Pre-reg: PRIMARY = keep gate60 long only if VIX[D-1] <= trailing-252d median (pead_v2 live-gate definition, lag-1). IS < 2025-01-01, OOS >= 2025-01-01. VIX ffilled to UTC-daily grid (Friday covers weekends; weekend shocks unseen — stated limitation). Non-OK days BLOCKED, counted separately. Credit leg deferred to F4b (data incident).
Window: 2022-12-02 .. 2026-08-31 (1369 days). Status mix: OK 1365, WARMING 0, NO_DATA 0, STALE 4
PRIMARY diag (full window): kept 263/364 longs; stress skips 97; blocked 4 days. Strict-p75 diag: kept 330/364 longs; stress skips 30; blocked 4 days. Spike diag: kept 326/364 longs; stress skips 34; blocked 4 days.
| Segment | Variant | LongDays | Ret | Sharpe | Net10 Sharpe | Turnover |
|---|---|---|---|---|---|---|
| IS | baseline (gate60) | 256 | +86.70% | +1.03 | +1.00 | 19 flips |
| IS | VIX<=252d median (PRIMARY) | 182 | +98.91% | +1.29 | +1.24 | 30 flips |
| IS | VIX<=252d p75 (strict) | 230 | +69.08% | +0.92 | +0.89 | 25 flips |
| IS | VIX z30<=1.5 (spike) | 230 | +67.28% | +0.91 | +0.85 | 41 flips |
| OOS | baseline (gate60) | 108 | +65.18% | +1.11 | +1.08 | 15 flips |
| OOS | VIX<=252d median (PRIMARY) | 81 | +57.81% | +1.11 | +1.08 | 16 flips |
| OOS | VIX<=252d p75 (strict) | 100 | +59.90% | +1.07 | +1.04 | 16 flips |
| OOS | VIX z30<=1.5 (spike) | 96 | +49.51% | +0.94 | +0.89 | 22 flips |
| FULL | baseline (gate60) | 364 | +208.40% | +1.06 | +1.03 | 33 flips |
| FULL | VIX<=252d median (PRIMARY) | 263 | +213.90% | +1.21 | +1.17 | 46 flips |
| FULL | VIX<=252d p75 (strict) | 330 | +170.35% | +0.98 | +0.95 | 40 flips |
| FULL | VIX z30<=1.5 (spike) | 326 | +150.10% | +0.92 | +0.87 | 62 flips |
OOS verdict (PRIMARY, gross)
- lift +0.00, bootstrap p 0.637
- beta vs BH: -0.04 | corr: -0.20 | resid Sharpe -0.25 (ret -7.18%, maxDD -15.01%)
- OOS blocked-day fraction 0.7%
Verdict: REJECT (lift +0.00 < +0.10; p 0.637 NOT < 0.05; resid Sharpe -0.25 <= 0; blocked 0.7% < 50%)
Forward returns of OK gate60-long days by vix_ratio quartile
| Quartile (vix_ratio) | Mean daily ret | Days |
|---|---|---|
| Q1(low) | +0.177% | 90 |
| Q2 | +0.879% | 90 |
| Q3 | +0.282% | 90 |
| Q4(high) | +0.167% | 90 |
Reading
- PASS needs lift ≥ +0.10 with p < 0.05 on OOS, positive residual, and <50% blocked days. A pure drawdown cut at matched Sharpe is a REJECT here (sizing-study trade, not a skip-filter). IS shown for consistency.
- Net@10/25bps alongside gross; weekend gate state = Friday's regime.
- Next on PASS: walk-forward + shadow overlay; on REJECT: F4b credit leg after the data incident closes, then F5 news/LLM shrink-only.
Time: 2s. Gate: VIX<=252d median (min_periods=180), spike z(30d)>1.5.