The thesis
Mean-reversion is one of the oldest ideas in systematic trading, and this strategy applies it in its most textbook form: buy the oversold, sell the overbought. Concretely, it enters when the 14-day RSI drops below 30 and exits when RSI climbs above 70, across a universe of 24 US large caps spanning tech, financials, healthcare, staples, and energy — names like AAPL, NVDA, JPM, JNJ, XOM, and MCD. The premise is that short-term price dislocations in liquid, well-covered stocks tend to snap back. It is a clean, interpretable thesis with no black-box discretion.
Recent activity
The strategy is now live, running on its scheduled daily cadence. Over the past week and a half it executed three buys: 42 shares of BAC at $58.21 (Sep 17), 8 shares of HD at $299.12 (Sep 21), and 10 shares of MCD at $238.58 (Sep 24). Most scheduled runs produced no trades at all — unsurprising for a strategy that only acts on genuine RSI extremes. Notably, the recent tape is all buys and no sells, consistent with a market where few of its holdings have reached overbought territory.
The portfolio total has drifted down over the window, from roughly $9,841 on Sep 22 to $9,681 on Sep 24, leaving it modestly below its $10,000 notional. Zero trades have been rejected, so execution is clean; the softness is market-driven, not operational.
Backtest and validation
On paper the strategy looks appealing. Over 451 days the backtest returned 14.73% (7.98% CAGR), with a 70.59% win rate across 38 trades, a Sharpe of 0.58, and a manageable 15.64% max drawdown. Fees were negligible ($38 total) despite a heavy 879% turnover.
The validation layer tells a more sober story, and it did not pass. A four-fold walk-forward test returned positive results in three folds — including a standout +11.1% (Sharpe 1.32) in early 2025 — but the most recent, out-of-sample fold (Dec 2025–May 2026) lost 2.84% with a negative Sharpe of −0.33 and its deepest drawdown yet (14.96%). The probabilistic Sharpe ratio is a healthy 0.785, but the deflated Sharpe ratio — which penalizes for the six trials run — collapses to 0.304, a red flag for how much of that backtest edge may be luck.
The balance
Strengths: a high win rate, low costs, an intuitive thesis, and clean live execution. Risks: the edge is fragile out-of-sample, the newest data is where it broke down, and the deflated statistics suggest overfitting cannot be ruled out. A high win rate paired with a mediocre Sharpe also hints at a payoff profile that wins small and loses big — exactly the shape that mean-reversion strategies suffer in trending or crashing markets.
The honest read: a promising, transparent strategy that earned its live slot on backtest merit but has not yet cleared the validation bar. Watch the live drawdown closely — the last fold is a warning, not a footnote.