Analyzing Passing Metrics in Ice Hockey using Puck and Player Tracking Data

Authors

  • David Radke Cheriton School of Computer Science, University of Waterloo, Canada
  • Jaxin Lu Cheriton School of Computer Science, University of Waterloo, Canada
  • Jackson Woloschuk Cheriton School of Computer Science, University of Waterloo, Canada
  • Tin Le Cheriton School of Computer Science, University of Waterloo, Canada; University of Calgary, Canada
  • Daniel Radke Cheriton School of Computer Science, University of Waterloo, Canada
  • Charlie Liu Cheriton School of Computer Science, University of Waterloo, Canada
  • Tim Brecht Cheriton School of Computer Science, University of Waterloo, Canada

DOI:

https://doi.org/10.3384/ecp201.3

Abstract

Traditional ice hockey statistics are inherently biased towards offensive events like goals, assists, and shots. However, successful teams in ice hockey require players with skills that may not be captured using traditional measures of performance. The adoption of puck and player tracking systems in the National Hockey League (NHL) has significantly increased the scope of possible metrics that can be obtained. In this paper, we compute recently proposed passing metrics from 1221 NHL games from the 2021-2022 season. We analyze the distributions of values obtained for each player for each metric to understand the variance between, and within, different positions. We find that forwards tend to complete fewer passes with smaller passing lanes, while defensemen pass to forwards significantly more than their defensive partners . Additionally, because these new metrics do not correlate well with traditional metrics (e.g., assists), we believe that they capture aspects of players’ abilities that may not appear on the game sheet.

Downloads

Published

2023-09-08