Date of Completion

2024

Document Type

Honors College Thesis

Department

Mathematics and Statistics

Thesis Type

Honors College

First Advisor

Jacob Martin

Second Advisor

Karen Benway

Keywords

Statistics, Regression, Sports, Analytics, Stepwise, Modeling

Abstract

Professional sports are one of the most consumed forms of entertainment in the world today. Professional sporting events are some of the most watched broadcasts worldwide each year. The 2022 FIFA World Cup Final garnered about 1.5 billion views worldwide, almost 20% of our planet’s population (Jones, 2023). The National Football League is the most popular professional sport in the United States. Recent polling data shows that a clear majority of the country, 72% of Americans, self-identify as football fans (“St. Bonaventure”, 2023). The NFL runs from September to February and regularly draws 15-20 million viewers weekly during the regular season, which lasts 18 weeks. The regular season is followed by the playoffs, which last five weeks and draw more viewers than any regular season games. The playoffs culminate in the most watched single broadcast in the United States each year: the Super Bowl (Stoll, 2024). This project took a data-based statistical approach to predicting the outcome of NFL playoff games using historical performance data combined with modern-day statistical modeling techniques. Specifically, this project used a stepwise variable selection approach to determine statistically significant predictors and integrate them into a multiple linear regression model to predict the final score of NFL playoff games. This project’s final regression model predicted the correct winner of NFL playoff games 64.41% of the time and the correct winner against the spread 56.78% of the time over the past ten years.

Creative Commons License

Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License.

Share

COinS