What the Strategy Does
RSI Snap-Back is a mean-reversion system built around one repeatable observation: large-cap tech names — AAPL, MSFT, NVDA, GOOGL, AMZN, META, and TSLA — tend to recover sharply after short-term selling extremes. The rules are intentionally simple: enter when RSI drops below 35, exit when it climbs back above 70, and never hold more than four names at once. That hard book limit is a structural risk control, not an afterthought — it caps concurrent drawdown exposure on what can be a volatile universe.
The strategy passed a code review on 2026-05-31 with a risk score of 0.35. The reviewer flagged three minor defensive-coding gaps but confirmed no look-ahead bias, sound position-sizing, and correct cash tracking — all the failure modes that disqualify a strategy before it trades a single dollar.
A Rocky Start Before Deployment
The timeline shows one earlier iteration that never made it out of the gate: a prior backtest run returned 0% with 0% max drawdown, a degenerate result that correctly triggered an automatic rejection. The version that shipped the same afternoon cleared that bar decisively. The quick turnaround suggests the logic was structurally sound but needed a data or parameter fix rather than a fundamental redesign.
Backtest Scorecard
| Metric | Value |
|---|---|
| Total Return | 20.95% |
| CAGR | 11.21% |
| Sharpe Ratio | 0.61 |
| Max Drawdown | −23.73% |
| Win Rate | 66.67% |
| Trades (451 days) | 37 |
| Turnover | 773% |
| Fees | $37 flat |
The numbers tell a mixed story. A 66.7% win rate and a positive return over 15 months are real signals. An 11.21% annualised return beats a basic cash benchmark and the fee drag is negligible at $37 total across 37 trades.
The friction points are harder to dismiss. A Sharpe of 0.61 is below the 1.0 threshold most systematic desks treat as the floor for live capital. A 23.73% max drawdown is aggressive for a strategy running on a $10,000 paper account — it means a drawdown period would erase nearly a quarter of equity before the system rotates out. And 773% annual turnover implies the strategy is cycling its entire book roughly eight times a year, which flatters the backtest in low-spread environments but could widen meaningfully under stress.
What's Missing: Validation
The validation field is null. There is no walk-forward test, no out-of-sample period, and no Monte Carlo stress run on record. The backtest runs on 451 days of presumably in-sample data, which is enough to fit RSI thresholds to a specific market regime but not enough to confirm they generalise. Given that 2024–2025 was broadly favourable for Mag-7 mean reversion, the strategy may be capturing a regime rather than a durable edge.
Early Live Signals: Nothing Yet
Two scheduled runs since deployment — the initial seeding and the first end-of-day cycle — both reported zero trades executed. That is not surprising: RSI < 35 on any Mag-7 name is a genuine extreme, and the market has to offer the signal before the strategy can act. A few quiet sessions at launch is normal operating behaviour, not a red flag.
Bottom Line
RSI Snap-Back is a clean, rule-based strategy with a coherent thesis and a passing grade from the automated reviewer. Its backtest return is real, but the drawdown depth and absent out-of-sample validation mean it earns a watchlist position rather than a conviction allocation. The next meaningful data point is its first live trade — and whether the entry timing looks as clean in real time as it did in the backtest.