The Thesis
RSI Snap-Back targets one of the most-watched corners of the market: the so-called Magnificent 7 (AAPL, MSFT, NVDA, GOOGL, AMZN, META, TSLA). The core premise is straightforward — large-cap tech names experience sharp short-term dislocations that tend to revert. The strategy enters when RSI drops below 35 (oversold) and exits when RSI climbs above 70 (overbought). A hard cap of four concurrent positions enforces book discipline and limits the strategy's exposure to simultaneous drawdowns across correlated names.
The logic is sound on first principles. Mag-7 stocks attract enormous liquidity, algorithmic attention, and institutional rebalancing flows — conditions that can amplify short-term momentum extremes while also providing reliable mean-reversion fuel once the dislocation exhausts itself.
Backtest Performance
Over 451 days, RSI Snap-Back returned 20.95% on a starting equity of $10,000, ending at $12,095. Annualised, that's a CAGR of 11.21% — a respectable number in isolation, though context matters enormously for any momentum-adjacent strategy.
The win rate of 66.67% across 37 trades is the headline strength: two out of every three exits were profitable. That kind of hit rate provides psychological and structural durability — a losing streak is less likely to blow through the position limit before the strategy recovers.
The concerns sit in the risk metrics. A Sharpe ratio of 0.61 is mediocre; the strategy is capturing return, but not efficiently relative to the volatility it absorbs. The max drawdown of 23.73% is the figure that demands attention — for a four-slot, large-cap-only book, a peak-to-trough loss approaching a quarter of the portfolio is steep. It suggests that when multiple Mag-7 names sell off simultaneously (a common occurrence — these stocks are highly correlated in risk-off environments), the strategy's drawdown protection is limited despite the slot cap.
Turnover of 773% across the period also implies meaningful transaction friction in any live deployment; the backtest records $37 in total fees (flat $1 per trade), which is benign at the paper-trading scale but worth monitoring as the book grows.
Recent Activity: A Strategy in Waiting
The live record tells a quiet story. From June 8 through June 15, RSI Snap-Back ran on schedule every business day and executed zero trades. Cash remains at the full $10,000 starting balance. No positions have been opened or closed in the observable window.
This is not necessarily a failure — it may be the strategy working exactly as designed. If Mag-7 RSI readings have remained in neutral territory (35–70), the model correctly passes rather than forces trades. Discipline in a rules-based system means accepting idle periods.
That said, a six-day consecutive no-op streak is worth flagging. Either the current market is too placid to generate RSI extremes in this universe, or the thresholds (sub-35 entry, above-70 exit) are calibrated tightly enough that the strategy will remain dormant for extended stretches — which in turn compresses the realised CAGR versus the backtest figure.
What to Watch
- Validation data is absent. The
validationfield is null, meaning the 20.95% return has only been confirmed on in-sample backtest data. An out-of-sample walk-forward or hold-out test is the next critical step before drawing strong conclusions. - Correlation risk. In a broad tech selloff, multiple Mag-7 names can breach RSI 35 simultaneously. The 4-slot book limits position count but not correlation — a coordinated drawdown remains the primary tail risk.
- Trigger frequency. If the current quiet period extends, it may warrant reviewing whether the RSI thresholds reflect realistic entry frequencies for this universe, or whether slight relaxation (e.g., RSI < 38) would improve capital utilisation without materially degrading the edge.
RSI Snap-Back has the bones of a credible mean-reversion strategy. The win rate is its calling card; the drawdown and Sharpe are the homework assignments still outstanding.