Analyzing the Impact of Season Formats on Team Ranking Accuracy in North American Professional Sports

Headshot of Lathan Gregg

Author: Lathan Gregg | Major: Industrial Engineering | Semester: Spring 2023

This past semester I have begun my research investigating the effect of different formats on the outcome of seasons in North American professional sports under the mentorship of Dr. Richard Cassady. For my research, I am developing a simulation to explore the effect of factors such as season length, number of teams, and number of postseason rounds on the season’s ability to accurately rank teams. For each simulated game, both team’s probability to win will be determined by a strength value assigned during the simulation and the outcome will be randomly determined based on this probability. To determine how accurately each league’s season format is at ranking teams, the standings resulting from the simulation will be compared to the rank of each team’s assigned strength.

I chose to research this topic after hearing Dr. Cassady present about a former student’s research on a related topic in the Honors Research Experience class I took last year with Dr. Kelly Sullivan. As part of the course, we had the opportunity to hear from the Industrial Engineering faculty about their research with goal of identifying an undergraduate research topic and mentor by the end of the semester. Dr. Cassady’s research particularly interested me because it combined my interest in sports and statistics and allowed me to investigate if a championship is a reliable indicator that the champion was the best team.

This semester, my research has consisted of investigating related studies and developing the simulation. This background research has prepared me to develop my simulation by showing how past researchers have explored the efficacy of tournament formats. To develop my simulation, I am using Java to model relevant factors of seasons for the National Football League (NFL), Major League Baseball (MLB), the National Basketball Association (NBA) , and the National Hockey League (NHL). A challenge I have faced this semester is figuring how to translate the relevant factors of a season to my simulation. One aspect I am modeling that has been particularly challenging is the generation of a regular season schedule for each team. To overcome this issue, I have had to identify what aspects of a schedule are relevant to the conclusions I am trying to draw. I have also had to learn more about coding in Java to identify fast and effective ways to implement the relevant aspects to my simulation.

At this point in my research, I am nearing the completion of my simulation model for the NFL. Once this is complete, I will replicate the code for the MLB, NBA, and NHL. From here, I will refine the simulation by adjusting the parameters to accurately reflect historical data. For example, I will compare the simulated distribution of regular season winning percentages to historical data and modify the parameters for assigning team strength to ensure these align. Once this is complete, I will investigate the results and work with Dr. Cassady to determine the best way to present my findings.

Although this experience has been challenging at times, it has been rewarding to apply what I have learned in my classes to my research while expanding my knowledge and abilities. I am ready to continue my research next semester and am excited to see what the results will indicate about the different season formats.