Never Tell Me The Odds

Liptrap, Keller

Never Tell Me The Odds

Persistent URL

https://hdl.handle.net/10456/59230

Author(s)

Liptrap, Keller

Date Issued

April 18, 2025

Abstract

This project developed a tool that uses data from the past five NBA seasons and machine learning models to predict the outcomes of NBA games. This tool leverages the growing access to and interest in advanced sports analytics. The tool named “Never Tell Me the Odds” provides a user-friendly dashboard to display predictions of NBA games for sports fanatics all around the world. The tool uses the nba_api to gather historical data, which is used and processed by machine learning algorithms to generate predictions. The backend of the tool handles all of the processing and training of the data, while the front end is a straightforward dashboard to show each of the predicted winners made by the models. The various machine learning algorithms, including linear regression, random forest, and support vector machines, were compared to find the most effective for NBA game prediction. Several experiments evaluate the tool’s performance to determine what influences the accuracy of predicting real-world NBA games. The result of the project indicates that there is potential for machine learning to predict NBA game outcomes. The tool and project offer insights for fans, analysts, and sports bettors. With this being said, there is an acknowledgment of the limitations of the machine learning models, such as player injuries or trades, as well as ethical considerations when it comes to sports betting.

Major

Computer Science

First Reader(s)

Jumadinova, Janyl A.