The thesis
Mean-reversion is about as classic as systematic trading gets: buy oversold names (RSI below 30), sell overbought ones (RSI above 70), and collect the snap-back. The strategy runs on a 24-name large-cap universe spanning tech, financials, healthcare, staples, and energy — AAPL, MSFT, NVDA, JPM, V, JNJ, PG, KO, WMT, XOM, and more. It is currently live.
Recent activity
Trading has been quiet. The last six scheduled runs (2026-09-25 through 10-01) each reported 0 executed, 0 rejected — unsurprising, since mean-reversion only acts when RSI hits an extreme, and a trendless or steadily rising tape simply doesn't trigger entries. The most recent fill was a 10-share MCD buy at $238.58 on 2026-09-24, following HD (8 @ $299.12) and BAC (42 @ $58.21) earlier in September, and a WMT position opened back in May.
More sobering is the equity trajectory embedded in those run logs. Total account value slipped from $9,692 on 2026-09-25 to $9,356 on 2026-10-01 — roughly a 3.5% drawdown over the window, and notably below the implied $10k starting stake. Cash has sat flat at $329.71, so the account is effectively fully invested and riding the holdings down. That is the live reality against which the headline backtest should be read.
Backtest and validation
On paper the numbers are appealing. Over 451 days the backtest returned 14.73% (7.98% CAGR), with a 70.6% win rate across 38 trades, a Sharpe of 0.58, and a max drawdown of 15.64%. Turnover of 879% is high but fees were a trivial $38.
The honest caveat is validation, which the strategy did not pass. Walk-forward testing across four folds shows three positive — folds two and three did the heavy lifting (+11.1% and +2.21%) — but the most recent fold (2025-12-16 to 2026-05-29) lost 2.84% with a negative Sharpe of -0.33. That out-of-sample weakness is the tell. The Deflated Sharpe Ratio of 0.304 sits well below the confidence we'd want after six trials, even though the Probabilistic Sharpe Ratio (0.785) looks healthier in isolation.
Balance of evidence
Strengths: a genuinely high win rate, intuitive and well-understood mechanics, low fee drag, and a diversified universe that avoids single-sector concentration.
Risks: the strategy's edge decays in exactly the environment it has met most recently — the newest fold and the live account are both red. A 70% win rate paired with a middling Sharpe implies the losers are larger than the winners, which is the signature failure mode of mean-reversion when a position keeps falling instead of reverting. The high turnover also leaves it exposed if fills or spreads worsen.
Verdict
Mean-reversion earns its place as a live experiment, but the failed validation and the recent live drawdown argue for treating it as unproven rather than deployable at size. We'll be watching whether the next RSI-extreme entries revert — or whether this is a regime the strategy simply isn't built for.