Pre-reg: resid L126/H21 top-50 next-open, covered universe, IS<2016<=OOS, 10/25bps. Distinct from edge_hunt25's fwd-2/5/10d announcement-jump skip (REJECTED): this is the multi-week pre-earnings window (absent corporate bid) + the post-earnings resumption control. Random-skip placebo (100 draws) required.
| Variant | OOS Ret | OOS Sh | Net10 Sh | Turn/yr | CorrMkt | CorrBase | ResidSh | p vs BASE | IS Sh |
|---|---|---|---|---|---|---|---|---|---|
| BASE resid covered | +1215.66% | 1.21 | 1.21 | 8.13 | 0.79 | 1.00 | 0.67 | - | 0.56 |
| BB PRIMARY blackout[-2,+21] | +683.80% | 1.02 | 1.02 | 10.38 | 0.83 | 0.97 | 0.34 | 0.9998 | 0.56 |
| BB blackout[-2,+14] | +919.33% | 1.13 | 1.13 | 9.65 | 0.82 | 0.98 | 0.53 | 0.9826 | 0.56 |
| BB blackout[-2,+30] | +615.69% | 0.99 | 0.99 | 10.76 | 0.84 | 0.97 | 0.28 | 1.0000 | 0.56 |
| BB control resumption[-21,-2] | +958.90% | 1.14 | 1.14 | 9.27 | 0.81 | 0.99 | 0.56 | 0.9914 | 0.56 |
PRIMARY mean names-in-blackout/day (OOS): 15.16%. PRIMARY OOS Sharpe lift -0.18; random-skip placebo mean -0.11, p = 1.000.
Verdict
REJECT — primary lift -0.18 (need >=+0.10) AND placebo p 1.000 (need <0.05). A control lift comparable to PRIMARY means the effect is not the absent bid.
Time 200s, placebo draws 100.