How well can random walkers rank football teams?

That's a question that two of my collaborators and I asked about two and a half years ago as part of an undergraduate math project at Georgia Tech. Actually, the other faculty member (Peter Mucha, now at UNC at Chapel Hill) first asked the question, the student we co-advised (erstwhile undergrad Thomas Callaghan, now an applied math grad student at Stanford) did nearly all of the real work on the project, and I just tried to contribute some knowledge of network theory. I want to emphasize that I was the third person out of three on the project. Also, Thomas is one of the best students I have ever had the pleasure to advise. He has a very bright future ahead of him even though he chose to go to Stanford for grad school rather than Caltech. :)

Anyway, you can find our answer an explanation, as well as rankings for 2003--2005, at this project's official web site. I'm posting this now because we've finally started posting our 2005 rankings, which you'll notice are pretty similar to the ones that the BCS and the other methods give.

At our site, you'll find a brief explanation of how the ranking system works, links to Ken Massey's comparison page, our two manuscripts (one published, the other under review), and some of the press coverage we got (and references to some of the other press coverage). (I still have never seen the CNN Headline News interview that came out of this project! We were supposed to get a copy of this...) I am probably the only member of Caltech's physics department who has been featured (however briefly) in ESPN: The Magazine.