Cleveland -18.5 WIN: Model correctly identified Washington as league-worst team (-12.4 BPI) and properly weighted Cleveland's home court advantage against inferior competition
MEM @ DEN UNDER 240.5 WIN: Injury-depleted Memphis (missing Morant, Bane, Smart) combined with Denver's defensive capabilities at altitude produced the expected lower-scoring affair
WHAT MISSED
Golden State -2.5 LOSS: Massively overweighted Wembanyama's absence (+5 points), failing to account for Spurs' depth and Warriors' home struggles. 13-point loss suggests injury impact calculation was fundamentally flawed
Portland @ Minnesota UNDER 236.5 LOSS: Model underestimated pace factors - Minnesota's recent offensive surge and Portland's defensive vulnerabilities led to 242 total, 5.5 points over projection
MODEL CALIBRATION NOTES
Injury weighting severely miscalibrated: The 7-point swing for Wembanyama + Paul was excessive, especially against a Warriors team that's been inconsistent at home
Home court bias evident: Consistently favored home teams (GSW, DEN) without adequately weighing recent form and matchup-specific factors
METHODOLOGY ADJUSTMENTS TO CONSIDER
Cap single-player injury impact at 3-4 points maximum: Even MVP-level players shouldn't create 5+ point swings in modern NBA depth-driven basketball
Incorporate recent form metrics: Add 10-game rolling performance data to counter stale season-long BPI ratings that don't reflect current team trajectory