🐻⬇️🏀

Model Card

RAPM on the per-segment margin of turnovers (opponent minus own, per 100 possessions), fit only on games whose play-by-play turnovers equal the box exactly. The ridge penalty is chosen by game-level cross-validation. Positive values mean a player's minutes raise the opponent's turnovers relative to their team's own. Held-out check (date split: fit before a date, score later segments; ncaam 2022-23 and 2023-24, ncaaw 2023-24): correlation with the actual per-segment margin 0.062-0.095 for this model, 0.051-0.083 for the Box Elo of the same family. Small in absolute terms: a segment holds about one such event.

Validation is computed for team models, APM, RAPM and the Box Elo forms. This page shows output coverage and site placement for this model.

Selected League / Season

Change league with the League selector at the top of the page.

Held-out player predictions

Plot Turnovers Lineup RAPM against any counting stat (one point per player; choose any season with a served list).

No current held-out results for this selection. Held-out player validation covers APM, RAPM and the Box Elo forms.

Lineup RAPM vs Box Elo, per season

Both forms of this family are shown side by side wherever a season publishes them; this table is information, not a switch. Each column is the held-out correlation of that form's predictions with the actual per-segment margin (fit before a date, scored on later segments); within two paired standard errors the two predict equally well. Correlations are small in absolute terms: a segment holds about one such event. Held-out correlation does not penalize a Box form for not telling teammates apart; the season links show each form's team share.

Output Coverage

Rows in each model's current published run (or legacy rows where no run is selected).