Player Availability Rating (PAR) - A Tool for Quantifying Skater Performance for NHL General Managers

Abstract: This project aims to assess the performance of various regression models in predicting the performance of hockey players. The measure of performance is chosen to be points scored (sum of goals scored and assists made) by individual players per game (PPG). This paper thus considers the offensive performance of the players and uses PPG as a metric to assign a value to the player. This predicted value can be used to rank players and is similar to what TSN.com and NHL.com use to rank players before each hockey season. A combination of physical characteristics, shooting percentage, and usage during games are used as training features. A novel metric, Player Availability Rating (PAR) is proposed which utilizes the offensive predictions to predict the availability of a player during a season. These metrics can be valuable in the NHL as it would allow for general managers to track a players performance and their availability during the course of a season.