An Analysis of “Weak Goals” as an Additional Tool for Evaluating Ice Hockey Goalies

• Activity Number: 473 - For the Love of the Game: Applications of Statistics in Sports
• Type: Contributed
• Date/Time: Wednesday, July 31, 2019 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Sports
• Author(s): Ryan Savitz* and Helen Cooney and Mya McQueen
• Companies: Neumann University and Neumann University and Neumann University
• Keywords: Ice hockey; logistic regression; weak goals; Cronbach's alpha; Sabermetrics

Abstract

This paper introduces a statistic that may be of potential value in evaluating the value of ice hockey goalies to their teams. This new statistic is the number of weak goals allowed per game. The concept of a weak goal has been used casually by hockey enthusiasts in the past, but we formalize the definition of such. The newly formalized definition of a weak goal was statistically validated, using a novel application of Cronbach's Alpha, to be a good one. The value added of this new statistic was then analyzed by conducting several logistic regressions. The dependent variable for each regression was win/loss status, and several sets of predictors, including weak goals, were included as the independent variables. The new weak goal statistic was found to be of borderline statistical significance, and to add a small amount of marginal value to the ability of extant goalie statistics to predict wins.