NOP +4.5 hit decisively (covered by 19.5), validating the injury impact assessment on Philadelphia without Embiid/PG
SAC +18.5 barely covered (by 1.5), showing the model correctly identified market overreaction to Sacramento's injuries
HOU @ NYK UNDER 218.5 hit cleanly, demonstrating solid total evaluation on back-to-back scheduling impacts
WHAT MISSED
CHI +10.5 (HIGH confidence) lost badly - model severely underestimated Detroit's ability to perform on B2B, missing by 5.5 points
SAC @ SAS UNDER 230.5 blown out by 30.5 points - completely whiffed on pace/scoring despite injury considerations
The back-to-back adjustment for San Antonio (+2.5 points) was wholly inadequate given they won by 17
MODEL CALIBRATION NOTES
Confidence ratings were backwards - the HIGH confidence CHI pick lost while MEDIUM confidence SAC spread hit
Systematic overweighting of rest advantage - both Detroit (B2B) and San Antonio (B2B) performed much better than model predicted
Total predictions inconsistent - nailed the low-scoring HOU/NYK game but badly missed two overs
METHODOLOGY ADJUSTMENTS TO CONSIDER
Reduce back-to-back penalties from current 3.0 points to 1.5-2.0 points - elite teams (Detroit/San Antonio) are less affected by fatigue
Separate injury impact by team quality - strong teams (Spurs) may absorb opponent injuries better and actually increase pace/scoring rather than decrease totals
Record: 3-4 spreads/totals, highlighting calibration issues with situational adjustments.