The annual NCAA Division I men’s basketball tournament is one of the most prominent sporting events in the United States and the subject of much interest among fans and statisticians alike. In this work, we propose a new method for predicting theresults of the tournament based on a semiparametric quantile regression model. The idea is to estimate a win probability by averaging across multiple conditional quantiles. To demonstrate the finite sample performance of the proposed methodology, we predict the tournament results for the years 2016 - 2019 and 2022. The results are then compared with other commonly used methods and rankings. Our method is competitive and offers a novel approach for use in bracket prediction.