Thesis
The mean-reversion strategy runs on a simple, well-worn premise: buy the oversold, sell the overbought. Concretely, it enters when RSI falls below 30 and exits when RSI climbs above 70, applied across a 24-name universe of large-cap US equities spanning tech (AAPL, MSFT, GOOGL, NVDA), financials (JPM, BAC, V, MA), healthcare, staples, and energy. It is currently classified as live.
Backtest Performance
On paper, the full-sample backtest is encouraging. Over 451 days the strategy returned 14.73% (7.98% CAGR), turning a $10,000 book into $11,473. The headline win rate is a striking 70.59% across 38 trades — mean-reversion systems typically win often and small, and that pattern holds here.
The caveats sit in the risk column. Sharpe is a modest 0.58, and maximum drawdown reached 15.64% — larger than the annualised return, which is not a comfortable ratio. Turnover ran hot at 879%, though fees stayed negligible at $38 total. High turnover with a middling Sharpe suggests the edge is thin relative to the churn.
The Validation Problem
This is where enthusiasm should cool. The strategy failed walk-forward validation. Across four out-of-sample folds, three were positive, but the aggregate out-of-sample return was -2.84% with an OOS Sharpe of -0.33.
The fold-by-fold record tells the story. Folds 1–3 delivered +2.06%, +11.10%, and +2.21%. But the most recent fold (Dec 2025 – May 2026) lost -2.84% and suffered a 14.96% drawdown — the strategy's weakest window is also its freshest. A deflated Sharpe ratio (DSR) of just 0.304 against six trials, versus a probabilistic Sharpe (PSR) of 0.785, points to an edge that shrinks materially once you account for selection across trials. In plain terms: the recent regime has not been kind to buying dips in these names.
Recent Live Activity
Live behaviour has been unremarkable — arguably too quiet. The last six scheduled runs (18–25 Aug) each executed zero trades, with cash parked at $7,569.25 and total equity drifting from $9,998 down to $9,791. The last actual fill was a WMT buy (21 shares at $115.75) back on 31 May. With no RSI extremes triggering, the strategy is effectively sitting on its hands while its paper book slips modestly below its $10,000 starting line.
Verdict
Mean-reversion has a genuine strength — a high, consistent win rate — but two flashing amber lights: a drawdown that exceeds its return, and a validation failure driven by a deteriorating most-recent fold. The signal that worked in 2024–early 2025 has not carried into 2026. Until the out-of-sample record stabilises, this is a strategy to watch rather than to scale. The current dormancy is a reasonable posture; forcing trades in the absence of clear oversold signals would only add risk without evidence of edge.