Strategies for Pulling the Goalie in Hockey

• Activity Number: 482
• Type: Contributed
• Date/Time: Wednesday, August 4, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Sports
• Author(s): David Beaudoin*+ and Tim B. Swartz
• Companies: Universite Laval and Simon Fraser University
• Keywords: National Hockey League ; Simulation ; Markov chain Monte Carlo ; Bayes constrained estimation

Abstract

This project develops a simulator for matches in the National Hockey League with the intent of assessing strategies for pulling the goaltender. Aspects of the approach that are novel include breaking the game down into finer and more realistic situations, introducing the effect of penalties and including the home-ice advantage. Parameter estimates used in the simulator are obtained through the analysis of an extensive data set using constrained Bayesian estimation via Markov chain methods. Some surprising strategies are obtained which do not appear to be used in practice. For example, results suggest that the best strategy for a team trailing by one goal with 3 minutes left in the game consists of pulling its goaltender until the score is tied, no matter the game situation (i.e. 5-on-5, 4-on-3, 4-on-5, etc.).