premium = Coinbase/Binance - 1 | z(30) | IS<2025-01-01/OOS>= | close-to-close forward day | 15bps/flip | 30d block placebo x2000.
| leg | exposure | IS net SR | OOS net SR | OOS B&H | net DD | OOS IC | flips |
|---|
| z>+0.5 (PRIMARY) | 31% | 1.04 | -0.08 | 0.58 | -54% | +0.073 | 549 |
| z>0.0 | 49% | 0.90 | -0.04 | 0.58 | -56% | +0.073 | 583 |
| z>+1.0 | 16% | 0.83 | -0.31 | 0.58 | -31% | +0.073 | 384 |
| z<-0.5 (reverse) | 33% | -0.71 | -1.66 | 0.58 | -95% | +0.073 | 598 |
- block placebo z>+0.5: OOS obs -0.08 vs null -0.62/p95 0.42 -> p 0.202
- partial IC(fwd1 | tr5,tr21) = +0.0755; partial IC(fwd5) = +0.0646
- DSR(K=4) OOS net vs B&H 0.58: 0.13 | haircut -0.08
| z quintile | mean fwd1 | n |
|---|
| Q1 | -0.187% | 484 |
| Q2 | +0.002% | 484 |
| Q3 | -0.038% | 484 |
| Q4 | +0.345% | 484 |
| Q5 | +0.329% | 484 |
| leg | exposure | IS net SR | OOS net SR | OOS B&H | net DD | OOS IC | flips |
|---|
| z>+0.5 (PRIMARY) | 31% | 0.90 | -0.48 | 0.53 | -62% | +0.058 | 519 |
| z>0.0 | 49% | 1.04 | -0.08 | 0.53 | -63% | +0.058 | 637 |
| z>+1.0 | 16% | 0.91 | -1.84 | 0.53 | -65% | +0.058 | 396 |
| z<-0.5 (reverse) | 33% | -1.04 | -0.55 | 0.53 | -99% | +0.058 | 634 |
- BTC PRIMARY: OOS net-vs-BH -0.66, placebo p 0.202, IS edge +0.46, partial IC +0.075
- REJECT -> the premium relation does not survive the placebo / control bars on this sample; feed stays research infrastructure. PASS would graduate to shadow only, no live wiring.