Predicting game or season outcomes is important for clubs as well as for the betting industry. Understanding the critical factors of winning games and championships gives clubs a competitive advantage when selecting players for the team and implementing winning strategies. In this paper, we work with NBA data from 10 seasons and propose an approach for predicting game outcomes that is then used for predicting which team will be champion and which stages a team will reach in the playoffs. We show that our approach has a similar performance as the odds from betting companies and does better than ELO.
© Copyright 2022 Machine Learning and Data Mining for Sports Analytics. MLSA 2021. Communications in Computer and Information Science. Published by Springer. All rights reserved.