The thesis
Mean-reversion is one of the oldest ideas in systematic trading, and this agent implements the textbook version: buy oversold names (RSI below 30), sell overbought ones (RSI above 70). It runs across a 24-name universe of US large caps — the megacap tech complex (AAPL, MSFT, GOOGL, NVDA), payments (V, MA), healthcare (JNJ, UNH, PFE, ABBV), and defensive consumer staples (PG, KO, WMT, COST). The bet is simple: extreme short-term moves overshoot and snap back. When it works, it works often; the discipline is entirely mechanical.
Backtest performance
On paper, the numbers flatter the strategy. Over 451 trading days it returned 14.73% (final equity $11,473 on a $10k base), a 7.98% CAGR, and a 70.59% win rate across 38 trades. That hit rate is the headline — more than two in three trades closed green.
But the quality metrics temper the enthusiasm. The Sharpe ratio of 0.58 is modest for a strategy carrying a 15.64% max drawdown, meaning returns came with real volatility rather than smooth compounding. Turnover of 879% signals heavy churn; at that pace, slippage and spread in the real world would eat meaningfully into the reported edge, even though modeled fees were just $38.
Validation: the red flag
This is where the story turns. Walk-forward validation failed. Across four out-of-sample folds, three were positive — folds 1 through 3 returned 2.06%, 11.10%, and 2.21% — but the most recent fold (Dec 2025–May 2026) lost 2.84% with a negative Sharpe of −0.33 and a 14.96% drawdown. The out-of-sample results are the honest ones, and they trend the wrong way.
The statistical gates say the same. With 6 trials tested, the Deflated Sharpe Ratio sits at just 0.304 — below the threshold that would give confidence the edge survives multiple-testing bias. A PSR of 0.785 is respectable but not decisive. In plain terms: the impressive full-sample return may be partly luck harvested from an in-sample fit.
Live activity
The live book reflects this caution. The last executed trade was a buy of 21 shares of WMT at $115.75 on May 31. Since then, every scheduled run — August 7 through 14 — reported 0 executed, 0 rejected. The account holds $7,569 in cash against a ~$9,995 total, so it is largely parked and marginally underwater versus its starting balance. No RSI extremes have triggered, which is exactly what you'd expect in a calm, range-bound tape.
The verdict
Mean-reversion has a genuine strength — a high win rate and intuitive, well-understood logic. But the failing validation, deteriorating final fold, and low DSR are warnings that its backtest edge is fragile and possibly overfit. The strategy is behaving responsibly by not forcing trades in the absence of signal. Until an out-of-sample stretch confirms the edge holds forward, this one belongs in observation, not conviction.