The Spirit of the Game: Developing an Algorithm to Create a Competitively Balanced NFL Schedule

Abstract

This thesis will consider what it means to develop a competitively balanced NFL schedule. Using data from previous NFL seasons, this thesis will model the expected performance of a team in a game based on a variety of variables controllable by the schedule makers. These performances will be compared to the baseline metrics of a team, and develop what will become "fairness factors" for each NFL game. A model will then be made to develop a 17 week NFL schedule that minimizes the overall fairness factor magnitude for the season using different approaches. Analysis and discussions will consider trends in the model, the strengths of the model, areas for improvement within the model, and how the model can be applied to other sports.