Schemer Soft launch · daily batch

Method

Box Plus/Minus 2.0

Method
bpm2-v0
Version
bpm2-v0
As of
2026-10-08
Family
Value metrics
Grain
player x season (and player x team stint)

Cohort rule

z against the season's rotation cohort (rel-v0 is_rotation: 15+ min/g, half-of-max games rule); rendered in the 2023-24 rotation mean and spread (the era-v0 transform)

Stats this method computes

KeyNameDefinitionUnitTransform
bpmBox plus/minus (BPM)Box Plus/Minus 2.0: points per 100 possessions above a league-average player, estimated from the box score and adjusted so a team's players sum to its opponent-adjusted net rating. 0 is league average.points per 100 possessions vs average
obpmOffensive box plus/minus (OBPM)The offensive part of BPM 2.0, from its own regression and adjusted to the team's opponent-adjusted offensive rating.points per 100 possessions vs average
dbpmDefensive box plus/minus (DBPM)The defensive part of BPM 2.0: BPM minus OBPM.points per 100 possessions vs average
bpm_positionBPM position estimate (BPM Pos)The position BPM 2.0 estimates from the box (1 point guard to 5 center): rebounding, steals, fouls, assists and blocks shares, blended with a prior and centered on 3 per team. Used to interpolate the BPM coefficients.position (1-5)
bpm_roleBPM offensive role estimate (BPM Role)The offensive role BPM 2.0 estimates from assist and threshold-scoring shares (1 creator to 5 receiver), blended with a prior and centered on 3 per team.role (1-5)
bpm_eraBox plus/minus, era-normalized (BPM era)BPM rendered in the 2023-24 league: the season's z-score against its rotation cohort, times the 2023-24 spread, plus the 2023-24 mean (the era-v0 rung-3 transform). Ranks within a season never move; talent pool and dilution are not adjusted (rung 4, deferred).points per 100 possessions (2023-24 scale)era
post_bpmBox plus/minus, postseason (BPM post)Box Plus/Minus 2.0: points per 100 possessions above a league-average player, estimated from the box score and adjusted so a team's players sum to its opponent-adjusted net rating. 0 is league average. Postseason scope: bpm2-v0p1, the same method over the season's postseason games only with its opponent adjustment solved directly (team sums, possessions, the opponent solve and league constants from those games; VORP prorates by postseason team games over 82). Play-in rows carry none; no era variant.points per 100 possessions vs average
post_obpmOffensive box plus/minus, postseason (OBPM post)The offensive part of BPM 2.0, from its own regression and adjusted to the team's opponent-adjusted offensive rating. Postseason scope: bpm2-v0p1, the same method over the season's postseason games only with its opponent adjustment solved directly (team sums, possessions, the opponent solve and league constants from those games; VORP prorates by postseason team games over 82). Play-in rows carry none; no era variant.points per 100 possessions vs average
post_dbpmDefensive box plus/minus, postseason (DBPM post)The defensive part of BPM 2.0: BPM minus OBPM. Postseason scope: bpm2-v0p1, the same method over the season's postseason games only with its opponent adjustment solved directly (team sums, possessions, the opponent solve and league constants from those games; VORP prorates by postseason team games over 82). Play-in rows carry none; no era variant.points per 100 possessions vs average
post_bpm_positionBPM position estimate, postseason (BPM Pos post)The position BPM 2.0 estimates from the box (1 point guard to 5 center): rebounding, steals, fouls, assists and blocks shares, blended with a prior and centered on 3 per team. Used to interpolate the BPM coefficients. Postseason scope: bpm2-v0p1, the same method over the season's postseason games only with its opponent adjustment solved directly (team sums, possessions, the opponent solve and league constants from those games; VORP prorates by postseason team games over 82). Play-in rows carry none; no era variant.position (1-5)
post_bpm_roleBPM offensive role estimate, postseason (BPM Role post)The offensive role BPM 2.0 estimates from assist and threshold-scoring shares (1 creator to 5 receiver), blended with a prior and centered on 3 per team. Postseason scope: bpm2-v0p1, the same method over the season's postseason games only with its opponent adjustment solved directly (team sums, possessions, the opponent solve and league constants from those games; VORP prorates by postseason team games over 82). Play-in rows carry none; no era variant.role (1-5)

This page carries the definitions only. The full method note, its coverage and its calibration record publish with the data pages. Every value a data page shows will name this method and its version. All methods