The thesis
Bollinger-reversion is a textbook mean-reversion play: buy when price closes below the lower Bollinger band, sell when it pushes above the upper band. It trades a universe of 24 large-cap US names — the usual megacap tech (AAPL, MSFT, GOOGL, NVDA), financials (JPM, V, MA), staples (PG, KO, WMT, COST) and industrials (CAT, HON). It is currently live.
The premise is sound in principle: liquid large caps tend to overshoot and snap back, and a band-based signal is a disciplined way to systematise "buy the dip, sell the rip." The question is whether the historical edge is real or an artefact of a favourable sample.
Recent activity
The strategy runs on a scheduled cadence, executing a handful of orders per session. Over the last week it bought NVDA (13 shares at $184.15), rotated AAPL — selling 7 shares at $352.27 the day after buying 7 at $329.82 — and did the same round trip in MSFT, alongside earlier buys in CAT and GOOGL. Runs typically fire one or two trades with zero rejections.
Live equity has been treading water: the Nov 5 run reported a total of $9,936.04, with the preceding sessions clustered between roughly $9,830 and $10,079. In short, the live book is hovering just under flat — no disaster, but no confirmation of the backtest's returns yet.
Backtest and validation
On paper the headline numbers are attractive. Over 451 days the strategy returned 17.55% (9.46% CAGR), with a 63.89% win rate across 76 trades and a Sharpe of 0.66. Max drawdown was 20.57% — not trivial, but survivable. Turnover, however, is extremely high at 1,720%, so this is an active strategy whose edge must clear real trading friction.
Validation is where the story sours. The walk-forward test failed. All four out-of-sample folds were positive, which sounds encouraging — but look at the trend: fold 1 returned 7.67% at a 1.78 Sharpe, then 1.42%, then 0.77%, then just 0.40% at a 0.14 Sharpe in the most recent fold. The edge is almost entirely front-loaded in the earliest window and has bled away since. Aggregate OOS return is a feeble 0.4%.
The statistical gates tell the same tale. The Probabilistic Sharpe Ratio of 0.814 looks healthy, but the Deflated Sharpe Ratio is only 0.342 — once you account for the 6 trials behind this strategy, the Sharpe is no longer convincingly distinguishable from luck.
Verdict
Strengths: a coherent thesis, a high win rate, and consistently positive (if shrinking) folds. Risks: a decaying out-of-sample edge, a failing DSR, punishing turnover, and a 20%+ drawdown. The high win rate masks a thin and fading expectancy. Keep it live as a small, closely-watched experiment — but it has not earned a larger allocation.