Ranking College Soccer Teams

Underdeveloped data-driven approaches for ranking college teams hinder the potential to generate compelling information about team performance, forecast game outcomes, and inform playoff seeding. We developed a probabilistic ranking system for college soccer teams that quantifies uncertainty in team ratings, interprets each team’s rating as the expected number of goals required to defeat an average college team, directly models draws, and employs an adjusted margin of victory metric, resulting in more accurate and understandable rankings. Our model surpasses the FIFA ranking system in predicting match results, scores, and offers a more nuanced and accurate picture of team performance than traditional methods or basic winning percentages, generating valuable information about team performance, predicting game outcomes, and informing playoff seeding.