Thesis
The mean-reversion strategy is one of the oldest in systematic trading: when a stock becomes statistically oversold — RSI falling below 30 — it is a candidate to buy; when it crosses above 70, it is time to exit. The assumption is that large-cap prices revert toward equilibrium after short-term dislocations caused by sentiment swings, thin liquidity, or macro noise.
Headmars's implementation applies this logic across a 24-stock universe of blue-chip names spanning technology, financials, healthcare, consumer staples, energy, and industrials — tickers like AAPL, MSFT, JPM, JNJ, and XOM. The diversification is intentional: RSI extremes in a single sector can cluster around macro events, so a broad universe reduces idiosyncratic timing risk.
Recent Activity
The strategy has been quiet. Daily runs from June 8 through June 15 recorded zero executions — RSI levels across the universe have not touched entry or exit thresholds in over two weeks. Portfolio equity has drifted modestly between roughly $10,062 and $10,107, with $7,569 sitting in cash.
The last executed trade was a buy of 21 shares of WMT at $115.75 on May 31. That position, along with the elevated cash balance, reflects a stance that the current market environment simply is not generating the kind of oversold extremes this strategy requires. In a trending or low-volatility regime, mean-reversion systems naturally go dormant — which is exactly what is happening here.
Backtest Performance
Over 451 days, the strategy returned 14.73% on a starting equity of roughly $10,000, compounding at a 7.98% CAGR. The win rate of 70.6% across 38 trades is genuinely strong — most trades that triggered did resolve in the expected direction. Fees were minimal at $38 total, and there were no FX costs given the all-USD universe.
The Sharpe ratio of 0.58 and max drawdown of 15.64% paint a picture of a strategy that earns its returns unevenly and with meaningful peak-to-trough risk. Turnover of 879% is also notable — positions cycle frequently, which amplifies slippage and execution risk in live trading beyond what the backtest captures.
Validation: Where the Story Gets Complicated
The strategy failed validation, and the walk-forward fold data explains why.
Of four time-sliced folds, three were profitable. Fold 2 (Jan–Jul 2025) was the standout, delivering +11.1% with a Sharpe of 1.32. But fold 4 — the most recent window, December 2025 through May 2026 — produced −2.84% with a Sharpe of −0.33 and a drawdown approaching 15%. This is the period the strategy would have been live, and it underperformed.
Two statistical flags reinforce the concern. The Probabilistic Sharpe Ratio (PSR) of 0.785 means there is roughly a 22% chance the observed Sharpe is a fluke of zero or below — acceptable but not comfortable. More tellingly, the Deflated Sharpe Ratio (DSR) of 0.304 adjusts for the fact that this is one of six tested configurations. After accounting for multiple testing bias, fewer than one in three trials would be expected to show a real edge at this level. That is a meaningful red flag against treating the backtest results as a reliable forecast.
Assessment
Mean-reversion on large-cap equities is a legitimate and time-tested approach. The strategy's high win rate and diversified universe are genuine strengths. The concern is not the thesis — it is the evidence that recent market conditions have eroded the edge, and that the backtest Sharpe may reflect parameter fitting more than a durable signal. Until the strategy demonstrates positive out-of-sample performance across the next validation window, it warrants cautious position sizing and close monitoring.