sweep: match.result D
← lab · AUC 0.530 (no real pattern — honest result) · ran 7/4/2026
What this is: Asks which pre-match factors drive one specific outcome, using a walk-forward model and permutation importance.
| Factor | Importance | Direction | Survives all eras? |
|---|
| diff__team.season_ppg | 0.0049 | ↓ -0.034 | ⏳ yes |
| market.p_over25 | 0.0046 | ↓ -0.085 | ⏳ yes |
| away__team.corner_diff_avg_l5 | 0.0024 | ↑ +0.004 | no |
| ratio__team.goals_volatility_l10 | 0.0020 | ↓ -0.014 | no |
| away__t3__team.defensive_leak_l5 | 0.0019 | ↓ -0.002 | no |
| home__team.season_ppg | 0.0019 | ↓ -0.041 | ⏳ yes |
| home__t3__team.state_index | 0.0017 | ↑ +0.009 | no |
| away__t3__team.corners_volatility_l10 | 0.0015 | ↓ -0.002 | no |
| home__team.dominance_l5 | 0.0014 | ↓ -0.034 | ⏳ yes |
| market.ah_line | 0.0014 | ↑ +0.043 | ⏳ yes |
| ratio__team.comeback_rate_l20 | 0.0014 | ↓ -0.015 | no |
| ratio__team.season_ppg | 0.0014 | ↓ -0.029 | ⏳ yes |
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
{
"name": "sweep: match.result D",
"sport": "football",
"target": {
"equals": "D",
"metric": "match.result"
},
"features": "all"
}