The thesis
Mean-reversion is about as textbook as a systematic strategy gets: buy oversold names when RSI drops below 30, sell overbought names when RSI pushes above 70. It trades a 24-name universe of US large caps — the usual megacap tech (AAPL, MSFT, GOOGL, NVDA), financials (JPM, BAC, V, MA), healthcare, staples, and energy. The bet is simple: short-term price extremes snap back, and a disciplined RSI trigger harvests that snap.
Backtest performance
On paper, the numbers flatter the thesis. Over 451 days the strategy returned 14.73% (7.98% CAGR), finishing at $11,473 on a $10k base, with a 70.59% win rate across 38 trades. Drawdown topped out at 15.64% and the Sharpe landed at a middling 0.58. Two things stand out. First, turnover is extremely high at 879% — this strategy churns, though fees stayed trivial at $38 total. Second, a 70% win rate paired with a sub-0.6 Sharpe tells you the losers are bigger than the many small winners. That asymmetry is exactly what you'd expect from fading trends: you're right often and occasionally very wrong.
Validation: the strategy fails the gate
This is where the story turns. Walk-forward validation did not pass. Across four sequential folds, three were positive — folds 1 through 3 returned 2.06%, 11.10%, and 2.21% — but the most recent fold (Dec 2025 to May 2026) returned -2.84% at a -0.33 Sharpe and a 14.96% drawdown. The out-of-sample read is negative on both return and risk-adjusted return.
The deflated statistics confirm the concern. The probabilistic Sharpe ratio (0.785) is reasonable, but after accounting for 6 trials the deflated Sharpe ratio drops to 0.304 — weak evidence that the edge is real rather than a product of selection. A strategy that earns most of its backtest return from one strong middle period (fold 2's 11.1%) and then stalls or loses in the most recent window is the signature of a fading edge, not a robust one.
Live activity
The strategy is running live and recently active: it executed 1 trade on Oct 29 (48 shares of BAC at $51.15), 2 on Oct 30 (GOOGL and CAT), and ran flat on Nov 2 with nothing executed. These are genuine oversold buys, consistent with the thesis. But the account tells the cautionary half: total equity has drifted $9,976 → $9,877 → $9,851, leaving it modestly below the $10k start. Early days, small sample — but it rhymes with the failed final fold.
Verdict
Mean-reversion is a coherent, cheap-to-run strategy with a genuinely high hit rate. The risk is equally clear: it failed validation on its most recent out-of-sample window, its deflated Sharpe is thin, and live performance has started soft. Worth watching, not yet worth trusting. The next few weeks of live fills will say more than the backtest did.