Skip to content
Pioneering the Use of AI in Analyzing and Tracking Professional Hockey Data

Pioneering the Use of AI in Analyzing and Tracking Professional Hockey Data

The team of researchers from the University of Waterloo has engineered artificial intelligence (AI) tools that enable them to analyze and track data from professional hockey games more effectively and expediently than any previous efforts. These innovative tools are set up to have substantial ramifications for the sports industry.

An existing problem in hockey analytics is the extensive reliance on manual video analysis of games. This information becomes a significant factor when professional hockey teams such as those in the National Hockey League (NHL) have to make crucial decisions about players' careers.

Dr. David Clausi, a professor in Waterloo's Department of Systems Design Engineering, explained their research goal as, "Interpreting a hockey game through video more effectively and efficiently than a human." The difficulty lies in the non-linear movement of hockey players at high speeds and the resultant change in the visibility of jersey numbers and last names. The resulting manual tracking of each player during a game often result in human errors.

The AI tool developed by Dr. Clausi, Professor John Zelek from Waterloo's Department of Systems Design Engineering, Assistant Research Professor Yuhao Chen, and their team of graduate students uses deep learning techniques to automate and revamp player tracking analysis. Stathletes, an Ontario-based professional hockey performance data and analytics company, partnered with the research team in this endeavour.

The researchers manually annotated teams, players, and players' movements from NHL video clips. Machine algorithms were then designed to analyze this data, opening up new pathways for more accurate analysis and predictions.

The researchers have achieved remarkable results with their system. The algorithms can track players with an impressive accuracy of 94.5%, identify teams correctly at a rate of 97%, and correctly identify individual players 83% of the time.

The prototype is currently under further refinement, but Stathletes has already adopted the system to annotate hockey manoeuvres. The potential for the application of these techniques stretches beyond hockey and into sports like soccer or field hockey.

Professor Zelek explained the multifaceted benefits of the system: "Coaches can use it to craft winning game strategies, team scouts can hunt for upcoming players, and statisticians can identify ways to give teams an additional advantage." Drawing from this, we can confidently say that this system is bound to alter the landscape of the sports industry.

Disclaimer: The above article was written with the assistance of AI. The original sources can be found on ScienceDaily.