While the sample size may be large enough to rank teams based on merit, it does not happen. OkSt lost in OT, while Alabama lost at home in regulation....
That's a very good point. Well, at least, it would be, except that of course Alabama lost in OT as well. But other than that small detail, it's very good indeed. ;)
Voters may not consciously take preseason rankings into their decisions, but polls pretty much work as a ladder, and as such knowingly or not, the voters rank them such as that. I think a lot of stuff points to pre-season rankings mattering a lot. What say you when 3 teams are undefeated....? Voters rarely if ever lower a team that does not lose. So what you think the voters would look at the 3 teams seasons as a whole and make some informed judgement? I dont, I think they would do what they do every season.
You mean like this season, when the top four in the preseason (AP) were: Oklahoma, Alabama, Oregon, LSU? But, wait, week 2 had Oklahoma, LSU, Alabama. Week 3 was Oklahoma, Alabama, LSU. Week 4 was Oklahoma, LSU, Alabama. Week five was LSU, Oklahoma, Alabama. Weeks 6-8 were LSU, Alabama, Oklahoma. All three were still undefeated, all three had been ranked 1, 2 and 3. So the thought that teams never get lowered if they don't lose fails that test right there.
Of course, the motion in the coach's poll wasn't as dramatic, and Oklahoma never entirely fell out of #1, though their margin fell to 11 points. Alabama and LSU did still flip, though, and there were plenty of other flips early in the season - just looking at Boise State, they passed Stanford in week 2, FSU in week 3, and then got passed by Stanford in week 5, and Wisconsin in week 6 (who also leaped ahead of OkSt that week).
Heck, just last year Florida was ranked #4/#3 (AP/Coaches) in the preseason and by week 2, was ranked 10/7 (at 2-0). Even after going 4-0, they were #7 in both polls -- seems a far cry from teams keeping their spot automatically. In the past four years in the coaches poll, an undefeated team has dropped in rank 36 times.
I would even argue, at least as of recently, more correcting of the pre-season rankings happen early rather than late. If you look at first few weeks this season, you will see some teams who won but dropped a bit in the rankings. Simply put you do not see that towards the end of the season, you just dont.
I wish I had read this before the last few paragraphs. ;)
In the final coaches' poll, Nebraska, Penn State and Florida State dropped despite not playing; West Virginia won and dropped. There's more movement in the AP poll, of course, as they actually seem to take it more seriously. But I'm confused -- if you think the polls should measure the full body of work of the season, wouldn't it make a lot more sense for there to be less movement late? By the last weeks of the season, you've got a pretty solid body of work for each team and can evaluate them fairly well, so the late games shouldn't change the evaluation that dramatically. Early on, of course, the preseason rankings are going to be compared with the performance on the field and you'll see more volatility.