proven-logic retro: derivative classes + simple parlays at real prices
← lab · hot ref threshold 3.6 · ran 7/7/2026
What this is: A one-off study — its numbers below are exactly what the run reported.
| note | Real opening prices, 7 leagues x 21 seasons, per-season $1000 restarts. Cards/corners legs excluded (no historical odds exist) — their synthetic 5yr grading is soft_retro #185. Parlay legs assumed independent (different matches). |
| hot ref threshold | 3.6 |
classes
| ref draw | {"bets":2876,"seasons":21,"hit_rate":0.26,"roi_flat":-0.029, |
| unders edge | {"bets":2940,"seasons":7,"hit_rate":0.396,"roi_flat":-0.0528 |
| mid dog home | {"bets":3633,"seasons":21,"hit_rate":0.296,"roi_flat":-0.036 |
simple parlay
| bets | 1,709 |
| seasons | 21 |
| hit rate | 0.093 |
| roi flat | -0.1284 |
| worst season | 730.61 |
| winning seasons | 5 |
| median season end | 923.48 |
combined portfolio
| bets | 11,158 |
| seasons | 21 |
| hit rate | 0.282 |
| roi flat | -0.0531 |
| worst season | 67.08 |
| winning seasons | 7 |
| median season end | 589.17 |
control bet every edge
| bets | 37,395 |
| seasons | 21 |
| hit rate | 0.331 |
| roi flat | -0.0511 |
| worst season | 0.4 |
| winning seasons | 3 |
| median season end | 8.1 |
Reading the columnswhat each number actually means
| AUC | predictability: 0.50 = coin flip, ~0.70 = ceiling for sports |
| Importance | how much the model leans on this factor (permutation importance) |
| Direction | sign of the raw correlation with the outcome |
| Survives all eras | effect points the same way in every historical era |
Spec · the reproducible recipe
{
"kind": "proven_retro",
"mid_dog": [
2.6,
4.5
],
"run_ids": [
21,
22,
25,
26,
27,
28
],
"under_edge": 0.05
}