BPM estimates a player's net contribution in points per 100 possessions above a league-average player, derived entirely from box score statistics. Developed by Daniel Myers and published on Basketball-Reference, BPM decomposes into OBPM (offensive) and DBPM (defensive) components. A BPM of 0.0 represents exactly league average; +8.0 or above is historically MVP-caliber.
Available historically back to 1973-74, enabling long-run comparisons across eras. VORP is derived directly from BPM. Validated by Jewell et al. (JQAS) as the most robustly significant box score predictor of player value. Free and publicly available.
BPM captures only what the box score measures — it is fundamentally a compression of visible counting stats. It cannot capture off-ball screening value, defensive positioning, gravity effects, or any contribution that doesn't register as a countable event. DBPM is particularly unreliable: it uses assists as a defensive proxy and systematically penalizes defenders who take difficult assignments. BPM is also not adjusted for opponent quality.
BPM is scored in APEX's Overall Impact pillar at approximately 10% within-pillar weight (~3.3% of composite score). Its primary value in APEX is historical coverage — BPM is available back to 1973-74, enabling the full 40-season backtest that modern metrics like EPM cannot cover. Jewell et al. (JQAS) confirm BPM's primacy among box score metrics, validating its inclusion. It is intentionally weighted lowest within the Impact pillar because EPM, LEBRON, and DARKO DPM all use on/off data in addition to box score inputs, making them more methodologically complete. See also DBPM's role-assignment confound in Known Limitations.