The Thesis
Mean-reversion is one of the oldest ideas in systematic trading, and this strategy states it plainly: buy oversold (RSI below 30), sell overbought (RSI above 70). It runs across a 24-name universe of US large-caps spanning tech (AAPL, MSFT, NVDA), financials (JPM, BAC, V, MA), healthcare (JNJ, UNH, PFE, ABBV), and staples/industrials (PG, KO, WMT, COST, CAT, XOM). The premise is that liquid mega-caps overshoot in the short term and snap back — a bet on noise, not trend.
Backtest Performance
On paper, the numbers look inviting. Over 451 days the strategy returned 14.73% (CAGR 7.98%), with a 70.59% win rate across 38 trades and final equity of $11,473. But the supporting stats temper the enthusiasm: a Sharpe of just 0.58, a 15.64% max drawdown, and eye-watering 879% turnover. A high win rate paired with a mediocre Sharpe is a classic mean-reversion signature — many small wins offset by occasional large losses when a name keeps falling. Fees were negligible ($38 total), so friction isn't the story here.
Validation: The Warning Sign
This is where the strategy stumbles. Walk-forward validation fails. Three of four folds were positive (+2.06%, +11.1%, +2.21%), but the entire edge is concentrated in fold 2, which alone returned 11.1% at a Sharpe of 1.32. The most recent fold (Dec 2025–May 2026) turned negative at -2.84% with a -0.33 Sharpe and the deepest drawdown of the set (14.96%).
The risk-adjusted diagnostics reinforce the caution. The Probabilistic Sharpe Ratio sits at 0.785, but the Deflated Sharpe Ratio is only 0.304 after accounting for 6 trials — meaning once we discount for selection, confidence that the true Sharpe exceeds zero is weak. In short, the headline return leans heavily on one favourable regime rather than a durable, repeatable edge.
Recent Live Activity
The live paper book tells a quieter story. The last executed trade was a buy of 21 WMT shares at $115.75 on 31 May. Every scheduled run since late July has produced zero executions and zero rejections — the RSI extremes simply aren't triggering. The account holds $7,569 in cash against a total value hovering around $9,900, below the $10,000 starting mark. The strategy is effectively parked, waiting for oversold conditions that a grinding, low-volatility tape hasn't delivered.
Verdict
Strengths: an intuitive, well-understood edge; high hit rate; low fee drag. Risks: failed out-of-sample validation, a deflated Sharpe near zero, regime-dependent returns, and a live book that is both underwater and inactive. This is a strategy worth watching, not scaling. Until the out-of-sample folds stabilise and the deflated Sharpe improves, mean-reversion should stay in observation — a promising sketch that hasn't yet proven it can trade through a regime it didn't train on.