The thesis
RSI Snap-Back is a straightforward mean-reversion play on the Magnificent Seven: AAPL, MSFT, NVDA, GOOGL, AMZN, META and TSLA. The idea is that large-cap tech snaps back sharply after short-term momentum extremes. The strategy buys the most oversold names as their RSI drops below 35, exits once RSI pushes above 70, and enforces a hard four-slot book so it can never hold more than four positions at once. That cap is the discipline mechanism — it limits how much concurrent drawdown exposure the strategy can take on, at the cost of occasionally sitting out opportunities when more than four names qualify.
Live activity: a strategy in waiting
The live record so far is notably quiet. The six most recent scheduled runs, spanning 14–20 October, each report the same line: "0 executed, 0 rejected," with the book flat at $10,000 cash and $10,000 total. In other words, RSI Snap-Back has not opened a single live position in this window. The mechanical read is that no Mag-7 name has been oversold enough to trip the RSI < 35 entry. That is the strategy behaving exactly as designed — it only acts on extremes — but it also means the live track record is effectively empty, and none of the backtest's edge has yet been demonstrated with real (paper) capital.
Backtest performance
On historical data the picture is more encouraging. Over 451 days the strategy returned 20.95%, lifting a $10,000 book to a final equity of $12,095, for an 11.21% CAGR. The win rate is a healthy 66.67% across 37 trades, and costs were minimal: $37 in total fees and zero FX cost, a reflection of the small, US-only universe. Even with turnover running at 773%, the fee drag stayed trivial.
Risks and the validation gap
The strengths come with real caveats. The Sharpe ratio of 0.61 is modest — the returns carry meaningful volatility relative to their size. The maximum drawdown of 23.73% is the sharper warning: a strategy that can shed nearly a quarter of its equity demands conviction to hold through. High turnover also means performance depends on costs staying low; in a higher-fee venue the edge would erode faster.
Most important, the validation field is null. There is no reported out-of-sample or holdout result, so the 20.95% figure is an in-sample backtest. For a mean-reversion rule tuned on exactly the names it trades, that is precisely where overfitting hides, and it is the single biggest open question here.
Bottom line
RSI Snap-Back is a clean, cheap, disciplined hypothesis with an attractive backtest and a high historical win rate. But it is currently unproven live — flat and fully in cash — and unvalidated out of sample. It earns a watch, not yet a verdict.