The thesis
Mean-reversion runs a textbook contrarian rule across a 24-name large-cap universe (AAPL, MSFT, JPM, XOM, and peers): buy when RSI falls below 30, sell when it climbs above 70. The premise is that short-term extremes in momentum tend to snap back. It is simple, interpretable, and cheap to run — the full backtest incurred just $38 in fees and zero FX cost.
Backtest performance
On paper, the numbers are inviting. Over 451 days the strategy returned 14.73% (7.98% CAGR), turning $10,000 into $11,473 across 38 trades. The headline strength is a 70.59% win rate — most positions closed green. Risk was moderate: a 15.64% max drawdown and a Sharpe of 0.58. The one caveat buried in the stats is turnover of 879%, meaning the book churned roughly nine times over — acceptable here only because fees stayed trivial.
Where validation pushes back
This is where the story turns cautious. Walk-forward validation failed. Across four out-of-sample folds, three were positive, but the most recent — spanning December 2025 to May 2026 — returned -2.84% with a Sharpe of -0.33 and a 14.96% drawdown. That is not a rounding error; it is the strategy's own reported out-of-sample result, and it lands in the period closest to today.
The robustness statistics reinforce the concern. The Probabilistic Sharpe Ratio sits at a respectable 0.785, but the Deflated Sharpe Ratio is only 0.304 after adjusting for six trials. A DSR that low signals the in-sample edge may be substantially the product of selection rather than genuine signal. In plain terms: the backtest flatters a rule that has not held up cleanly when tested on unseen data.
Recent activity: parked in cash
Live behaviour matches the caution. The last executed trade was a single WMT buy (21 shares at $115.75) back on 31 May 2026. Every scheduled run since — through mid-September — reports 0 executed, 0 rejected, with the book holding $7,569 in cash against a total value drifting between roughly $9,788 and $9,872. That total sits below the $10,000 starting line, so the live account is modestly underwater while it waits. The RSI triggers simply are not firing: a placid, range-bound market gives a mean-reversion rule little to do.
The verdict
Mean-reversion is a clean, low-cost strategy with a genuinely high hit rate and a mostly-positive fold history — reasons it earned a live slot. But the balance of evidence urges restraint. The failing validation, the negative most-recent fold, and a deflated Sharpe near 0.3 all point to an edge that is thinner than the 14.73% headline suggests. Its current idleness is arguably prudent rather than broken: no oversold names, no forced trades. We would keep it live on a short leash, watching whether the next batch of signals confirms the backtest or the out-of-sample warning.