Ryan Dodds Thesis

Open Access
Dodds, Ryan
Area of Honors:
Industrial Engineering
Bachelor of Science
Document Type:
Thesis Supervisors:
  • Soundar Kumara, Thesis Supervisor
  • Catherine Harmonosky, Honors Advisor
  • Data Analysis
  • Visual Basic
  • Excel
The purpose of this study is to attempt to level the playing field for the average daily fantasy football user. Despite the appearance of a fair, reasonable contest, almost all of the daily fantasy sport winnings are won by only the very, very top performers who are often well-funded, highly adept at statistics, and benefitting from very advanced models that predict the performances of players. In order to help average users gain an advantage over the competition, several models were created that require little statistical prowess to operate consistently. While they were quite complicated to create and optimize, the common daily fantasy football participant would likely have little trouble downloading, setting up, and running these models on a consistent basis to further their understanding of the statistics and data analysis in play behind the sport. Unfortunately, the average weekly score was not able to be statistically significantly increased. However, there were very promising results in the area of increasing prediction accuracy and reducing error to only 0.62 fantasy points on average. Given how important it is for daily fantasy football users to minimize uncertainty, this is an impressive result that would benefit many of those in the target audience of this paper. While it wouldn’t necessarily be able to outperform the competition in any given week, it would allow users to pick and choose when they should enter the competitions in order to minimize their losses and maximize their return on investment. Although the original desired outcome was unable to be achieved without the use of models beyond the scope of this experiment, the models that were created delivered impressive results that would benefit many daily fantasy football participants.