sweep: match.total_corners >= 10

← lab · AUC 0.524 (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?
market.p_over25	0.0059	↑ +0.084	⏳ yes
home__team.corners_against_avg_l5	0.0044	↑ +0.035	⏳ yes
match.ref_fouls_avg	0.0043	↓ -0.003	no
away__team.goals_volatility_l10	0.0034	↑ +0.005	no
match.state_gap	0.0033	↑ +0.021	⏳ yes
away__team.state_index	0.0032	↓ -0.014	no
home__team.dominance_l5	0.0030	↑ +0.004	no
away__team.cards_avg_l5	0.0027	↓ -0.025	no
away__team.elo	0.0027	↓ -0.021	no
home__team.ht_lead_rate_l20	0.0024	↑ +0.017	no
match.humidity_pct	0.0020	↑ +0.007	no
home__team.goals_volatility_l10	0.0019	↑ +0.018	no

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.total_corners >= 10",
  "sport": "football",
  "target": {
    "op": ">=",
    "value": 10,
    "metric": "match.total_corners"
  },
  "features": "all"
}