Pre-reg: split IS<2024-01-01/OOS>=2024-01-01 (feed starts 2021-06 — NOT the house split; both sides ~45 formations, power caveat). Resid L126/H21 top-50 H21 next-open. PIT via availableDate<=d-1d. Uncovered names never skip. Costs 10bps/25bps-stress. sqrt252.
Feed: 122 settlements 2021-06-15..2026-08-14; mapped 499/501 names; OOS day coverage 91.0%.
Scoreboard (OOS)
| Variant | Ret | Sharpe | MaxDD | Net10 | Net25 | Turn/yr | CorrMkt | ResidSh | p vs BASE |
|---|---|---|---|---|---|---|---|---|---|
| BASE resid L126/H21 | +150.17% | 1.51 | -25.36% | 1.51 | 1.46 | 8.12 | 0.66 | 0.78 | - |
| S PRIMARY skip dDTC>=+1.0 | +147.86% | 1.52 | -24.94% | 1.52 | 1.46 | 8.55 | 0.67 | 0.79 | 0.6162 |
| S strict dDTC>=+2.0 | +149.40% | 1.50 | -25.52% | 1.51 | 1.45 | 8.21 | 0.66 | 0.78 | 0.5440 |
| S control skip dDTC<=-1.0 | +143.41% | 1.48 | -25.29% | 1.48 | 1.43 | 8.49 | 0.67 | 0.74 | 0.7742 |
| S velocity rel>=+50% | +144.55% | 1.51 | -25.15% | 1.51 | 1.45 | 8.47 | 0.68 | 0.77 | 0.7364 |
- S PRIMARY skip dDTC>=+1.0: skip-flag 14.26% of name-days; nonzero days 655.
- S strict dDTC>=+2.0: skip-flag 3.56% of name-days; nonzero days 655.
- S control skip dDTC<=-1.0: skip-flag 12.35% of name-days; nonzero days 655.
- S velocity rel>=+50%: skip-flag 9.91% of name-days; nonzero days 655.
BASE 21d-forward by held-names mean dDTC quartile (OOS)
| Quartile | Mean fwd21 | N |
|---|---|---|
| Q1(cooling) | +8.244% | 6 |
| Q2 | +3.173% | 6 |
| Q3 | -3.160% | 6 |
| Q4(heating) | +5.260% | 6 |
IS reference
| Variant | IS Sharpe | IS lift |
|---|---|---|
| BASE resid L126/H21 | 0.77 | - |
| S PRIMARY skip dDTC>=+1.0 | 0.76 | -0.01 |
| S strict dDTC>=+2.0 | 0.76 | -0.00 |
| S control skip dDTC<=-1.0 | 0.76 | -0.01 |
| S velocity rel>=+50% | 0.76 | -0.01 |
PRIMARY verdict: REJECT (gross lift +0.01 / net10 +0.01 (need >=+0.10); p 0.6162 (need <0.05); resid +0.79 (need >0); IS lift -0.01 (need same sign); control lift -0.03 (need < PRIMARY); DSR(K=4) 0.99)
Time 28s.