Not how he has it setup. Basically he mutes the early season game data such that it is rather small in deciding ranking compared to preseason predictions. As each progressive game is played, the amount of muting gradually goes down. If I recall correctly, by about the 12th game of the season, game results are close to having full weight.
Statistically this makes sense as it's difficult to balance real results on very low data points and your predictive data. Pomeroy opts to be more conservative early putting a lot of weight of preseason prediction and less weight on in game results, because otherwise, we're a 4-0 team with a win over a 3-1 team on a neutral court. Good, but there's also about 50 additional undefeated teams right now, so each day rankings would be a lot more fluid (very very fluid early season) and the average person hates that, haha, so his reasoning makes a lot of sense.
That said, personally I love the only real results, no predictive component to stats and I think college football would be much much much better if this were how rankings were handled instead of an AP poll, as teams would be ranked entirely on their on-field performance. It would get rid of all jersey and conference bias that voters have. That said when you have a team like Alabama ranked 50something after 1 game their fanbase would kill people, so there's a give and take there too I guess.
