Daily SPY (adj), signals lag-1 (PIT), IS<2016<=OOS, 10bps/switch. corr p = block-permutation (block 5). Placebo = random gate with matched exposure. Required per reports/placebo_audit.md.
| signal | target | n | corr | perm p |
|---|
| VIX3M/VIX | 1d | 4260 | -0.035 | 0.015 |
| VIX3M/VIX | 5d | 4260 | -0.067 | 0.011 |
| VIX9D/VIX | 1d | 3934 | +0.028 | 0.062 |
| VIX9D/VIX | 5d | 3934 | +0.037 | 0.138 |
| VVIX_z | 1d | 5034 | +0.013 | 0.295 |
| VVIX_z | 5d | 5034 | +0.014 | 0.576 |
| variant | scope | Sharpe | buy-hold | exposure | placebo p |
|---|
| long contango (ts>1) | IS | 0.52 | 0.46 | 48% | 0.698 |
| long contango (ts>1) | OOS | 0.87 | 0.88 | 92% | 0.005 |
| long backwardation (ts<1) | IS | 0.13 | 0.46 | 4% | 0.032 |
| long backwardation (ts<1) | OOS | 0.13 | 0.88 | 8% | 0.024 |
| term | metric | horizon | n | corr | perm p |
|---|
| 10-Year | bid_to_cover_z | 1d | 151 | -0.008 | 0.917 |
| 10-Year | bid_to_cover_z | 5d | 151 | +0.082 | 0.345 |
| 10-Year | bid_to_cover_z | 10d | 151 | +0.007 | 0.943 |
| 10-Year | indirect_share_z | 1d | 108 | -0.121 | 0.215 |
| 10-Year | indirect_share_z | 5d | 108 | -0.174 | 0.066 |
| 10-Year | indirect_share_z | 10d | 108 | -0.067 | 0.486 |
| 30-Year | bid_to_cover_z | 1d | 89 | +nan | nan |
| 30-Year | bid_to_cover_z | 5d | 89 | +nan | nan |
| 30-Year | bid_to_cover_z | 10d | 89 | +nan | nan |
| 30-Year | indirect_share_z | 1d | 85 | +nan | nan |
| 30-Year | indirect_share_z | 5d | 85 | +nan | nan |
| 30-Year | indirect_share_z | 10d | 85 | +nan | nan |
| POOLED | bid_to_cover_z | 1d | 1133 | +0.032 | 0.287 |
| POOLED | bid_to_cover_z | 5d | 1133 | +0.001 | 0.980 |
| POOLED | bid_to_cover_z | 10d | 1133 | +0.004 | 0.882 |
| POOLED | indirect_share_z | 1d | 914 | -0.033 | 0.321 |
| POOLED | indirect_share_z | 5d | 914 | -0.047 | 0.148 |
| POOLED | indirect_share_z | 10d | 914 | -0.024 | 0.454 |
| term | weak-quartile | strong-quartile | n |
|---|
| 2-Year | +0.303% | +0.319% | 321 |
| 5-Year | +0.403% | +0.658% | 300 |
| 7-Year | +0.331% | +0.147% | 207 |
| 10-Year | -0.064% | +0.309% | 137 |
| 30-Year | +0.532% | -1.031% | 89 |
| POOLED | +0.372% | +0.266% | 1054 |
- Adoptable only if an overlay beats buy-hold AND placebo p<0.05.
- Auction weak-quartile > strong-quartile supports the demand
mechanism; the reverse kills it.