Evidence
Across 75 historical occurrences, the QQQ / SMH / SOXX basket returned on average 14.9% over the next 63 trading days (median 11.1%), against SPY — an excess of 7.6%. The forward window was positive 82.7% of the time.
Interpretation
On this evidence the setup historically added excess return versus SPY at the 63-day horizon.
Metrics by holding period
| Holding (days) | n | Avg | Median | Win rate | Excess | Worst |
|---|---|---|---|---|---|---|
| 21 | 75 | 2.5% | 1.4% | 56.0% | 2.1% | -20.8% |
| 63 | 75 | 14.9% | 11.1% | 82.7% | 7.6% | -28.6% |
| 126 | 72 | 22.9% | 23.8% | 86.1% | 9.5% | -32.2% |
Validation (63-day) — FAIL
| Stage | Result |
|---|---|
| training | ✅ passed |
| validation | ❌ n=18 < 20 |
| oos | ❌ n=9 < 20 |
| random-entry null | ✅ beats random entry · 95% CI [4.9%, 11.5%] |
| ±15% robustness | ✅ holds · edge holds same-sign under all ±15% parameter shifts |
Considerations
Ways investors commonly express a view like this include broad sector/theme ETFs (e.g. QQQ). Position sizing, entry timing and risk control remain the investor's own decisions.
Reality check
n=75. A historical observation, not a prediction or recommendation; regimes change and past behaviour need not repeat.