.12 |
| Win rate edge over 50% | 12.5pp | -2.4pp |
Metric Comparison by Regime
Win Rate
VIX ≤ 25 (Low vol)
62.5%
VIX > 25 (High vol)
47.6%
Avg Return / Trade
VIX ≤ 25 (Low vol)
0.285%
VIX > 25 (High vol)
-0.127%
Annualized Sharpe
VIX ≤ 25 (Low vol)
4.36
VIX > 25 (High vol)
-1.12
What This Means
Significant regime dependency detected. Win rate difference of 14.9pp between regimes. This strategy behaves materially differently depending on market conditions — consider running it only in the favorable regime.
Sharpe split: VIX ≤ 25 (Low vol): 4.36 vs VIX > 25 (High vol): -1.12. The risk-adjusted return gap between regimes is substantial — position sizing should reflect this.