The thesis
Mean-reversion is about as classic as systematic equity trading gets: buy when a name is oversold (RSI below 30), sell when it's overbought (RSI above 70). It runs across a 24-stock universe of US large caps — the usual megacap tech, financials, staples, healthcare and energy. The bet is that short-term price extremes in liquid, heavily-followed names tend to snap back. It's a clean, legible idea, which is part of its appeal.
Backtest: the flattering headline
Over a 451-day backtest the strategy returned 14.73% (final equity $11,473 on an implied $10k), a 7.98% CAGR, with a 70.59% win rate across 38 trades. Fees were trivial at $38 and there was no FX cost. On the surface this is the kind of record that gets a strategy promoted to live.
Look closer and the picture softens. The Sharpe ratio is only 0.58 — modest for a 70% hit rate, which tells you the losers are larger than the winners. Max drawdown was 15.64%, and turnover ran to 879%, so this is an active book, not a buy-and-hold proxy.
Validation: it failed
This is the part that matters, and the verdict is blunt: validation did not pass. Across four walk-forward folds, three were positive — +2.06%, +11.1% and +2.21% — but the most recent fold (Dec 2025–May 2026) lost -2.84% with a Sharpe of -0.33 and a 14.96% drawdown. The single big fold (Sharpe 1.32) is carrying the whole record.
The deflated Sharpe ratio (DSR 0.304) is the red flag. With 6 trials in the search, DSR discounts the headline Sharpe for the luck of selection, and 0.30 is weak. PSR of 0.785 is better but not decisive. In plain terms: the edge looks real in-sample but thin and possibly overfit once you account for how many variants were tried, and the freshest out-of-sample window went the wrong way.
Live activity: quietly stuck
The live paper account tells a consistent story. Recent executions were all buys — MCD, HD, BAC and WMT accumulated through September — and the book is now nearly fully invested, with only $329.71 cash. Since then, six consecutive scheduled runs (Sep 25–Oct 2) executed nothing: no new oversold entries, and critically no overbought exits. Meanwhile total equity drifted down from $9,692 to $9,357 — a roughly 3.5% slide while the strategy sat on its hands.
That's the mean-reversion trap in miniature: the sell leg only fires on RSI>70, so positions that keep sagging simply aren't sold, and the book rides the drawdown.
Verdict
The strengths are genuine — high win rate, three of four positive folds, disciplined low-cost execution. But the risks are more persuasive right now: a failed validation gate, a negative most-recent fold, a DSR that smells of overfitting, and a live account that is underwater and fully committed with no exit signals triggering. I'd treat the 14.73% as a ceiling, not an expectation, and keep this strategy on a short leash until a clean out-of-sample stretch earns it back.