The Impact of Puck Possession and Location on Ice Hockey Strategy

I create a state space within the game of ice hockey by noting which team has possession, and in what location of the rink the puck is located. This space is used to model the game as a semi-Markov process, as data from a series of games in 2004-2005 NCAA play suggest that the system cannot be modeled as a continuous time Markov process. The model is then used to determine the average number of goals scored by a team as a function of the starting state. These scoring probabilities are used to demonstrate the effectiveness of several commonly used tactics. ∗The author thanks George Lindsey, Gopi Goswami, Jim Greiner, Erol Pekoz and Carl Morris for their helpful suggestions, as well as his anonymous reviewers.