Strategy Thesis
Bollinger Reversion is a classical mean-reversion rule: buy when price closes below the lower Bollinger Band, sell when it closes above the upper band. The universe is deliberately blue-chip — 24 large-cap names spanning tech (AAPL, MSFT, NVDA, GOOGL), financials (JPM, BAC, V, MA), healthcare (JNJ, UNH, PFE, ABBV), consumer staples (PG, KO, WMT, COST, MCD, NKE, HD), energy (XOM, CVX), industrials (CAT, HON), and media (DIS). The thesis is that these names revert reliably because institutional flows and index rebalancing dampen sustained directional momentum.
Backtest Performance
Over 451 days, the full-sample backtest produced a 17.55% total return (9.46% CAGR) across 76 trades. The 63.89% win rate is a genuine strength — more than three in five trades closed positive. Turnover ran high at 1,720.89%, which is characteristic of a band-touch strategy that exits quickly once price recovers.
The risk picture is less clean. A 20.57% maximum drawdown is substantial for a strategy targeting stable large-caps, and the Sharpe of 0.66 reflects meaningful volatility relative to the returns generated. Total fees of $76 (flat per-trade) are modest at this scale.
Walk-Forward Validation: The Warning Sign
The walk-forward results are where the strategy's edge comes under pressure. Four folds were run, and while all four ended positive on returns, the trajectory is concerning:
| Fold | Period | Return | Sharpe | Max DD | Trades |
|---|---|---|---|---|---|
| 1 | Aug 2024 – Jan 2025 | 7.67% | 1.78 | 7.06% | 24 |
| 2 | Jan 2025 – Jul 2025 | 1.42% | 0.25 | 20.56% | 11 |
| 3 | Jul 2025 – Dec 2025 | 0.77% | 0.20 | 6.66% | 18 |
| 4 | Dec 2025 – May 2026 | 0.40% | 0.14 | 11.19% | 20 |
Fold 1 was exceptional — a Sharpe of 1.78 on a tidy 7.67% return with minimal drawdown. Every subsequent fold declined monotonically in both return and risk-adjusted performance. By Fold 4, the out-of-sample return had compressed to 0.40% with a Sharpe of 0.14. The Probabilistic Sharpe Ratio of 0.814 is reasonable, but the Deflated Sharpe Ratio of 0.342 — which penalizes for multiple testing across 6 trials — suggests the true edge may be weaker than headline numbers imply. The strategy does not pass validation under the current gate.
Recent Live Activity
The most recent executed trades occurred in late May and early June: entries into WMT, COST, GOOGL, and PG, followed by a PG exit nine days later at $148.40 (bought at $140.03, a clean 5.98% gain), and a DIS entry at $99.18. Since then, the daily scheduled runs through late June and into early July have logged zero executions — only rejected signals. The portfolio value has oscillated in a roughly $9,700–$9,900 range over that period.
The rejection streak suggests prices in this universe are not reaching band extremes frequently enough to generate trade opportunities — consistent with a lower-volatility environment.
Strengths and Risks
Strengths: Positive returns across all four walk-forward folds; win rate above 60%; disciplined universe of liquid large-caps; simple, interpretable logic.
Risks: Return and Sharpe both decay sharply over time, suggesting either market-regime sensitivity or that early backtest performance was partially in-sample luck. The Fold 2 drawdown of 20.56% — nearly matching the full-backtest max drawdown in a single period — points to vulnerability during trending or dislocated markets. The current live idle period raises a question about signal frequency adequacy.
Outlook
Bollinger Reversion is a legitimate starting point, not a finished product. The strategy warrants parameter exploration (band width, lookback, position sizing) and a re-run of validation after enough new live data accumulates. Until the walk-forward trend reverses, treat live allocation with caution.