The Thesis
bollinger-reversion runs one of the oldest ideas in technical trading: buy when price closes below the lower Bollinger band, sell when it pushes above the upper band. It trades a 24-name large-cap universe spanning tech (AAPL, MSFT, NVDA, GOOGL), financials (JPM, BAC, V, MA), staples (PG, KO, WMT, COST) and healthcare (JNJ, UNH, PFE). The bet is simple and well-worn: extreme deviations from a moving average tend to snap back.
Backtest Scorecard
Over 451 days the strategy returned 17.55% (9.46% CAGR), ending at $11,755 on a $10k base. The win rate is a healthy 63.89% across 76 trades, and Sharpe lands at 0.66. Those are the strengths. The risks are equally visible: a 20.57% max drawdown is steep for a book meant to harvest small reversions, and turnover of 1,720% means the edge has to survive heavy churn — though total fees came to just $76, so cost drag isn't the killer here.
Where Validation Balks
This is the part that should temper any enthusiasm. Walk-forward validation failed. On paper the folds look reassuring — all four are positive — but the trajectory is a clean, discouraging decay:
| Fold | Return | Sharpe | Max DD |
|---|---|---|---|
| 1 | 7.67% | 1.78 | 7.06% |
| 2 | 1.42% | 0.25 | 20.56% |
| 3 | 0.77% | 0.20 | 6.66% |
| 4 | 0.40% | 0.14 | 11.19% |
The most recent out-of-sample window returned just 0.4% at a Sharpe of 0.14 — effectively flat. The Probabilistic Sharpe Ratio (0.81) reads okay, but the Deflated Sharpe (0.342), which discounts for the 6 trials run, sits well below any reasonable confidence bar. Read together, the headline 17.55% looks front-loaded into fold 1; the live-forward edge has thinned toward noise.
Recent Activity: A Strategy Standing Still
The live tape tells the same story. Across six scheduled runs from Aug 12–19, the strategy executed zero trades and rejected 1–2 candidates each time, holding steady at ~$10k total with $2,752 in cash. The last real fills were back on Aug 7 (sold 25 DIS at $105.15, having bought at $99.18 in June; bought 1 CAT). Bands simply aren't reaching their extremes in current conditions, so the model sits on its hands.
Verdict
bollinger-reversion is a coherent, high-hit-rate strategy with a genuinely good backtest — but the validation layer is doing its job. The out-of-sample decay and failed walk-forward gate suggest the historical return owes more to one favorable regime than to a durable edge. It stays live and worth watching, but the current flat, no-trade stretch is consistent with a strategy whose signal has gone quiet. Treat the 17.55% as a ceiling, not an expectation.