This is a brief description of the Simpson Rating model. I've omitted the gory details.

The purpose of this system is to predict the outcome of the NCAA Division I men's basketball tournament. Note that this prediction is for the winning team, not the margin of victory. Although this system predicts the tournament better than any other system I've found, I've discovered no model that would have predicted Arizona in 1997 or Syracuse in 2003.

The weighting of the displayed metrics reflects the decaying effect of games as time passes. The more recent the game, the greater it's effect on a team's rating. If a team has played twenty five games, the weight of the twenty fifth game is 1.00, and the weight of the first game is .04.

All teams begin the season with equal standing; there is no weight given to prior years' results or a team's reputation. Early in the season, the ratings are very volatile due to the small number of games played. The SOS (strength of schedule) rating is especially rough until all teams are connected (think "Six Degrees of Separation").

The power rating accounts for home court advantage, and uses a modified least squares algorithm that compresses the margin of victory to minimize the effect of blowouts. It does not produce predicted point differences.

I began this many years ago in attempt to win the office March Madness pool. After several spreadsheet based versions, I've developed an OLAP system that allows me to publish the ratings as web pages. Version 2.2 was implemented January, 2007. Future refinements will reflect a continuing effort to more accurately rank the teams.

I do this for amusement only, and make no warranty, expressed or implied about the veracity of the numbers. Using the information found on this web site for gambling of any sort is strongly discouraged. You must accept complete responsibility for your actions based on the content of this web site. If you are foolish enough to risk money on the basis of computer ratings, you deserve the result.

I would like to acknowledge Jeff Sagarin, Kenneth Massey, Ken Pomeroy and Mike Greenfield for their efforts.

Send messages to with questions or comments.

(Despite what the numbers may say, ARIZONA is really #1)