Winning Metrics > Regression Points per Possession
Strong evidence exists that teams play slightly worse as score margin increases, and teams play slightly better as score margin decreases. A team leading by 10 points will play slightly worse than the same team leading by 20 points. Regression Points per Possession accounts for this, allowing us to forecast counterintuitive and non-competitive results that can occur in a game.
Strong evidence exists that teams play slightly worse as score margin increases, and teams play slightly better as score margin decreases. A team leading by 10 points will play slightly worse than the same team leading by 20 points. Regression Points per Possession accounts for this, allowing us to forecast counterintuitive and non-competitive results that can occur in a game. For example, a game with a 40-point halftime margin will have a non-competitive second half, likely with many bench players and reserves playing the majority of the game. This should be analyzes with care since it occurs in a different environment than the first half of the game, rendering those first half statistics less informative. Regression Points per Possession accounts for this phenomenon.