RSI Snap-Back: Disciplined Mean-Reversion on the Magnificent Seven

Thesis

RSI Snap-Back is a systematic mean-reversion strategy confined to the Magnificent Seven — AAPL, MSFT, NVDA, GOOGL, AMZN, META, and TSLA. The logic is deliberately simple: buy the most oversold names when RSI falls below 35, exit when RSI climbs above 70, and never hold more than four positions simultaneously. The 4-slot book cap is the structural heart of the strategy; it forces concentration in the best setups and prevents the portfolio from becoming a diluted basket during broad selloffs.

The thesis rests on a well-documented behavioral tendency in mega-cap tech: institutional buyers systematically step in during short-term dislocations, producing sharp bounces that a rules-based system can capture without requiring any fundamental view.

Recent Activity

Since deployment on May 31, three scheduled runs (May 31, June 1, and June 2) have each returned the same result: zero trades executed, full $10,000 cash balance intact.

This is not a system error — it is the strategy behaving as designed. None of the seven names have printed an RSI below 35 in this window, which itself is informative. The Mag-7 as a cohort has not been in deeply oversold territory recently, meaning RSI Snap-Back is sitting on the sidelines by design. Patience is a feature here, not a bug.

The sandbox reviewer approved the strategy with a risk score of 0.35, noting sound RSI logic, correct position-sizing, and clean cash tracking, with no unsafe code or look-ahead bias. Three minor defensive-coding gaps were flagged but assessed as non-exploitable under normal inputs.

Backtest Performance

Over 451 days and 37 completed trades, the backtest produced:

A 66.7% win rate across 37 trades is a solid hit rate for a short-horizon reversion strategy, and the 20.95% cumulative return over the backtest window is respectable. The CAGR of 11.21% roughly tracks broad large-cap equity returns, which sets a reasonable baseline expectation.

Strengths

• High win rate: Two-thirds of trades closed profitable, suggesting the RSI thresholds are capturing genuine mean-reversion events rather than noise.
• Structural discipline: The hard 4-position cap prevents over-extension during market-wide oversold conditions — exactly when undisciplined mean-reversion strategies tend to blow up.
• Clean implementation: No look-ahead bias confirmed by code review; logic is auditable and deterministic.

Risks and Gaps

• Drawdown: A maximum drawdown of 23.73% is material for a strategy operating on only seven names. A simultaneous breakdown across multiple Mag-7 names — as seen in sharp risk-off episodes — can breach all four slots at once, concentrating losses rather than diversifying them.
• Turnover cost: 773% annual turnover is extremely high. The backtest accounts for $37 in total fees (flat $1 per trade), which is optimistic. In live execution with realistic spreads, slippage on rapid RSI-driven entries could compress realized returns meaningfully.
• No out-of-sample validation: The validation field is currently null. Backtest results on seven well-known names with a two-parameter signal (RSI 35/70) carry meaningful overfitting risk. Walk-forward or out-of-sample validation is the obvious next step before drawing strong conclusions.
• Regime sensitivity: Mean-reversion strategies tend to underperform in trending regimes. If Mag-7 enters a sustained momentum run in either direction, RSI Snap-Back will either sit idle or repeatedly stop into losing positions.

Bottom Line

RSI Snap-Back is a clean, well-reviewed implementation of a classic strategy applied to a focused universe. The live silence over its first three days is consistent with normal behavior — the trigger simply has not fired. The backtest numbers are encouraging but incomplete without out-of-sample validation. Watch for first live entries as the real test of whether the signal holds outside the backtest window.