Professional Sports and Market Potential: A Machine Learning Approach to Analyzing Franchise Valuations

Abstract

The crossover between AI and sports data in the Big Four (NHL, MLB, NFL, NBA) professional sports leagues increases with every passing season. Yet, the usage of AI, specifically machine learning, hasn’t reached the financial operations of a sports franchise. Professional sports franchises in the Big Four are worth anywhere from hundreds of millions to billions of dollars. But what exactly are the main determinants of a franchise’s value? Is this value predictable using publicly available metrics about a professional sports franchise? This paper aims to answer both questions by combining a thorough dataset consisting of financial, franchise-performance, and economic metrics with modern machine learning models in order to predict franchise values and identify the strongest determinants of a franchise’s value. The findings of this research indicate that while models predict franchise value with excellent efficiency and demonstrated significant model evaluation metrics, the lack of previous models to provide a benchmark hold back the conclusion of an accurate prediction. All machine learning models relied heavily on financial data, indicating that financial metrics are more influential drivers of franchise value over economic and franchise-performance data. This thesis provides groundwork for future work looking to combine machine learning and the financial aspects of a professional sports franchise.