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 across a 24-name large-cap universe spanning tech, financials, staples, healthcare, and energy — AAPL, MSFT, NVDA, JPM, V, JNJ, PG, WMT, XOM, and more. The bet is that short-term price extremes in liquid blue chips revert, and that a disciplined RSI rule can harvest that reversion repeatedly.
Recent activity
The strategy is live. Recent scheduled runs have been quiet: most sessions from 22–29 September executed zero trades, with a single execution on 24 September. The last few buys were MCD (10 shares @ $238.58), HD (8 @ $299.12), and BAC (42 @ $58.21). Notably, recent fills are all buys — the oversold side of the book — with no overbought sells triggering, leaving the account nearly fully invested at roughly $329.71 cash.
That matters because the live paper account is drifting down: total equity slipped from $9,841 on 22 September to $9,542 on 29 September, sitting below a $10k baseline. This is a live-vs-backtest gap worth watching, since the backtest ended at $11,473.
Backtest and validation
On paper the backtest looks appealing: 14.73% total return over 451 days, a 7.98% CAGR, a 70.59% win rate across 38 trades, and a manageable 15.64% max drawdown. Fees were a trivial $38 and FX cost was zero.
But the win rate flatters. A Sharpe of 0.58 is modest, and turnover of 879% signals a lot of churn for that return. More importantly, validation failed. Across four walk-forward folds, three were positive — +2.06%, +11.10%, and +2.21% — but the most recent fold (Dec 2025 to May 2026) returned -2.84% with a -0.33 Sharpe and a 14.96% drawdown. That out-of-sample degradation is the headline risk: the strategy's edge weakened precisely in the freshest, unseen window.
The deflation statistics reinforce caution. The Probabilistic Sharpe Ratio is a respectable 0.785, but the Deflated Sharpe Ratio — which adjusts for the 6 trials run — collapses to 0.304, well short of confidence. In plain terms, once you account for selection across trials, the observed Sharpe is not convincingly greater than zero.
The balance sheet
Strengths: a clear, well-understood thesis; a high win rate; positive results in three of four folds; low trading costs; and a genuine second-best fold (+11.10%, Sharpe 1.32) showing the edge can appear.
Risks: it failed validation, the newest fold was negative, the deflated Sharpe is weak, turnover is high, and live equity is currently below baseline with only buy-side activity. The 70% win rate coexisting with a middling Sharpe hints at small wins and occasional larger losers.
Verdict
Mean-reversion is a credible idea with an honest backtest and a dishonest-looking one lurking underneath. It stays live, but the failed out-of-sample check means it should be treated as unproven — worth monitoring, not scaling.