A Machine Learning Approach to Throw Value Estimation in Professional Ultimate Frisbee

Research Paper will be posted in the coming weeks. Check back soon!
Download the
Full Paper Here
Authors

Eberhard, Braden; Miller, Jacob

Abstract

In the past decade, Ultimate Frisbee – commonly known as ‘ultimate’ – has transformed from a largely amateur sport to a professional arena with dedicated athletes and multiple leagues including the Ultimate Frisbee Association, the Premier Ultimate League, and the Western Ultimate League. Unlike established professional sports with sophisticated analytical frameworks like baseball's sabermetrics or football's Next Gen Stats, ultimate has historically relied on basic counting statistics such as goals, assists, and blocks, with analysis often limited to post hoc volunteer-tracked metrics. The emergence of professional leagues has been pivotal in driving more thorough data collection, with new tracking systems now capturing unprecedented detail – recording aspects of every throw, including thrower and receiver location, throw outcome, and game time. Despite these advancements, analytics in ultimate are still underdeveloped, leaving room for more refined methods to assess player contributions and team strategy.