The Thesis
RSI Snap-Back plays a simple, well-worn idea: large-cap tech names tend to snap back sharply after short-term momentum extremes. The rules are mechanical — buy the most oversold names when RSI drops below 35, exit when RSI climbs above 70, and rotate to keep a tight four-name book. The universe is the Mag-7: AAPL, MSFT, NVDA, GOOGL, AMZN, META, and TSLA.
The four-slot cap is the interesting design choice here. Rather than chasing every oversold signal, the strategy caps concurrent exposure at four positions, which enforces discipline and limits how much drawdown can pile up at once. It's a deliberately narrow, high-conviction structure.
Backtest Performance
Over a 451-day backtest, RSI Snap-Back returned 20.95%, growing a $10,000 book to $12,095. That annualizes to a CAGR of roughly 11.21%. The win rate is a healthy 66.67% across 37 trades, suggesting the reversion signal fired correctly about two times in three.
The less flattering numbers sit right next to those wins. Sharpe comes in at 0.61 — positive, but modest, telling us the returns came with meaningful volatility. Max drawdown reached 23.73%, a deep hole for a strategy that explicitly markets itself on limiting concurrent drawdown exposure. Turnover is high at 773%, though total fees over the full run were a trivial $37, so trading costs are not eating the edge here.
The Validation Gap
The most important caveat: there is no validation data. The validation field is null, meaning we have an in-sample backtest and nothing that tests the edge on unseen data. A 66.67% win rate and 20.95% return look attractive, but without out-of-sample or walk-forward confirmation, we can't distinguish a real reversion edge from a curve fit to one favorable 451-day window. For a strategy flagged live, that gap deserves attention before sizing up.
Recent Live Activity
And here's the twist: live, the strategy is doing nothing. Its last six scheduled runs — from July 17 through July 24 — each report 0 executed, 0 rejected, with cash and total equity both flat at $10,000. No RSI reading in the Mag-7 has crossed below the 35 entry threshold during this stretch, so the book has sat entirely in cash.
That's not a malfunction; it's the strategy behaving as designed. Mean-reversion systems are patient by nature and only act at extremes. But it's a reminder that the headline backtest figures reflect a full history, not the recent flatline — anyone watching the live equity curve should expect long idle spells punctuated by bursts of activity.
Verdict
RSI Snap-Back is a clean, disciplined idea with a respectable in-sample record and a sensible risk cap. The two things to watch are the unvalidated edge and that 24% drawdown, which is at odds with the strategy's own risk pitch. Promising, but unproven where it counts.