Thesis
Bollinger Reversion is a classic mean-reversion play: buy when price closes below the lower Bollinger Band, sell when it crosses above the upper band. The universe is 24 large- and mega-cap U.S. equities spanning technology, financials, healthcare, consumer staples, and energy — names with deep liquidity and well-documented reversion tendencies after short-term dislocations.
Backtest Snapshot
Over a 451-day full-sample backtest, the strategy produced a 17.55% total return (9.46% annualised) on 76 trades. A 63.89% win rate is genuinely respectable for a mean-reversion approach — more than three in five entries resolved in the strategy's favour. Maximum drawdown reached 20.57%, which is the number that demands attention: for a portfolio of blue-chips, a one-fifth peak-to-trough decline is meaningful and implies the strategy can sit underwater for extended stretches during trending regimes.
The Sharpe of 0.66 is modest. It reflects adequate but not exceptional risk-adjusted returns, and it sets up the more revealing picture that walk-forward testing paints.
Walk-Forward Validation
The four-fold walk-forward results are where the story gets complicated.
| Fold | Period | Return | Sharpe | Max DD |
|---|---|---|---|---|
| 1 | Aug 2024 – Jan 2025 | +7.67% | 1.78 | 7.06% |
| 2 | Jan 2025 – Jul 2025 | +1.42% | 0.25 | 20.56% |
| 3 | Jul 2025 – Dec 2025 | +0.77% | 0.20 | 6.66% |
| 4 | Dec 2025 – May 2026 | +0.40% | 0.14 | 11.19% |
All four folds are positive — that is the good news, and it distinguishes bollinger-reversion from strategies that simply overfit to a lucky in-sample window. But the trend is unmistakable: returns and Sharpe are compressing fold-over-fold. Fold 1 was a strong 1.78 Sharpe; by Fold 4 (the most recent and therefore most predictive period), the Sharpe has collapsed to 0.14 and the out-of-sample return sits at just 0.4%.
The probabilistic Sharpe ratio (PSR) of 0.814 indicates the strategy's Sharpe is likely above zero, but the deflated Sharpe ratio (DSR) of 0.342 — which corrects for the six trials tested — brings the adjusted confidence well below conventional thresholds. Headmars' validation gate has accordingly flagged this strategy as not passed.
Recent Activity
The strategy has been quiet in live paper trading. Runs on June 11–17 executed zero trades, with one or two signals rejected each day. The most recent execution was a 25-share buy of DIS at $99.18 on June 10, preceded by a sell of PG at $148.40 on June 9 — a clean round-trip entry below the lower band and exit above the upper band. Earlier in the period the strategy accumulated positions in GOOGL, PG, WMT, and COST during a brief cluster of signals at end of May. Portfolio value has drifted from $10,101 to $10,281 over the past week, a modest positive drift.
The silence since June 10 is consistent with the strategy's selective nature: in quiet or trending markets, price rarely breaches the outer bands long enough to trigger entries.
Assessment
Strengths: A positive win rate above 60%, all-positive walk-forward folds, and a focused large-cap universe keep this strategy worth monitoring. The thesis is grounded in a well-documented market microstructure phenomenon.
Risks: The decay in fold performance is a red flag for regime dependence. A max drawdown of 20.57% — occurring almost entirely in Fold 2 during early 2025 — suggests the strategy is poorly equipped for sustained directional moves. High portfolio turnover (1,720% annualised) also means transaction cost assumptions matter significantly at scale.
Bollinger Reversion earns its live status as a paper-trading laboratory, but it will need either a regime filter or a materially stronger recent fold before Headmars would consider graduating it toward real capital.