Research
Understand the math before you size up. 286 aligned sessions · sample prices, not a live signal
Hedge ratio (β)
-0.4721
OLS on log prices
Half-life
23.4 days
AR(1) on spread
Backtest Sharpe
-1.76
Rough, in-sample
Profit factor
0.60
5 simulated trades
Spread z-score
How this works
- Align QQQ and SPY daily adjusted closes.
- Regress log(QQQ) on log(SPY) → hedge ratio β.
- Spread = residual from that regression (should be stationary if pair is cointegrated).
- Z-score over 60 days. Enter at ±2, exit at 0, stop at ±3.
- Regime filter: half-life between 2–60 days. If spread drifts too long, the “edge” is probably gone.
Simulated trades (last sample)
| Entry | Exit | Dir | P&L% |
|---|---|---|---|
| 2026-09-16 | 2026-11-06 | long spread | 1.02% |
| 2026-06-26 | 2026-08-28 | short spread | -1.93% |
| 2026-04-20 | 2026-05-21 | long spread | 1.77% |
| 2026-04-16 | 2026-04-17 | long spread | -0.92% |
| 2026-04-03 | 2026-04-15 | long spread | -1.80% |