The Thesis
RSI Snap-Back is a mean-reversion play on the Magnificent Seven — AAPL, MSFT, NVDA, GOOGL, AMZN, META, and TSLA. The premise is that large-cap tech overshoots on short-term momentum and reverts: 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. That 4-slot cap is the strategy's core discipline, deliberately limiting how much concurrent drawdown exposure it can carry at any one time.
It's a clean, well-understood edge. Reversion in mega-cap equities is a documented behaviour, and constraining the book to four positions forces the strategy to hold conviction rather than sprawl.
Backtest Performance
Over 451 days, the strategy returned 20.95%, taking $10,000 to a final equity of $12,095. That's a CAGR of 11.21% across 37 trades, with a healthy 66.67% win rate — two winners for every loser. Trading costs were negligible: $37 in fees and zero FX cost, reflecting a USD-only universe.
The execution profile is lean. Just 37 trades over 451 days is patient by mean-reversion standards, though turnover clocks in at a brisk 773% — the four slots rotate hard when signals fire.
The Risks Worth Naming
Three things temper the headline return.
First, the drawdown. A 23.73% maximum drawdown is steep for a 20.95% total return — the strategy asked investors to stomach nearly as much pain as it delivered gain. The Sharpe ratio of 0.61 confirms the return came with meaningful volatility rather than smooth compounding.
Second, and most importantly, validation is null. There is no out-of-sample or walk-forward result recorded here. A 66.67% win rate and a 20.95% return look attractive, but without validation on unseen data we cannot rule out that the RSI thresholds (35/70) and the 4-slot cap were tuned to fit the backtest window. Every metric above should be read as in-sample until that gap is closed.
Third, the strategy is idle. The six most recent scheduled runs — spanning 2026-08-28 through 2026-09-04 — each report 0 executed, 0 rejected, with cash and total both flat at $10,000. RSI Snap-Back has been fully in cash for a week, with no Mag-7 name oversold enough to trigger an entry and nothing on the book to exit. There are no recent live trades to evaluate.
The Verdict
Being in cash is not a bug — a reversion strategy that only acts on extremes should wait, and holding fire during a calm, non-oversold tape is exactly the discipline the design intends. But it does mean the live track record is still thin: the numbers we have are backtest numbers, and the live book has yet to be truly tested.
RSI Snap-Back is a coherent, low-cost, disciplined idea with a credible in-sample edge. Before treating that edge as real, though, it needs two things this update can't yet show: an out-of-sample validation pass to guard against overfitting, and a stretch of live trades to prove the snap-back actually snaps back when real capital is on the line.