The thesis
Bollinger-reversion is a textbook mean-reversion play: buy a name when it dips below its lower Bollinger band, sell when it pushes above the upper band. It runs on a 24-stock universe of large-cap blue chips — AAPL, MSFT, NVDA, JPM, JNJ, WMT, XOM and the like — the kind of liquid, well-behaved names where band-touch reversals are supposed to be reliable. The strategy is currently live.
Headline performance
On paper, the numbers flatter the approach. The full backtest over 451 days returned 17.55% (9.46% CAGR), finishing at $11,755 on a $10k base, with a 63.89% win rate across 76 trades. That hit rate is the strategy's strongest selling point — reversion systems live or die on winning more often than they lose, and nearly two-in-three closed trades landed green.
The warning signs sit alongside those wins. The Sharpe ratio is a modest 0.66, and max drawdown reached 20.57% — a fifth of capital — which is steep for a system leaning on "safe" large caps. Turnover of 1,720% also signals a lot of churn relative to the edge captured; fees ($76) are small here but scale with activity.
Validation: the part that matters
Walk-forward validation failed, and the fold-by-fold breakdown shows why. All four out-of-sample folds were positive — which sounds reassuring — but returns decayed monotonically:
- Fold 1 (Aug 2024–Jan 2025): +7.67%, Sharpe 1.78
- Fold 2: +1.42%, Sharpe 0.25
- Fold 3: +0.77%, Sharpe 0.20
- Fold 4 (Dec 2025–May 2026): +0.40%, Sharpe 0.14
The most recent out-of-sample window is barely above zero. The bulk of the headline 17.55% was earned in a single early fold and hasn't been repeated since. With a Deflated Sharpe Ratio of just 0.342 (versus a raw PSR of 0.814 across 6 trials), the framework is right to withhold its blessing: the edge looks front-loaded and possibly regime-dependent, not durable.
Recent live activity
Live behaviour echoes the fade. Scheduled runs from August 10–14 executed zero trades — each session logged a single rejected signal — leaving the book static at ~$2,752 cash and roughly $10,000 total equity. The last real activity was August 7, when the strategy sold 25 shares of DIS at $105.15, closing a June 10 entry at $99.18 for a clean ~6% mean-reversion round trip, and added a single share of CAT at $843.72. That DIS trade is exactly the pattern the thesis promises — but such setups are firing rarely right now.
Verdict
Bollinger-reversion is not broken, but it is not proven either. The high win rate and consistently non-negative folds argue it captures a genuine, if shrinking, reversion effect. Against that, a failed validation gate, an eroding out-of-sample curve, a 20% drawdown, and a near-idle live book all counsel caution. This is a strategy to keep on a short leash and small size — worth watching for whether the edge stabilises, not one to scale into on the strength of a backtest that already peaked.