We’re starting on an ambitious path—making a ranking system for every event, every discipline, men and women. Can it be done? More crucially, can it be done well?

I recently spent a year analyzing predictive rating systems for the sport of college football for my data science master’s thesis. I can’t link it here because it’s not publically available yet, but I’m sure it will find its way to this website eventually. That is to say, I do have some experience with this kind of thing, and there seriously may not be another person on earth simultaneously more informed and critical of such rating models—I care deeply about getting these things right to the extent they can be.

I’m going to create something that doesn’t exist in any sport to my knowledge—create a ranking system comparing men and women together. The model will try to measure itself against current records for men and women at each distance and each discipline, from the 800m all the way to a 250-mile ultra, in order to determine runner rating.

Will I be using AI? Yes, a lot. This is something that would take far longer without that kind of help but would not be able to be done well without a certain amount of human expertise. There will not be anything AI does that I didn’t tell it to do and that won’t be double-checked.

One of the biggest concerns I have about existing models in sports is transparency. So I’m going to do something novel and build the model with you watching over my shoulder. It might take a while, but I will keep you updated on every step along the way. Pull up a chair, bring a friend—we’re in it for the long haul.