Application of PageRank Algorithm to Division I NCAA men’s basketball as bracket formation and outcome predictive utility

Abstract

This article examines the use of the PageRank algorithm to rank the teams and predict team performance in the tournament. This method has the potential to be utilized as an alternative method to choose tournament participants, as opposed to the traditional ranking and seeding process currently employed by the NCAA. PageRank allows for the consideration of all games played during the regular season (average of 5832 games per season) and for customizable performance weights in the prediction. The PageRank algorithm is a viable tool in predicting tournament outcomes due to depth and extensiveness of the data. The PageRank analysis helped to predict over half of the tournament game outcomes correctly in 2014-2018 and produced an average bracket score of 81.8 points out of 192 possible over the same 5 years.