The Thesis
RSI Snap-Back is built on a simple but historically defensible premise: the largest names in tech get hit harder than their fundamentals warrant during short-term selling waves, then recover sharply. The strategy enters any Mag-7 name (AAPL, MSFT, NVDA, GOOGL, AMZN, META, TSLA) when its RSI drops below 35 — signalling a meaningful but likely temporary oversold condition — and exits when RSI exceeds 70. A hard cap of four concurrent positions keeps the book concentrated and prevents the strategy from spreading itself thin across correlated names during broad market stress.
The design philosophy is explicitly disciplined: if no name is oversold enough, the strategy holds cash. That patience is a feature.
Backtest Performance
Over 451 trading days, RSI Snap-Back compounded starting equity of $10,000 into $12,095 — a 20.95% total return and an annualised CAGR of 11.21%. The win rate of 66.7% across 37 trades is genuinely encouraging for a mean-reversion system; it suggests the RSI thresholds are doing real filtering work rather than picking up noise on both sides.
The Sharpe ratio of 0.61 is modest but not dismissible for an equity-only strategy with no leverage. It reflects the inherent volatility of holding concentrated positions in high-beta tech names.
The most important number to sit with is the 23.73% maximum drawdown. For a strategy running on Mag-7 names, that figure is plausible — these stocks can move violently in both directions — but it means a live account would have experienced a stomach-churning drawdown before recovering. Anyone deploying capital here needs to be comfortable holding through that kind of paper loss.
Turnover of 773% annually is high, which matters for real-money execution. The backtest accounts for $37 in total fees (flat $1 per trade), a reasonable paper-trading proxy, but slippage on fast-moving large-caps during high-RSI extremes could meaningfully erode edge in production.
Recent Activity: Disciplined Patience
The live run log from June 18–25 shows the same result every day: zero trades executed, zero rejected, full $10,000 in cash. Six consecutive scheduled runs without a single signal firing.
This is not a malfunction. It is the strategy doing exactly what it was designed to do: refusing to trade when the Mag-7 names aren't sufficiently oversold. Given that large-cap tech has broadly held up well in recent weeks, the RSI < 35 threshold simply hasn't been breached. The strategy is waiting for its pitch.
What's worth monitoring is whether this dry spell extends significantly further. A mean-reversion system that never finds an entry point either means the market is unusually calm — or that the entry threshold is set too conservatively for the current regime.
Risks and Gaps
Two structural risks stand out. First, validation data is absent ("validation": null). The backtest covers a 451-day window, but without walk-forward analysis or out-of-sample testing, it's impossible to know how much of the 66.7% win rate is genuine signal versus parameter fit. RSI thresholds are notoriously easy to overfit to historical data.
Second, the Mag-7 universe is highly correlated. When the sector sells off broadly, all seven names may cross RSI < 35 simultaneously, forcing the 4-slot book to pick among equally stressed positions and potentially concentrating exposure at exactly the wrong moment.
Bottom Line
RSI Snap-Back is a clean, legible strategy with a sensible thesis and encouraging backtest numbers. The current cash-holding stretch reflects discipline, not dysfunction. The priorities before treating this as signal-grade are clear: add walk-forward validation to stress-test the win rate, and model the correlated-selloff scenario explicitly. A Sharpe of 0.61 with proper out-of-sample confirmation would be a solid foundation — without it, the backtest remains a promising hypothesis.