College Football BlogPoll: Week 1 Ballot Final

The RTT blog poll is PURELY results based. There is zero looking at what we thought last week and seeing how teams should move up or down. It’s just looking at the total body of work and seeing how it looks. After a few weeks, a computer comes in and does most of the poll, but there’s no data yet. If how they looked in week one is different than what we expected in pre-season, there’s naturally plenty of movement from the preseason poll. Also, in the first few weeks, the poll is extremely volatile. For instance, Florida was in the preseason top 10. Last week, there’s no way they played like a top 25 team. But if they play like a top 10 team next week, they’ll pop right back up there, even though they only beat a middling Big East team.

Are a generally poor judge of performance.

but there are basically two tenets: one is to not rely solely on computers (i.e. allow human override if you use computers). We do that, but not all computer users do; some won’t override for any reason whatsoever. You’ll see Cook rant on that nearly every week during the early going, though he rarely spikes those ballots anymore. (When Hinton always does it, Cook has a hard time banning such voters from the poll.)

The other is basically to not be stupid. In theory, we’re supposed to take more time and collect more data before voting – e.g. draft ballots and not having ballots due until Wednesday. Also, we’re supposed to avoid the momentum problem that the regular polls have; if two consecutively ranked teams win and the lower-ranked team appears to be the better team on the season, they should pass the other team. (Basically, don’t try to justify your preseason poll by stubbornly refusing to move teams.)

I infringe a bit by relying on computers more than Cook really cares for, but that’s fine with me. (For a former engineer, he’s surprisingly anti-cpu.)