Just for fun, a quick AP pre/post-season analysis

[Luftgekuehlt67]

It came up in another thread how funny it is to see AP voters apparently fall for the same washed up, former glory teams year after year in the preseason poll. That's certainly my impression, but it got me wondering if there is any truth to it.

Just for funsies, I took all the pre-season and post-season AP polls between 2000 and 2020 and then did a simple diff:

(Number of Times Making Post-Season Top 25) minus (Number of Times Making Pre-Season Top 25)

Top 5 "Most Disappointing" teams:

1) Tennessee (-6)
2) USC (-5)
3) Georgia (-4)
3) Florida State (-4)
3) Nebraska (-4)

Texas, the team that put the topic on the table, was next at -3.

Top 5 "Pleasant Surprise" teams:

1) Cincinnati (+6)
2) Boise State (+4)
2) Boston College (+4)
2) Utah (+4)
2) BYU (+4)

For the "Most Disappointing" list, none of those teams really surprised me. Tennessee, USC, and Nebraska I think I would have guessed as being quite high off the top of my head, for sure. On the "Pleasant Surprise" list, I don't think I would have had many guesses. I think maybe - MAYBE - if I really thought hard, I might have come up with Cincinnati, but probably not.

Iowa is +1 over this era, if you're curious. Since we've all got ISU on the brain right now, I'll point out they were at 0.

Human polls are pretty dumb.

 

[ssckelley]

Human polls are pretty dumb.

But they provide entertainment and that's how they should be looked at.

 

[NorthKCHawk]

Preseason polls are a crap shoot, and pollsters are human and human's have biases. Most AP writers have been around for decades and have seen a lot more USC success than Cincinatti.

The polls obviously get more accurate as the season goes on and the final polls are generally spot on.

I actually preferred the old BCS formula that balanced the best computer and human polls. I would prefer that we select playoffs by 50% AP and 50% computers to try and wring out the SEC bias. The coaches poll is stupid. The coaches themselves don't have time to do them, and they would not have time to be informed of teams in other conferences. An SID poll is not a reliable poll.

 

[Fryowa]

Preseason polls are a crap shoot, and pollsters are human and human's have biases. Most AP writers have been around for decades and have seen a lot more USC success than Cincinatti.

The polls obviously get more accurate as the season goes on and the final polls are generally spot on.

I actually preferred the old BCS formula that balanced the best computer and human polls. I would prefer that we select playoffs by 50% AP and 50% computers to try and wring out the SEC bias. The coaches poll is stupid. The coaches themselves don't have time to do them, and they would not have time to be informed of teams in other conferences. An SID poll is not a reliable poll.

Sagarin is the best one out there IMO but it’s a proprietary system, so it would always be accused of having some sort of bias.

They need to go to an open source RPI formula to determine rankings. You can accuse that of bud all day long and even though you may be right, if the formula being used is known by everyone and known from the start of the season, there’s no complaining that can be done.