The Predictive Power of Advanced NFL Analytics: Using Machine Learning Strategies and Advanced Analytics to Beat the Books

Abstract

Professional football has become increasingly competitive over the years, leading teams to turn towards advanced analytics to gain a new edge on their opponents. Advanced analytics put statistics within the context with which they occur to give a more accurate representation of a performance’s significance, aiding in the decision-making process. As teams pour more money into applying data analysis to game strategy, the stakes of professional football rise higher, drawing larger audiences and more gamblers. In the gambling world, the average bettor is expected to lose in the long run due to a sportsbook’s advantage in determining the odds. This thesis uses various machine learning algorithms with the intention to aid the average bettor by applying advanced analytics to the betting world via money line and spread wagers. While the models perform well in comparison to the average bettor, they are not able to perform well enough to overcome the advantage of the sportsbooks. Regardless of how accurately a sportsbook sets its lines or predictions on a game, the sportsbook will always come out on top so long as it balances the number of bets on each side appropriately.