This is a review of the NFL schedules research applying mixed-integer linear programming titled “Alleviating Competitive Imbalance in NFL Schedules: An Integer-Programming Approach“. The National Football League has a strong fan base and generates more revenue than any other sports league in the world. During a regular season, NFL games are scheduled primarily on Sundays with the exception of two or three additional games on Thursday and Monday nights. The NFL regular season lasts 17 weeks with each team playing sixteen weeks with one bye week. Having a scheduled game on a Thursday provides a team with three or four extra days for rest and practice while a bye week provides an extra week to prepare for the next match. Due to this, teams can play against more rested opponents, giving the opponent a competitive edge.
The NFL incorporates a variety of complex rules to schedule a season’s worth of games in order to maintain fairness for the teams. However, every year several teams feel that their schedule does not provide them with the same fair opportunities afforded to other teams.
When comparing average win percentages against all opponents versus average win percentages against rested opponents it becomes clear that teams are less likely to win when playing a rested opponent.
A mixed-integer linear programming model is developed to create a schedule which minimizes the number of games in which a team plays against a rested opponent as well as minimizing any long periods of consecutive home or away games.
This model incorporates the thirteen rules currently used by the NFL in determining its schedule as well as an additional nine rules to help create fairer NFL schedules. The model incorporates these rules into designing the schedule while looking to minimize the number of games teams play against a more rested opponent, the number of teams playing three consecutive road games, and the number of teams playing three different sets of back-to-back road games in the course of a season.
It is also essential that the model is able to generate alternative schedules in a reasonable length of time, which can be difficult due to the sheer amount of data and parameters. To handle this, a two-phase model is incorporated which breaks the full model into two simpler programs, solving them consecutively. The first phase assigns the games to weeks without determining the venue and also assigns bye weeks to each team. The second phase then determines who will host each game, taking into account all of the parameters. Breaking the model into two phases increases its efficiency, thus increasing the speed with which it can generate schedules.
Running experiments with the model determines that the schedules generated by the model, as compared to past NFL schedules, treat teams on a fairer basis in regards to playing better-rested teams.
The NFL can use this model to create a schedule that is fair to all teams, minimizing the bias regarding playing better-rested teams. This would help ensure teams win due to their skill and not because they had a longer time to recover between games.
Analysts could compare schedules created by the model with past NFL schedules to determine what effect, if any, it would have had regarding which teams made to the playoffs and which team was the ultimate winner of the season.
Analytics Used: Mixed-Integer Linear Programming Model, Two-Phase Method.
Find out how Sports Analytics Expert Victor Holman can give your team the competitive advantage.
How mature is your team’s analytics program? Take the Sports Analytics Maturity Assessment.
Discover the Groundbreaking Sports Analytics Application and Framework coaches and sports analysts are talking about!
Learn all about sports analytics in Victor Holman’s Sports Analytics Blog.