He bases that analysis completely off point differential in games (including a calculated strength of schedule based on the same information). Wins and losses don't directly affect it at all. It's not as "fair" or "sensible" as a system that relates all wins and losses equally, but it's a much better predictor of future performance (thus why he calls it his "predictor"). He actually runs three separate ratings, one on pure point differential, one on pure wins/losses, and one 50/50 combination of the two. Agree with some of it,but some is off I think, but that's a pretty cool ranking system The main drawback to this and Ken Pomeroy's system (http://www.kenpom.com) is that it's still early enough in the season that a blowout or two in either direction can really affect the overall standing of a team. As the season progresses, the systems get more and more accurate. OK I was kind of wondering why Kansas was rated ahead of both Oklahoma and Texas in the Sagarin system. Now it makes sense. Oklahoma hasn't really been blowing teams out this season.