Predictive Modeling and Analysis of Hockey Using Markov Chains

Abstract

Predicting the outcome of a hockey game can be challenging due to the fast paced and physical nature of the sport. In this paper, we share two versions of a continuous-time Markov process-based model that takes the certain state the home team is in at any point in the game and gives a winning probability statistic for that time. This state is based on the home team’s shot and goal differential relative to the opposing team and approximates the probability that the home team would win depending on the state they are currently in at a given time in the game.