The thesis
Mean-reversion runs one of the oldest ideas in technical trading: buy weakness, sell strength. Concretely, it accumulates names when the 14-day RSI drops below 30 (oversold) and trims them when RSI pushes above 70 (overbought). The universe is 24 large-cap US equities spanning tech (AAPL, MSFT, NVDA), financials (JPM, BAC, V, MA), staples (PG, KO, WMT, COST), healthcare (JNJ, UNH, PFE), and energy (XOM, CVX). It runs live as a paper-traded book.
Backtest performance
Over 451 days the strategy returned 14.73% (7.98% CAGR), lifting a $10,000 book to $11,473. The headline strength is a 70.59% win rate across 38 trades — better than two in three closed positions were profitable. Risk-adjusted return is more modest: a Sharpe of 0.58 with a 15.64% max drawdown. Turnover is high at 879%, meaning the book effectively recycled its capital nearly nine times, though total fees stayed low at $38.
A high win rate paired with a middling Sharpe is a familiar mean-reversion signature: many small wins, punctuated by the occasional sharp loss when an oversold name keeps falling. The drawdown figure is where that tail shows up.
Validation: the part that matters
This is where the picture darkens. Walk-forward validation fails. Across four folds, three were positive — +2.06%, +11.10%, +2.21% — but the most recent fold (Dec 2025 to May 2026) posted −2.84% with a −0.33 Sharpe and a 14.96% drawdown. The out-of-sample stretch is the one that counts most, because it's the regime closest to today, and it went the wrong way.
The robustness statistics are mixed. The Probabilistic Sharpe Ratio of 0.785 is respectable, but the Deflated Sharpe Ratio — which penalizes for the 6 trials run during search — sits at just 0.304, comfortably below the level that would justify confidence that the edge is real rather than selection noise. Three-of-four positive folds is encouraging; a losing, higher-drawdown final fold is not.
Recent activity
The live book has been quiet. Scheduled runs from Aug 12 through Aug 19 each executed zero trades — the strategy is holding $7,569.25 in cash against a total of roughly $9,984, leaving it slightly underwater versus a $10k start and largely uninvested. The last actual fill was a buy of 21 WMT at $115.75 on May 31. In other words, RSI simply hasn't crossed its thresholds; nothing in the universe has been oversold enough to buy or overbought enough to sell for nearly three months.
Verdict
Mean-reversion has an intuitive thesis and an attractive hit rate, but the evidence is not yet compelling. The failing validation, weak deflated Sharpe, and a negative most-recent fold all argue for caution, and the current all-cash posture means it is contributing little either way. The strengths are real; so are the risks. This is a strategy to keep on a short leash and watch — not one to lean on.