The thesis
Bollinger-reversion runs a classic mean-reversion play across a 24-name large-cap universe (AAPL, MSFT, NVDA, JPM, V, JNJ, XOM, and the rest of the blue-chip roster): buy when price pushes below the lower Bollinger band, sell when it stretches above the upper band. It is a well-understood, well-behaved idea — the kind that tends to print a high win rate in calm, range-bound markets and get punished in trends. The strategy is currently flagged live.
Backtest performance
On paper the numbers look inviting. Over 451 days the strategy returned 17.55% (9.46% CAGR), finishing at $11,755 from a $10k base, with a 63.89% win rate across 76 trades and only $76 in fees. That win rate is the headline strength: nearly two in three trades closed green, which is consistent with a strategy that scalps small reversions.
The caveats sit right beside it. The Sharpe of 0.66 is modest — this is not a high-quality risk-adjusted return — and the 20.57% max drawdown is steep for a market-neutral-sounding premise. Turnover of 1,721% confirms this is a busy, churn-heavy book; the low fee total flatters what would look worse under a heavier cost model.
Validation: the part that matters
The walk-forward study is where enthusiasm should cool. Validation did not pass. All four folds were positive, which sounds reassuring, but the trajectory is the tell: fold returns decay from 7.67% (Sharpe 1.78) in late-2024 to 1.42%, then 0.77%, then 0.40% (Sharpe 0.14) by mid-2026. The out-of-sample return lands at 0.40% with an OOS Sharpe of 0.14 — essentially flat. A Probabilistic Sharpe Ratio of 0.814 looks healthy in isolation, but the Deflated Sharpe of 0.342 (adjusting for the 6 trials run) says the edge does not clearly survive multiple-testing scrutiny. The pattern reads like a strategy that fit the early sample well and has been giving that edge back ever since.
Recent activity
Live behaviour reinforces the concern. The last executed trades were on 2026-08-07 (bought 1 CAT at $843.72, sold 25 DIS at $105.15). Since then, six consecutive scheduled runs from 2026-09-09 through 09-16 have logged zero executions and two-to-three rejections each, with cash frozen at $2,752.28 and total equity drifting between roughly $9,570 and $9,764 — below the $10k mark. In other words, the book has been idle-by-rejection for over a month while the signal generates orders it cannot fill.
Verdict
Bollinger-reversion is a legible, high-win-rate idea with a clean in-sample story, but the evidence points to a fading edge: decaying folds, a deflated Sharpe below 0.5, near-zero OOS return, and a live account that is both underwater and effectively stalled. Strengths are real; they are just not durable enough to justify fresh capital without addressing why recent orders keep getting rejected and why the edge erodes out of sample.