Wilson Computer Ratings for College Basketball

What is all this?

On this page I produce computer ratings and rankings for all NCAA men's basketball teams, using the algorithm developed by David Wilson (UW-Milwaukee '65, '66; UW-Madison '67). For many years, before I developed my own algorithm (available elsewhere on the site), the Wilson ratings were the primary ones that I shared. I've been producing ratings using his algorithm for over 20 years, and I'm honored to continue sharing his system here.

But why?

There are quite a few people who produce computer ratings for Division I basketball, and that is pretty big business (such ratings are the foundation for virtually all betting lines). But there are fewer sites that collect and share information for Division II and III, and hardly anyone provides ratings and rankings of all teams. So that's my contribution: a page that will take the games of all ~1000 NCAA teams and produce comprehensive rankings, plus easy-to-read conference standings.

Are the Wilson ratings any good?

I think it was in the 2010's, someone actually tracked how accurate this system was at predicting who receives Division I at-large bids. And the results were pretty good. Ultimately, "are they any good" is up to your own interpretation... my reply is always that if you don't think so, then just enjoy the standings part of the site!

What data do you use?

  • Game scores: The ratings are built with nothing other than scores of D-I, II, and III games played. Scores are collected by Ken Pomeroy (D-I), John Kean (Missouri S&T sports information director; D-II), and from d3hoops.com. I am indebted to their work; without their effort, my site would not exist. I have to do very little, other than some spell checking and checking for duplicate games. If you want the games file for your division, it's available on the left.

  • Making D-II & D-III ratings more accurate: I am occasionally importing NAIA games so that the strength of those opponents will be better represented. Unfortunately, my upstream source for those games, National Statistical, does miss a few games so the result is definitely not perfect. But overall this is still a net positive -- a more accurate strength of schedule assessment for D-II and D-III.

  • Early in the season: The original Wilson algorithm was designed to start every team at the same place (a rating of "zero," basically), but I don't believe that's accurate -- Kansas and Prairie View are probably not equal, not even in November. So until about Christmas, I keep some of last year's games in the calculations to give them a meaningful starting point.

Input and output files for the Wilson system:

Input files:

Output files (as "raw" text files):

Who the heck are you?

I am a faculty member at Indiana University in Bloomington, with a Ph.D. in Atmospheric Science from the University of Alabama in Huntsville. For my day job, in addition to teaching lots of courses I have research interests in the broad areas of thunderstorms and numerical weather prediction (forecasting the weather using computers).

I do love D-I basketball but thanks to my years at UAH, I still have quite the soft spot in my heart for D-II. (D-III folks, I need to go to more of your games, I'll admit!)

I found an error, or I want to get in touch.

Contact me using this email form if you have questions or non-hateful comments.