Strategy Thesis
The mean-reversion strategy operates on a straightforward premise: large-cap equities tend to snap back toward fair value after short-term extremes. When RSI drops below 30, the strategy buys the dip; when RSI exceeds 70, it exits the position. The universe is intentionally blue-chip — 24 names spanning tech (AAPL, NVDA, MSFT), financials (JPM, V, MA), healthcare (JNJ, UNH, ABBV), consumer staples (PG, KO, WMT), and industrials (CAT, HON) — reducing idiosyncratic blow-up risk while keeping liquidity ample.
The logic is sound in theory: institutional rebalancing and options market-making tend to dampen sustained momentum in mega-caps, creating the elastic behavior this strategy exploits.
Backtest Performance
Over 451 days, the full backtest returned 14.73% (CAGR ~7.98%), finishing with equity of $11,473 on a $10,000 starting base. The Sharpe of 0.58 is modest but acceptable for a rules-based long-only system with no leverage. The 70.6% win rate across 38 trades is the headline number — more than seven in ten entries resolved profitably.
The maximum drawdown of 15.64% is the main caveat on the backtest side. For a strategy trading only large-caps with tight RSI thresholds, that figure suggests the exits aren't always clean; a position can sit underwater for a meaningful stretch before RSI reaches the sell threshold.
Turnover of 879% annualized is high — a consequence of the binary entry/exit logic — but because this is paper trading with flat $1 per-trade fees and no FX exposure, transaction drag is minimal in simulation.
Walk-Forward Validation
This is where the picture gets more nuanced. The validation suite ran 4 chronological folds across the full period:
| Fold | Period | Return | Sharpe | Max DD | Trades |
|---|---|---|---|---|---|
| 1 | Aug 2024 – Jan 2025 | +2.06% | 0.57 | 4.24% | 8 |
| 2 | Jan 2025 – Jul 2025 | +11.10% | 1.32 | 10.53% | 10 |
| 3 | Jul 2025 – Dec 2025 | +2.21% | 0.46 | 8.80% | 6 |
| 4 | Dec 2025 – May 2026 | −2.84% | −0.33 | 14.96% | 10 |
Three of four folds were profitable, but Fold 4 — the most recent — was the weakest, posting a loss and the highest drawdown of any period. The out-of-sample return of −2.84% and Deflated Sharpe Ratio of 0.304 (well below the 0.5 threshold typically required to pass) are what pushed the validation gate to failed.
The Probabilistic Sharpe Ratio of 0.785 suggests the in-sample edge has some statistical basis, but the DSR adjustment for multiple trials (6 parameter combinations tested) erodes confidence significantly.
Recent Activity
The strategy has been conspicuously quiet in July 2026. Six consecutive daily runs from June 30 through July 7 executed zero trades — the RSI thresholds simply haven't been breached in the current universe. Cash sits at $7,569.25 of a $9,916 total portfolio, implying the sole open position is a 21-share WMT lot purchased at $115.75 on May 31. Portfolio value has fluctuated narrowly between $9,848 and $9,959 over this window.
This extended idle period is a feature, not a bug — the strategy is disciplined and won't chase. But it does highlight a real operational risk: low signal frequency means returns depend heavily on a handful of trades per quarter.
Verdict
Mean-reversion has a credible thesis and respectable historical mechanics, but the most recent fold's underperformance is a yellow flag. The strategy appears to work best in choppy, range-bound regimes (Fold 2's Sharpe of 1.32 stands out) and struggles when trends persist. Before increasing allocation, it would be worth examining whether the RSI thresholds need regime-conditional adjustment — or whether a momentum filter could sit alongside the current signal to reduce whipsaw in trending markets.