The thesis
Mean-reversion runs a classic textbook rule across 24 US large caps: buy when RSI drops below 30 (oversold) and sell when RSI climbs above 70 (overbought). The universe spans mega-cap tech (AAPL, MSFT, NVDA), financials (JPM, V, MA), staples (PG, KO, WMT), and energy (XOM, CVX) — a deliberately broad, liquid basket where extreme readings are more likely to snap back than trend away.
Backtest performance
Over 451 days the strategy returned 14.73%, ending at $11,473 on a $10k start, for a 7.98% CAGR. The headline strength is a 70.59% win rate across 38 trades — most positions closed green. Risk metrics are more sobering: a Sharpe of 0.58 is modest, and a 15.64% max drawdown means the equity curve was not smooth. Turnover ran hot at 879%, though fees stayed negligible at $38 total. 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 a stock keeps falling instead of bouncing.
Validation: the red flag
This is where the story turns. Walk-forward validation failed. Of four out-of-sample folds, three were positive — but the most recent (Dec 2025–May 2026) returned -2.84% with a -0.33 Sharpe and a 14.96% drawdown, its worst fold on record. The aggregate OOS return of -2.84% sits far below the 14.73% full-sample figure, a classic in-sample/out-of-sample gap.
The deflated statistics tell the same tale. With 6 trials searched, the Probabilistic Sharpe Ratio is 0.785 but the Deflated Sharpe Ratio is just 0.304 — once you correct for multiple testing, confidence that the true Sharpe exceeds zero drops sharply. In plain terms: some of that backtest edge may be luck dressed up as skill.
Recent activity
The live paper account has been dormant. The last executed trade was a 21-share WMT buy at $115.75 on May 31. Since then, six consecutive scheduled runs (July 21–28) executed zero trades and zero rejections — no names hit the RSI extremes. Cash has been parked at $7,569.25 while total equity drifted between roughly $9,837 and $9,956, leaving the book modestly underwater against its $10k baseline.
Verdict
Mean-reversion is not broken, but it is not proven either. The high win rate and reasonable long-run return are genuine strengths; the failed validation, deflated Sharpe, and negative most-recent fold are genuine warnings. The current inactivity is arguably healthy — a disciplined strategy that refuses to force trades. But until out-of-sample results stabilize, this one belongs in the 'promising, unconfirmed' bucket rather than a capital allocation.