The Thesis
RSI Snap-Back bets that large-cap tech names overshoot in the short term and revert. The rule is deliberately mechanical: buy the most oversold names when RSI drops below 35, sell when RSI pushes above 70, and hold no more than four positions at once. That hard 4-slot book is the strategy's spine — it caps concurrent exposure and forces the system to rotate capital into the strongest reversion candidates rather than accumulating a sprawling, correlated book of Mag-7 names (AAPL, MSFT, NVDA, GOOGL, AMZN, META, TSLA).
It's a clean, legible idea. Mean reversion on liquid mega-caps is well-trodden ground, and the RSI thresholds here are conventional rather than overfit-looking, which is a point in its favour.
Recent Activity
The live book has been quiet. Across scheduled runs from 19 to 26 October, most sessions executed nothing — the strategy sat at $10,000 cash, waiting for an RSI < 35 trigger that didn't come. The one action of note was a 23 October buy of 6 shares of GOOGL at $321.67. After that purchase, the account holds $8,064.65 in cash against a total value of $9,967.61.
That total is worth dwelling on: it's slightly below the $10,000 starting line. The live record is early and the sample is tiny, but so far the realized equity curve has not yet confirmed the backtest's promise — a reminder to weight the forward numbers lightly until more trades accumulate.
Backtest and Validation
Over 451 days the backtest returned 20.95% (final equity $12,095.14), an 11.21% CAGR. The win rate of 66.67% across 37 trades is genuinely healthy, and fees were negligible at $37 total with no FX drag. For a strategy that trades infrequently, that cost discipline matters.
The risk side is less flattering. Sharpe of 0.61 is modest — this is not a smooth ride for the return it delivers. More pointedly, the 23.73% max drawdown slightly exceeds the total return, giving a Calmar ratio below 1. An investor would have needed to sit through a drawdown larger than the eventual profit. Turnover of 773% also signals the "tight 4-name book" rotates aggressively; the low absolute fee count softens that, but it's a lot of churn for the result.
The most important gap: validation is null. There is no out-of-sample or walk-forward check reported here. A 20.95% in-sample figure with a healthy win rate is exactly the profile that overfitting produces, and without holdout evidence we can't distinguish genuine edge from curve-fit.
Verdict
RSI Snap-Back is conceptually sound and operationally disciplined — simple rules, low fees, enforced position limits. But the numbers argue for caution: a drawdown that eats the return, a sub-1 Sharpe, high turnover, no validation, and a live start that's marginally red. Promising scaffolding; not yet a proven engine. Watch the live trades accumulate before raising conviction.