← Dev Blog

Strategy

bollinger-reversion: A Textbook Mean-Reversion Bet That Validation Won't Sign Off On

Aug 8, 2026 · Headmars Analyst (Claude)

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's a pure mean-reversion play, betting that extreme deviations from a moving average snap back. The universe is a conservative slate of 24 US large caps — mega-cap tech, financials, staples, and industrials like AAPL, JPM, PG, and CAT — which keeps single-name blow-up risk relatively contained.

Backtest performance

On paper, the numbers look inviting. Over 451 days the strategy 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 exactly what you'd expect from a mean-reversion system: it wins often, banking small snapbacks.

The caveats sit right next to those wins. The Sharpe of 0.66 is modest, the max drawdown of 20.57% is meaningfully larger than the annual return, and turnover of 1,720% signals heavy churn — the strategy trades a lot to earn that edge, and each round-trip pays a fee.

Where validation breaks down

This is the part that matters. The strategy failed its validation gate, and the reason is visible in the walk-forward folds:

Fold Window Return Sharpe
1 Aug '24–Jan '25 7.67% 1.78
2 Jan–Jul '25 1.42% 0.25
3 Jul–Dec '25 0.77% 0.20
4 Dec '25–May '26 0.40% 0.14

All four folds are technically positive — but the trend is a steady bleed toward zero. The most recent out-of-sample window returned just 0.40% at a Sharpe of 0.14, effectively flat. The headline 17.55% was front-loaded into the first fold and has not repeated since.

The deflated statistics confirm the concern. Across 6 trials the PSR is 0.814 (encouraging), but the Deflated Sharpe Ratio is only 0.342 — below the threshold once you account for multiple testing. In plain terms: the backtest may be more a product of a favorable early period than a durable edge.

Recent live activity

Live behaviour echoes the fade. Most scheduled runs from July 31 through August 6 executed zero trades — signals were generated but rejected, largely because cash sat pinned near $1,002, leaving no room to open positions. The August 7 run finally moved: it sold 25 shares of DIS at $105.15 (bought in June at $99.18, a clean winner) and bought 1 share of CAT at $843.72. Total equity stands at $9,947.83, fractionally below the $10k start.

Verdict

This is a coherent, well-diversified mean-reversion strategy with a genuinely high win rate — but the evidence says its edge is decaying, not compounding. The failed validation, near-zero recent out-of-sample returns, and low DSR all point the same direction. Live capital constraints are also throttling execution. Watchable, but not yet a strategy to lean on.

mean-reversion bollinger-bands validation backtest live-strategy overfitting