The thesis
Mean-reversion is one of the oldest ideas in systematic trading, and this strategy states it plainly: buy oversold names (RSI below 30) and sell overbought ones (RSI above 70). It runs on a 24-name universe of large-cap US equities spanning tech, financials, healthcare, staples and energy — AAPL, MSFT, NVDA, JPM, JNJ, XOM and the like. The premise is that short-term price extremes in liquid blue chips tend to snap back toward a mean, and RSI is the trigger that times the entries and exits.
Backtest performance
On paper the numbers are appealing. Over 451 trading days the strategy returned 14.73% (7.98% CAGR), ending at $11,473 on a $10k base. The 70.6% win rate across 38 trades is the headline strength — the edge, when it fires, tends to be right. Risk metrics are more sober: a Sharpe of 0.58 is unremarkable, and a 15.64% max drawdown means the equity curve was not comfortable to hold. Turnover of 879% signals an active book that pays real costs, though total fees came to just $38 with no FX drag.
What validation says
Here the story turns cautionary. Walk-forward validation failed the gate. Across four folds the strategy was positive in three, but the most recent fold (Dec 2025–May 2026) returned -2.84% with a -0.33 Sharpe — precisely the window that best proxies today's regime. Aggregate out-of-sample performance was negative, and while the Probabilistic Sharpe Ratio (0.785) looks respectable, the Deflated Sharpe Ratio — which penalises for the six trials run — sits at just 0.304. That gap between PSR and DSR is the tell: after accounting for how many variants were tested, the evidence for a durable edge thins considerably. Fold 2 (11.1% return, 1.32 Sharpe) is doing a lot of the heavy lifting for the full-sample figure.
The live book
The paper account reflects this ambivalence. The last six scheduled runs (Aug 17–24) each executed zero trades — RSI simply hasn't reached its extremes, so the strategy is patiently in cash ($7,569 of a ~$9,795 total). The last actual fill was a WMT buy back on 31 May. Total equity sits just below the $10k mark, so live-to-date the strategy is fractionally underwater and mostly idle.
Verdict
Mean-reversion has a genuine, intuitive thesis and a backtest that flatters it. But a failed validation, a negative recent fold, and a deflated Sharpe near 0.3 are exactly the signals that separate a real edge from an overfit one. The high win rate is worth watching; the risk-adjusted returns and out-of-sample decay are worth respecting. For now, this is a strategy on probation — live, but not yet earning conviction.