A Good Shooter or Just a Good Shot Selector? Using Python to Find Out

Traditional basketball statistics like field-goal percentage fail to distinguish between shot difficulty and a player's actual shooting skill. This analysis proposes using a logistic regression model in Python to calculate an expected field-goal percentage (xFG%) for each shot based on features like distance and shot type. By converting xFG% into expected points per shot (xPPS), it becomes possible to separate a player's shot selection quality from their shot-making ability. The model demonstrates how to measure a player's performance in "points above expected," providing a more nuanced view of their contribution than simple percentages.