Whenever this guy is mentioned, it seems to be about how wrong he is. Which begs the question, why is he mentioned so much?
I'm not saying he's wrong, I'm saying his numbers are often quite interesting. In fact, I think he's right more often than not, and I typically use him for a benchmark of what teams real strengths are.
The numbers can be misleading, and they can be manipulated by anomalies (as in the relative strength of the Pac-10 last year), but more often than not I think he's a better barometer than the human polls.