Does Pulling the Goalie Lead to a Playoff Appearance?

Abstract

The analytics movement took over the sports world in the last five to ten years. The movement originated in baseball and moved to football and basketball with great success. Hockey has yet to undergo an analytical decision-making boom like the other “Big Four” North American sports leagues. The closest decision within hockey that can be examined analytically is pulling the goalie late in games. Pulling the goalie is one of the most researched topics related to hockey. The results of this research have evolved since the 1980’s, but the focus has been on what is the best time to pull the goalie to maximize win percentage. There is little research on if the strategy is worthwhile in the long run. This study uses six years of historical game summary data from the National Hockey League to evaluate if the strategy can help teams improve their place in the standings and their likelihood to make the playoffs. A panel regression with team fixed effects was used to evaluate the effect of optimal pull percent on standing points; a logit panel regression was utilized to evaluate the likelihood to make the playoffs. This study found that the optimal pull percentage of a team is not significant in both models. The team quality effects of goals for average and goals against average were significant in the improving the number of standing points while only goals for average was significant in the likelihood of making the playoffs. What this means for the future is discussed within the discussion section of this study.