The thesis
Bollinger-reversion is a textbook mean-reversion play: buy when price closes below the lower Bollinger band, sell when it pushes above the upper band. The idea is that price extremes snap back toward a moving-average center. It runs on a diversified 24-name universe of US large caps spanning tech, financials, healthcare, staples, and energy — AAPL, MSFT, NVDA, JPM, V, JNJ, PG, WMT, XOM, and CAT among them. The status is marked live.
Backtest performance
Over 451 trading days the strategy returned 17.55%, ending at $11,755 on a $10k base — a 9.46% CAGR. The win rate is genuinely good at 63.89% across 76 trades, and the deflated- and probabilistic-Sharpe headline (PSR 0.81) initially looks encouraging. But the risk profile is less flattering: a Sharpe of just 0.66 and a 20.57% max drawdown mean the returns came with real volatility. Turnover of 1,721% is high for a 24-name book, signaling heavy churn — and $76 in fees confirms the trading is not free.
Validation: the strategy did not pass
This is where the case weakens. Walk-forward validation across four folds failed. All four folds were positive, which sounds reassuring, but the magnitudes tell a story of decay:
| Fold | Return | Sharpe |
|---|---|---|
| 1 (2024-08→2025-01) | 7.67% | 1.78 |
| 2 (2025-01→2025-07) | 1.42% | 0.25 |
| 3 (2025-07→2025-12) | 0.77% | 0.20 |
| 4 (2025-12→2026-05) | 0.40% | 0.14 |
The edge is almost entirely concentrated in the earliest fold and fades monotonically thereafter. The out-of-sample return of 0.4% and OOS Sharpe of 0.14 are effectively flat. With a Deflated Sharpe of 0.34 against 6 trials, the honest read is that much of the backtest gain may be in-sample fitting rather than a durable signal.
Live activity: stalled
Recent paper trading reinforces the caution. The last executed trades were in early-to-mid June — buys in DIS, GOOGL, PG, WMT, and COST, plus a PG trim. Since then, the six most recent scheduled runs (July 21–28) each report 0 executed, 1 rejected, with cash pinned at $1,001.98. Total equity has drifted between roughly $9,354 and $9,818 — below the $10k mark — so the live book is currently underwater relative to its start, in contrast to the rosy backtest.
Verdict
Bollinger-reversion has a clean, interpretable thesis and a strong nominal win rate, and it stayed positive across every validation window. But the deteriorating fold returns, sub-1 Sharpe, 20%+ drawdown, and a validation gate it failed to clear all point to a thin, possibly overfit edge. The repeated trade rejections and sub-par live equity suggest the strategy is struggling to find qualifying setups in the current regime. It earns continued observation on paper — not fresh capital.