The Thesis
Mean-reversion is about as classic as a quant signal gets: buy what's oversold, sell what's overbought. Here the rule is mechanical — enter when 14-day RSI drops below 30, exit when it climbs above 70 — applied across a 24-name universe of US large caps spanning tech (AAPL, MSFT, GOOGL, NVDA), financials (JPM, BAC, V, MA), healthcare, staples and energy. The bet is that short-term price extremes in liquid blue chips tend to snap back. It's a hand-written, built-in strategy, approved by the reviewer at risk 0 back in May.
Backtest: Flattering, But Read the Fine Print
Over 451 days the strategy returned 14.73% (7.98% CAGR), turning $10,000 into $11,473 across 38 trades. The headline numbers look inviting: a 70.59% win rate and total fees of just $38. But the texture is less rosy. The Sharpe ratio is a modest 0.58, max drawdown reached 15.64%, and turnover ran to a hefty 879% — this is an active strategy that trades its whole book many times over. A high win rate paired with a mediocre Sharpe usually means the losers, when they come, are large relative to the many small wins.
Validation: The Part That Matters Failed
This is where I'd pump the brakes. The walk-forward validation did not pass. Across four folds, three were positive, but the most recent — Dec 2025 to May 2026 — returned -2.84% with a negative Sharpe of -0.33 and a 14.96% drawdown. The aggregate out-of-sample figures inherit that weakness: OOS return of -2.84% and OOS Sharpe of -0.33.
The deflated metrics tell the same story. With 6 trials in the search, the Probabilistic Sharpe Ratio is a decent 0.785, but the Deflated Sharpe Ratio is only 0.304 — below the confidence we'd want before trusting the edge is real rather than a product of selection. The strong in-sample result against weak recent out-of-sample performance is the textbook signature of a strategy that fit the past better than it predicts the future.
Live Activity: Mostly Waiting
The signal is patient by design, and lately it has been dormant. The Oct 29–30 runs put money to work — buying BAC (48 shares @ $51.15), then GOOGL (8 @ $280.28) and CAT (4 @ $622.54). Since then, four consecutive scheduled runs (Nov 2–4) executed zero trades: nothing is oversold enough to trigger. The paper account currently holds $2,742 in cash and a total value around $9,873 — fractionally below its $10,000 starting line.
Verdict
Mean-reversion has a clean, defensible thesis and a seductive backtest, but it's live despite a failed validation and a weak deflated Sharpe. The high win rate is real; so is the risk that the edge has decayed, as the latest fold suggests. Worth watching — with position sizing that respects the drawdown history, not the win rate.