Fight Prior MMA Data Science and Statistical Analysis


Our Goal

Our goal is to better understand the factors that differentiate fighters. When two fighters enter the ring, anything can happen. Nevertheless, the likely outcome of a fight is largely determined by each fighters’ past performances and training, as well as broader-trends in the sport as a whole. We chose the name “Fight Prior” to reflect that a fight’s outcome is determine not only on fight night, but also by the the fight’s prior: the history leading up to the bout.

Every fan builds their own “fight prior”; opinions leading up to a fight establish betting odds and subjective opinions of who you think is going to win a fight.

We want to use data to build our fight prior.

Many factors, both measureable and unmeasurable contribute to a fight’s prior. We will work to shed light upon these factors using approaches steeped in data science, statistics and machine learning. We want these results to help enrich the sport of MMA, better informing fans and fighters alike. Accordingly, we will be publically presenting results on authored by both the site creators and other members of the community.

Our current focus areas are:

  • Understanding how the fan-base shapes fighter popularity
  • Summarizing the trajectories of fighters’ careers
  • Quantifying fighter demographics and styles


Fight Prior was created and is edited by Sean Hackett and John Storey. Sean has a purple belt in Brazilian Jiu-Jitsu and a PhD in quantitative and computational biology. John has a PhD in statistics and is a professor of genomics and statistics. He’s a fan of combat sports such as MMA and hockey.

Why Fight Prior

There has been an explosion of both interest and participation in MMA to the extent that the field can now be considered a “big data” problem. We, the authors, want to bring tools from data science that have revolutionized other fields into this space.


The data analyses and views published here are our own; they do not represent those of our employers or funding agencies.