Home

Comments on the AP College Football Poll: November 14, 2010

1) Oregon 10-0
2) Auburn 11-0
3) Boise State 9-0
4) Texas Christian 11-0
5) Louisiana State 9-1
6) Wisconsin 9-1
7) Stanford 9-1
8) Ohio State 9-1
9) Nebraska 9-1
10) Alabama 8-2
11) Michigan State 9-1
12) Oklahoma State 9-1
13) Arkansas 8-2
14) Virginia Tech 8-2
15) Missouri 8-2
16) Oklahoma 8-2
17) South Carolina 7-3
18) Texas A&M 7-3
19) Nevada 9-1
20) Southern Cal 7-3
21) Iowa 7-3
22) Mississippi State 7-3
23) Arizona 7-3
24) Miami-Florida 7-3
25) Utah 8-2

Others Receiving Votes
Northwestern 7-3
Florida State 7-3
North Carolina State 7-3
Northern Illinois 8-2
San Diego State 7-3
Temple 8-2
Michigan 7-3
Syracuse 7-3
Tulsa 7-3
Maryland 7-3
Delaware 9-1, FCS

If you've been wondering why I haven't been fixing the AP college football poll every week, it's because I don't like to fix the AP poll's top 25 until after the final edition is released in January. When there are no more games to be played, there is no longer any wiggle room for ignoring head-to-head results and season-long performance. But before that, there can be plenty of room for wiggling.

For example, I suppose I could have gotten on voters' cases for not ranking South Carolina ahead of Alabama after the Gamecocks beat the Crimson Tide 35-21 on October 9th, but if the voters believed South Carolina's win to be an upset, who am I to say that they were wrong? The games yet to be played will prove them right or wrong. And as it turns out, they appear to have been right (so far). But we won't really be able to judge one way or another until after all of the games have been played.

However, though I will not fix the AP poll until January, it is late enough in the season that I now feel comfortable commenting on the weekly poll. I will go ahead and point out problems with the poll that would be outright errors if the season had ended last weekend, and no more games were to be played. This can also give voters (in the AP or in any other poll) an idea of potential errors to avoid going forward.

Michigan State

9-1 Wisconsin is #6, 9-1 Ohio State is #8, and 9-1 Michigan State is #11. And it is no coincidence that that is the exact order in which they each lost their one game: Wisconsin losing theirs October 2, Ohio State October 16, and Michigan State October 30.

This is a habit the AP poll would do well to break out of-- there is way too much reliance on when teams lose. Now, I can understand differentiating between a mid-November loss and a mid-September loss, but these three teams all took their loss in October. The exact week it occurred in October should not have such an overriding impact.

What should have an overriding impact is the fact that Michigan State beat Wisconsin 34-24 (MSU and OSU do not play this year). After all, the AP poll asks its voters to pay attention to head-to-head results. It does not ask them to pay attention to when teams took their last loss.

On the other hand, the games yet to be played may well prove voters correct. For example, if Wisconsin wins out, and Michigan State loses their finale to Penn State, then the voters were right all along. Furthermore, if MSU barely beats both Purdue and Penn State in their next two games, while Wisconsin stomps on their next two opponents, then there will be enough of a performance difference to dismiss the head-to-head result and rate Wisconsin higher.

But if Michigan State wins out, and beats at least Purdue by more than a touchdown, then they should be the highest rated Big Ten team headed into the bowls. Unfortunately, they almost certainly would not be. I just don't see the voters suddenly moving the Spartans up that high-- it would be an admission of their own ranking error all these weeks. But better to fix an error than to keep stubbornly making it.

The worst part is, this same illogical ranking in the Harris and coaches' polls could cost MSU a Rose Bowl bid-- as well as a BCS bid entirely.

Texas A&M

Here is another head-to-head issue. #18 Texas A&M defeated #16 Oklahoma 33-19, but this one is a little more complex because Oklahoma is 8-2 and Texas A&M is 7-3. However, Texas A&M has more losses than Oklahoma only because they played Oklahoma State and Arkansas (The AP poll's #12 and #13 teams), and Oklahoma did not (Both Texas A&M and Oklahoma lost to #15 Missouri). So Texas A&M's extra losses are irrelevant.

This is another problem with the AP poll-- focusing too much on straight record and not enough on strength of schedule.

However, the AP voters could be proven right in this case too. If Oklahoma wins out, which would include a win over Oklahoma State (who beat Texas A&M), then maybe Oklahoma will have turned out to deserve a higher ranking than Texas A&M. And Oklahoma winning out certainly does seem quite plausible, so I can't get on voters' cases for ranking Oklahoma higher. Yet.

However, I think it is likely that Texas A&M will deserve to be ranked higher come January (and thus probably deserves it now). Time shall tell the tale.

Arizona

Yet another head-to-head issue: #23 Arizona defeated #21 Iowa 34-27, and both teams are 7-3. This ranking is probably the result of Iowa being ranked much higher than Arizona in the preseason poll. But there are some potentially valid reasons for it. Both teams have one upset loss to an unranked opponent, but Iowa's came to a better team (7-3 Northwestern, rated #26), and on the road. And Iowa defeated #11 Michigan State 37-6. Iowa's performance has been better in the rest of their games too. And Arizona's win over Iowa was close, at home, and came in September.

So maybe there is already ample reason to rank Iowa higher. Once again, I will wait until January, when the dust has cleared, to decide. But I hope voters at least look closely at both if they end up with the same record after the bowl games. Don't vote Iowa higher only because you did in the preseason.

Miami-Florida and Utah

I find these teams, ranked #24 and #25, rather dubious inclusions in the top 25. 7-3 Miami has two losses to unranked opponents, which is pretty ugly for a ranked team in mid-November (though it will be less ugly come January). 8-2 Utah only has 1 such loss, which is more reasonable, but their schedule is very weak (close wins over Pitt and Air Force are all they have going for them, and both of their losses were by huge scores).

Better candidates? I guess I would submit 6-4 Florida, 6-4 Penn State, and/or 7-3 Hawaii. Sure, their records are worse than Miami's and Utah's, but that is because they have played more ranked opponents.

Florida, for example, may be 6-4, but unlike Miami and Utah, they have no losses to unranked opponents (their worst loss is to #22 Mississippi State by 3 points-- whereas Miami and Utah have both been trounced by unranked opponents). But I suppose that we won't really know what to make of Florida until their finale against Florida State (FSU beat Miami 45-17).

6-4 Penn State is not even listed in the "also receiving votes" section, which is unfair because they beat 7-3 Northwestern 35-21, and Northwestern is the top team listed amongst the also-rans. In fact, PSU has defeated three teams that are "also receiving votes"! Sometimes I think AP poll voters stop paying attention entirely when filling out the bottom of their ballots. PSU has one more loss than NW only because they played Alabama, whereas Northwestern played nothing but patsies in nonconference play. AP voters really need to stop rewarding teams for playing weak schedules.

As for 7-3 Hawaii, sure they have 1 more loss than Utah, but what matters is that they only have one loss to an unranked opponent, the same as Utah, except that Hawaii's happened in a close game in September, while Utah's happened in a rout last week. And Hawaii has a win over a ranked opponent (beating #19 Nevada 27-21), whereas Utah hasn't even beaten anyone in the "also receiving votes" section.

Andy Staples' Ballot

This week, Sports Illustrated's Andy Staples submitted perhaps the best post-October ballot I have ever seen in the history of the AP poll. First of all, he got all of the head-to-head cases above (Michigan State, Texas A&M, and Arizona) correct, which is itself remarkable compared to other voters. Plus he did not rank either Miami or Utah! But on top of all that, his willingness to radically reshuffle the teams at the top of his ballot this week, whether he is right or wrong, is very admirable. Mr. Staples is truly an advanced specimen on the evolutionary scale of AP poll voters.

As to his putting Boise State #1, well, it doesn't really matter whether you select Auburn (toughest schedule), Oregon (better performance, weaker schedule), or Boise State (best performance, weakest schedule). They all have their merits. I would love to see Boise State play Oregon and TCU play Auburn at the end of the regular season though, just to see what's what. Because as it stands, if all four win out, we may never get to know what was what. Yet again.

The only quibbles I have with Andy "Hopeful Monster" Staples' top 25 is his inclusion of 7-3 Northwestern and 7-3 North Carolina State. See my comments above on why I think 6-4 Penn State belongs ahead of Northwestern, and as for NC State... they have no more business being ranked than Miami does, as they also have 2 losses to unranked opponents. In fact, I think Miami is probably a better choice than NC State. But we shall soon find out, perhaps as soon as next week.

Mark Anderson's Ballot

Mark Anderson of the Las Vegas Review-Journal has taken a lot of flak this season for his odd ballots, especially early in the season. His ballots looked very different from others because he was using a simple math formula to rank teams. He was doing this as a means of being more objective. However, I don't think the gain in objectivity is worth the loss in accuracy. You can read my general review of math-based college football rankings here, but I will also briefly touch on specific problems I have with Anderson's system.

The problem with any formula, especially a simple one, is that it will make valuations that almost no one would agree with, even the system's creator! For example, Anderson's system will grant 10 points to a team that defeats an 0-8 team 21-20, and it will grant 8 points to a team that loses to an 8-0 team 21-20. Would Anderson himself really agree with this? Beating the worst team in the country by 1 point is better than losing to the best team in the country by 1 point?

How about a more specific example, drawn from this season? Wisconsin gets 19 points for stomping on loser Purdue 83-20. But LSU only gets 18 points for beating top ten Alabama 24-21. Sound fair to you?

Anderson gives a team 8 points just for playing Utah, but only 5 for playing Notre Dame, even though ND stomped on Utah. Likewise, playing Fresno State is worth more than playing Mississippi, though Ole Miss whipped Fresno. Anderson's system equates a 7-3 Sun Belt team with a 7-3 SEC team in terms of "strength of schedule." And that is just wrong.

For all that, Anderson's current ballot isn't really different at this point from a typical AP voter's ballot. But it makes all of the questionable choices I covered at the outset of this article, the same as the AP poll as a whole does.

But the real point is that I don't think any human voter should be using a math formula to rate teams. That's what the hundreds of computer ratings are for. A poll of humans is supposed to be just that: a poll of humans.

Jon Wilner's Ballot

Jon Wilner of the San Jose Mercury News has to be this season's most ridiculed voter at Pollspeak.com. However, I am more interested in lauding good voters than in calling out bad ones, so I am not bringing him up to pile on, but rather to defend him.

In fact, not only do I think that the readers at Pollspeak are wrong to tab him the worst voter, I think he is one of the better voters in the AP poll. After all, he was one of the few voters this week to correctly rank Texas A&M ahead of Oklahoma, as well as Arizona ahead of Iowa, respecting those head-to-head results just like the AP poll's instructions asked him too. Unfortunately, he did place Michigan State well behind Wisconsin, but then, so did practically every AP poll voter.

The main reason people have a problem with him is the fact that he has Auburn #1 and LSU #2. Granted, that makes him odd, and obviously few of us agree with him, but putting LSU #2 is not logically invalid! Auburn at #1 is of course a perfectly legitimate choice, and once that choice is made, well, LSU's only loss came to Auburn, so they are just as valid a choice for #2 as Oregon, Boise State, or TCU.

His bottom teams, San Diego State and Maryland, are pretty stinky, but the same can be said for the bottom teams on most ballots.

Home