The thesis
bollinger-reversion runs a classic mean-reversion play across a 24-name universe of US large caps: buy when price dips below the lower Bollinger band, sell when it pushes above the upper band. The bet is that stretched prices snap back toward their moving average. It's a well-trodden idea, and on the surface the numbers flatter it.
Backtest performance
Over 451 days the strategy returned 17.55% (9.46% CAGR), finishing at $11,755 on a $10,000 start. The 63.89% win rate across 76 trades is the headline strength — the entries do tend to catch bounces. But two figures temper the enthusiasm. The Sharpe of 0.66 is modest, and the 20.57% max drawdown means the ride was rough relative to the reward. Turnover of 1,720% also flags a high-churn approach; the strategy is trading a lot to earn that return, and every round trip pays a fee.
Where it breaks: validation
This is the part that matters, and it's where bollinger-reversion fails our validation gate. Walk-forward analysis ran four folds — all four were positive, which sounds reassuring — but the trend is the story. Fold returns decay monotonically: 7.67% → 1.42% → 0.77% → 0.40%, with Sharpe falling from 1.78 to 0.14. The out-of-sample return is effectively flat at 0.4% with an OOS Sharpe of 0.14.
The risk-adjusted probability metrics say the same thing. The Probabilistic Sharpe Ratio sits at a healthy 0.81, but the Deflated Sharpe Ratio — which penalises for the six trials run — drops to 0.34. Once you account for how many variations were tested, the edge no longer clears the bar. The strong backtest looks increasingly like a fold-1 artifact that the market has since arbitraged away.
Recent activity: idling
Live behaviour reinforces the caution. The last executed trades were on 2026-08-07 (a 1-share CAT buy and a 25-share DIS sell). Every scheduled run since 2026-08-20 reports the same pattern: 0 executed, 1–2 rejected. The book is holding roughly $2,752 in cash against a total equity that has drifted from ~$9,815 down to ~$9,646 over the past week. In other words, the signal isn't firing, and the existing positions are bleeding slightly.
The verdict
bollinger-reversion is a textbook case of why a good backtest is a starting point, not a conclusion. The high win rate and positive folds are genuine strengths, but the out-of-sample decay, the deflated Sharpe, and the current run of rejected orders all point the same direction: the edge is thin and fading. It stays live for observation, but on the evidence it is not a candidate for scaled capital until it demonstrates fresh out-of-sample lift.