The thesis
RSI Snap-Back is a mean-reversion play on the seven largest US tech names — AAPL, MSFT, NVDA, GOOGL, AMZN, META, and TSLA. The premise is that large-cap tech snaps back sharply after short-term momentum extremes: the strategy buys the most oversold names as RSI drops below 35 and exits into strength above 70. A hard four-slot book caps concurrent exposure and forces rotation, which is a sensible discipline against the classic mean-reversion failure mode of averaging into a name that keeps falling.
It is a coherent, well-scoped idea. The universe is liquid, the signals are transparent, and the position cap is an explicit risk control rather than an afterthought.
Backtest performance
Over 451 days, the backtest returns 20.95% (final equity $12,095 on $10k), an 11.21% CAGR, and a 66.67% win rate across 37 trades. Fees were a negligible $37 and FX cost was zero. On the surface, this is a solid mean-reversion profile: more winners than losers, low frictional drag.
Two numbers temper the enthusiasm. The Sharpe of 0.61 is modest — the returns came with meaningful volatility, not a smooth ride. And the max drawdown of 23.73% is steep for a strategy whose selling point is limiting concurrent drawdown exposure. A near-quarter peak-to-trough decline suggests the four-slot cap did not fully insulate the book when the whole cohort sold off together, which is exactly what correlated mega-cap tech tends to do. Turnover of 773% also signals a busy, rotation-heavy approach whose edge depends on those frictions staying small.
Recent activity
Here the picture gets quieter. The last six scheduled runs — from 4 through 11 September — each report 0 executed, 0 rejected, with cash and total equity pinned at $10,000. In other words, the live book is currently empty and idle. No name in the universe has crossed the RSI < 35 entry threshold recently, so the strategy is patiently waiting rather than forcing trades. That is arguably correct behaviour for a signal-gated system, but it means the impressive backtest figures are not yet being earned in live conditions.
The gap that matters
The single most important line in the data is validation: null. The headline metrics are backtest metrics — in-sample by construction. Without an out-of-sample or walk-forward validation record, we cannot distinguish a genuine edge from a curve fit to the 2025–2026 tech tape, and this platform has repeatedly flagged overfitting as the primary risk in any auto-deploy gate.
Verdict
RSI Snap-Back is a clean, disciplined idea with an encouraging but unvalidated backtest. The strengths — transparent signals, a hard position cap, tiny fees — are real. The risks are equally real: a mediocre Sharpe, a 24% drawdown that undercuts the discipline narrative, and, critically, no validation to prove the edge survives out of sample. For now it sits live but flat. The right next step is not more backtesting but a proper walk-forward validation before this strategy earns any meaningful capital.