Modeling NFL Football Outcomes
Three modeling techniques were used to develop models that explain the outcomes of NFL football games. The modeling techniques applied included ordinary least square regression, logistic regression, and proportional odds. Two seasons of NFL football games were used as the training data set and one season of NFL football games was used as the testing data set. The OLS model developed explained approximately 83% of the variation in the point spread of a football game. The logistic regression model developed estimates the probability of the home football team winning given the differences of significant in-game statistics. The proportional odds model estimates the probability of a home team winning the game by 10 or more points or less than 10 points, or losing the game by these same amounts given differences of in-game statistics found to be significant. The models all did well under the testing data set.
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
- All contributor(s) agree to transfer the copyright of this article to EPH Journal.
- EPH Journal will have all the rights to distribute, share, sell, modify this research article with proper reference of the contributors.
- EPH Journal will have the right to edit or completely remove the published article on any misconduct happening.