Model Comparison for Professional Basketball’s Player Efficiency Rating

The Player Efficiency Rating (PER) is a statistic commonly used by Sports Networks to compare and contrast Professional Basketball players by attempting to quantify an individual player’s value while on the court. Though it is widely used, there are still several flaws that need to be addressed which Hollinger himself has openly admitted. For example, it is evident in the formula that offense is overvalued whereas defense is undervalued since there are more intangible and unrecognized factors on defense that contribute to a player’s success. This research will seek to develop a comparable model that will be used to compare against existing models. Using the existing models as the standard, I will compare each developed model using linear regression, residual sum of squares, and finding the coefficients of determination — which will attempt to prove that there is a better model that addresses these flaws accordingly and more accurately quantifies a player’s contributions.