De-constructing a Popular Preseason Poll

[mb21]

For what it's worth, Hall of Fame voters in baseball often publish their ballots now and when criticized or asked to defend a choice, they get snippy and often resort to personal attacks for the person simply asking a question. I think we'd end up with a situation like that if there was more transparency. The voters, at least most of them, feel an entitlement to voting and feel their opinion is better than anyone else's and that they shouldn't be judged. Perhaps I'm underestimating them and using the MLB HoF comparison poorly, but I've often wondered in recent years why some who vote for the HoF even bother to publish their ballots if they are thin-skinned and not willing to be challenged.

It does provide great humor though. Some of my favorite articles throughout the year on baseball are where the writers explain their ballots and the justifications for their picks. It's comedy and I think we'd at least end up with that so I'm all for it.

 

[mb21]

His point was that the poll specifically said it ONLY was going to consider how good a team is RIGHT NOW, then predicted the team's record. But there was an awful lot of projection/predictions going on. Florida and Texas were the two prime examples. They are NOT top 5 teams right now. You don't lose as much as those two teams did and be THAT good right now. They may very well end up that way by the end of the season. But that's projecting how good the team might be, not how good they are right now.

Fair enough, but I don't agree that you can't lose that much and still be a top 5 team. It makes it a lot harder, but recruiting is a factor and Texas' recruiting over the last 5 years has been in the top 5 in the nation. It stands to reason they have very talented players to fill in. I'm not saying Texas is a top 5 team. I don't know.

 

[tm3308]

Fair enough, but I don't agree that you can't lose that much and still be a top 5 team. It makes it a lot harder, but recruiting is a factor and Texas' recruiting over the last 5 years has been in the top 5 in the nation. It stands to reason they have very talented players to fill in. I'm not saying Texas is a top 5 team. I don't know.

Okay, we'll just say it's exceedingly difficult to lose that much and still be THAT good, particularly when both of those teams lost QB's that will go down as 2 of the best QB's in CFB history. Texas was VERY reliant on McCoy, both as a passer and as a rusher. And he leaned on Shipley pretty heavy. Both of them are now gone. Gilbert is going to be a very good QB, IMO. But he's not there yet. I'm willing to give Brantley the benefit of the doubt on that one, as he's older and has more playing experience than Gilbert. So Florida I could see a fairly reasonable argument for. Texas has no business being that high at this point, if the poll is only going on how good the team is right now.

 

[JonDMiller]

I am not a big fan of there being preseason polls that are a part of the BCS process. That is a biased process, as the only way you can do it is with past notion carry over, historical exceptions, etc.

Starting the polls in Mid October like the BCS and the Harris Poll do is the way to go, IMO.

If you are ranked in the preseason top ten and lose a game early to a solid team, and win the rest, there is a great chance you are going to be near the top four, which means automatic qualification for a BCS bid. If you are not one of the historical hotties, and you go 8-0 to start the year and then lose your last regular season game of the year, you will be ranked lower than the team that started the year in the top ten. Should that matter? I don't think it should matter as much as it does.

Think about this...the 2002 Iowa football team...they were one of the best in school history, but they didn't show up in the rankings until the WEEK 6 Poll! There is little doubt in my mind they were one of the three or four best teams in America that year...yet they didn't sniff the rankings until September 29th

 

[mb21]

Jon, I don't disagree with you. I don't even think polls should start as soon as you do. To be honest, I see no reason for them to come out until week 9 or week 10. At that point we have a pretty good idea where teams rank relative to one another.

I'm wondering if anybody has ever looked back at the major pre-season rankings and compared them to the post-season Sagarin ratings. I don't think citing one or two examples, or even three or four is evidence of bias, though I also believe it is a huge issue. There will always will be outliers (2002 Iowa for example), but we remember that one because it was an extreme example of the bias we're talking about. I think it exists. I'm just interested to know to what extent it really exists.

If we have to settle on human polls to do the rankings, we consciously accept some degree of bias. What is that degree? How does it compare to the actual bias of pre-season ratings and post-season Sagarin? How far off is it currently? How can it be improved? People aren't going to accept an all-computerized system so that's the important question to ask and answer.