relational: does the better-X side get more Y? (football)

← lab · 148 side-vs-side questions tested · ran 7/4/2026

What this is: Side-vs-side questions: does the team better at X actually get more Y? Tested for every factor × every countable outcome, conditioned on how big the X advantage was.

"Does the side better at X get more Y?" — every factor (including invented composites) against every countable outcome, split by how big the X gap was. 0.5 = coin flip; the by-gap column shows the probability when the gap is LARGE (top tercile).

Better at	Gets more	P(better side)	…when gap is big	n	Stable
team.elo	goals	0.6828	0.8325	39,486	⏳ yes
team.state_index	goals	0.6271	0.7344	19,898	⏳ yes
team.season_ppg	goals	0.6233	0.7186	18,680	⏳ yes
team.venue_ppg_l5	goals	0.62	0.7141	17,457	⏳ yes
team.shot_diff_avg_l5	goals	0.6111	0.7171	18,801	⏳ yes
team.dominance_l5	goals	0.6093	0.7144	19,055	⏳ yes
team.goal_diff_avg_l5	goals	0.6041	0.6962	18,238	⏳ yes
team.ht_lead_rate_l20	goals	0.6058	0.7156	16,029	⏳ yes
team.form_points_l5	goals	0.5981	0.6795	17,797	⏳ yes
team.corner_diff_avg_l5	goals	0.5955	0.6715	18,594	⏳ yes
team.shots_for_avg_l5	goals	0.5942	0.6893	18,597	⏳ yes
team.h2h_win_rate_l5	goals	0.6214	—	9,416	⏳ yes
team.goals_for_avg_l5	goals	0.5867	0.6663	17,486	⏳ yes
team.h2_goals_l10	goals	0.5869	0.6638	17,202	⏳ yes
team.corners_against_avg_l5	goals	0.4161	0.3555	18,326	⏳ yes
team.elo	shots_on_target	0.6912	0.807	3,494	no
team.goals_against_avg_l5	goals	0.416	0.3531	17,459	⏳ yes
team.state_index	shots_on_target	0.6749	0.7865	3,494	no
team.corners_for_avg_l5	goals	0.5742	0.644	18,306	⏳ yes
team.season_ppg	shots_on_target	0.6675	0.7821	3,200	no
team.dominance_l5	shots_on_target	0.6602	0.7697	3,408	no
team.shot_diff_avg_l5	shots_on_target	0.6609	0.7582	3,377	no
team.skill_v1	shots_on_target	0.6587	0.7879	3,410	⏳ yes
team.elo	corners	0.6537	0.7504	3,607	no
team.ht_lead_rate_l20	shots_on_target	0.6639	0.775	3,044	no
team.state_index	corners	0.6501	0.7631	3,607	no
team.goals_volatility_l10	goals	0.5648	0.6081	18,913	⏳ yes
team.sot_for_avg_l5	shots_on_target	0.6517	0.7541	3,290	⏳ yes
team.possession_avg_l5	shots_on_target	0.6466	0.7799	3,384	⏳ yes
team.pass_acc_avg_l5	shots_on_target	0.6462	0.7511	3,375	⏳ yes
team.skill_v1	goals	0.6561	0.7858	2,905	⏳ yes
team.skill_v1	corners	0.6407	0.769	3,471	⏳ yes
team.season_ppg	corners	0.6439	0.7533	3,286	no
team.goal_diff_avg_l5	shots_on_target	0.6443	0.7475	3,216	no
team.venue_ppg_l5	shots_on_target	0.6451	0.7553	3,150	no
team.venue_ppg_l5	corners	0.6435	0.7536	3,189	no
team.shots_for_avg_l5	shots_on_target	0.6393	0.7323	3,355	no
team.matches_since_win	goals	0.4358	—	15,640	⏳ yes
team.sot_for_avg_l5	goals	0.6468	0.7516	2,800	⏳ yes
team.sot_against_avg_l5	shots_on_target	0.3663	0.2654	3,265	⏳ yes
team.pass_acc_avg_l5	corners	0.6299	0.7526	3,434	⏳ yes
team.pass_acc_avg_l5	goals	0.6419	0.7492	2,871	⏳ yes
team.possession_avg_l5	corners	0.6296	0.7571	3,445	⏳ yes
team.goal_diff_avg_l5	corners	0.6321	0.7122	3,286	no
team.h2h_win_rate_l5	shots_on_target	0.6567	—	2,298	⏳ yes
team.possession_avg_l5	goals	0.6396	0.7678	2,883	⏳ yes
team.dominance_l5	corners	0.6268	0.719	3,465	no
team.shot_diff_avg_l5	corners	0.6273	0.6983	3,434	no
team.matches_since_blank	goals	0.5574	0.6323	16,709	⏳ yes
team.form_points_l5	shots_on_target	0.6311	0.7474	3,128	no
team.corner_diff_avg_l5	shots_on_target	0.6252	0.7224	3,322	no
team.ht_lead_rate_l20	corners	0.6301	0.7277	3,079	no
team.goals_for_avg_l5	shots_on_target	0.6292	0.7345	3,107	no
team.h2_goals_l10	corners	0.6244	0.6979	3,187	no
team.corners_against_avg_l5	shots_on_target	0.3783	0.3038	3,291	no
team.form_points_l5	corners	0.6223	0.7258	3,185	no
team.goals_for_avg_l5	corners	0.6227	0.7082	3,162	no
team.sot_against_avg_l5	goals	0.3692	0.2797	2,779	⏳ yes
team.sot_for_avg_l5	corners	0.6186	0.7037	3,346	⏳ yes
team.h2_goals_l10	shots_on_target	0.6212	0.7179	3,147	no

Reading the columnswhat each number actually means

P(better side)	probability the side better at X ends up with more Y; 0.5 = no relationship
…when gap is big	same probability in the top tercile of X gaps — monotone rise = real leverage
Stable	same direction in both time halves of the data

Spec · the reproducible recipe

{
  "kind": "relational",
  "sport": "football"
}