The thesis
Bollinger-reversion is a classic mean-reversion play: buy when price dips below the lower Bollinger band, sell when it pushes above the upper band. It runs on a 24-name large-cap universe spanning tech (AAPL, MSFT, NVDA), financials (JPM, V, MA), staples (PG, KO, COST) and healthcare (JNJ, UNH). The logic is intuitive — bet that stretched prices snap back to their moving average — and it's the kind of strategy that tends to look good on paper.
Backtest performance
Over 451 days the strategy returned 17.55% (CAGR 9.46%), lifting a $10k book to $11,755. The win rate is a healthy 63.89% across 76 trades, and fees were negligible at $76 total with no FX drag. Those are genuine strengths.
The warning signs are in the risk and activity numbers. The Sharpe is only 0.66, meaning returns came with meaningful volatility, and the max drawdown of 20.57% is steep for a book that returned 17.55% — you'd have stomached a drop larger than your full-period gain at the worst point. Turnover of 1,721% also tells us this is a busy strategy churning its capital roughly 17x over, which raises the bar for any live slippage or fee assumptions to hold.
Validation: the honest red flag
This is where the picture turns. Walk-forward validation failed. On the surface all four folds were positive — a clean 4/4 — but that masks a clear decay: fold 1 returned 7.67% at a Sharpe of 1.78, then fold 2 (1.42%), fold 3 (0.77%) and fold 4 (0.40%) marched steadily toward zero, with Sharpe collapsing from 1.78 to 0.14. The aggregate out-of-sample return was just 0.40% at a 0.14 Sharpe. A deflated Sharpe ratio of 0.34 — adjusted for the 6 trials run — says the edge does not clear the bar once you account for how many variants were tested. The 0.81 PSR offers only faint comfort against that.
The read: the early backtest window flattered this strategy, and its edge has been eroding in more recent, unseen data.
Recent live activity
Live, the book sits at roughly $9,998 on $2,677 cash — essentially flat since going live, having clawed back from ~$9,681 a week earlier. Execution has been sparse: most scheduled runs over the past week logged 0 executed with 1–2 rejected orders, with the 15 Oct run the exception (2 executed). Notably, it round-tripped NKE — bought 70 shares at $35.64 on 18 Sep, sold at $37.56 on 15 Oct — a small, clean reversion win that's the thesis working as designed.
Verdict
Strengths: a coherent, well-understood edge, a strong win rate, and low trading costs. Risks: a sub-1 Sharpe, a drawdown bigger than its total return, and — most importantly — a validation failure showing the out-of-sample edge fading toward zero. This is a strategy to watch and keep on a tight leash, not one to size up.