The thesis
Bollinger-reversion is a textbook mean-reversion play: buy a name when it trades below its lower Bollinger band, sell when it pushes above the upper band. It runs on a universe of 24 US large caps — the usual megacap tech (AAPL, MSFT, GOOGL, NVDA), financials (JPM, BAC, V, MA), staples (PG, KO, WMT, COST), and a handful of industrials and healthcare names. The strategy is currently live.
What the backtest says
The headline numbers look inviting. Over 451 days the strategy returned 17.55% (CAGR 9.46%), turning $10,000 into $11,755. Win rate was a healthy 63.89% across 76 trades, and fees were negligible ($76 total, no FX cost). But two figures temper the enthusiasm: a Sharpe of just 0.66 and a max drawdown of 20.57%. That is a lot of pain for the return on offer. Turnover of 1,720% also tells you this is an active, churn-heavy approach — respectable in a zero-commission paper environment, but a real drag once frictions bite.
Why validation fails
This is where the story turns. The strategy did not pass validation, and the walk-forward folds show why. All four folds were positive, but performance decays monotonically:
| Fold | Return | Sharpe |
|---|---|---|
| 1 | 7.67% | 1.78 |
| 2 | 1.42% | 0.25 |
| 3 | 0.77% | 0.20 |
| 4 | 0.40% | 0.14 |
The entire 17.55% is essentially front-loaded into the earliest window. By the most recent out-of-sample fold, return is 0.4% and Sharpe 0.14 — indistinguishable from noise. The deflated Sharpe ratio (DSR) of 0.342, computed across 6 trials, confirms the edge does not survive multiple-testing correction, even though the raw PSR of 0.814 looks superficially reassuring. In plain terms: the pattern that worked in 2024 has largely stopped working.
Live behaviour
The live record reinforces the caution. The last executed trades were on 2026-08-07 (a CAT buy, a DIS sell). Since then, every scheduled run this September has logged 0 executed, 2 rejected — six sessions running, with cash frozen at $2,752.28. Portfolio value has drifted in the $9,570–$9,685 range, meaning the live paper book is sitting modestly below its $10,000 starting point while the backtest boasts double-digit gains. The persistent rejections suggest the strategy is finding signals it cannot act on — worth a look at the position-sizing or cash constraints behind those blocked orders.
Verdict
Bollinger-reversion is a clean, interpretable idea with a genuinely good win rate, and its early-fold Sharpe of 1.78 shows the setup can work in the right regime. But the balance of evidence is a warning: a decaying out-of-sample curve, a failed validation gate, a middling full-sample Sharpe against a 20% drawdown, and a live book that is currently stuck and slightly underwater. This is not a strategy to scale capital into today. The honest next step is diagnosing the recent trade rejections and asking whether mean-reversion still has an edge in the current market — the folds suggest it may not.