Home

Fixing the Final 2009 AP College Football Poll

1) Alabama 14-0
2) Texas 13-1
3) Florida 13-1
4) Boise State 14-0
5) Ohio State 11-2
6) Texas Christian 12-1
7) Iowa 11-2
8) Cincinnati 12-1
9) Penn State 11-2
10) Virginia Tech 10-3
11) Oregon 10-3
12) Brigham Young 11-2
13) Georgia Tech 11-3
14) Nebraska 10-4
15) Pittsburgh 10-3
16) Wisconsin 10-3
17) Louisiana State 9-4
18) Utah 10-3
19) Miami-Florida 9-4
20) Mississippi 9-4
21) Texas Tech 9-4
22) Southern Cal 9-4
23) Central Michigan 12-2
24) Clemson 9-5
25) West Virginia 9-4

Here is the final AP poll college football top 25 for the 2009 season. The fixed final AP poll top 25 follows the article below.

The voters did a pretty good job this season compared to most years: congrats to them on that. However, this ranking is still broken in a few places, so let's fix it.

But before I get to correcting the AP poll's errors, let me make one thing clear. This is not about what I would personally prefer to see in the rankings. I myself would probably rank Texas around #5, drop Nebraska to about #20, and knock Texas Tech out entirely. And I probably wouldn't rank Mississippi either. But all those choices the AP poll made are fine-- logically valid choices. This is about fixing only those AP poll choices that are not logically valid or fair.

Terrence Cody blocks Tennessee field goal attempt 2009

Congratulations to the Alabama Crimson Tide, 2009 national champions! Or "national co-champions" to some Boise State fans. This is my painting of what I think of as the 2009 play of the year: Alabama defensive tackle Terrence Cody blocking a field goal to preserve a win against Tennessee. Without that play, Alabama... well, they still would have won the national championship. But hey, not every play of the year can be Vince Young scoring on 4th down in the final seconds of the national championship game.

Georgia Tech

Speaking of "not logically valid or fair," how did ACC champion Georgia Tech (11-3) wind up ranked three places behind a 10-3 ACC team they defeated in the regular season? Unfortunately, this happens nearly every year in the final AP poll. One team wins their conference, then plays a BCS bowl and loses to a top ten opponent. Another team, from the same conference and who lost to the conference champ, goes to a minor bowl and defeats a lowly ranked or unranked opponent, then passes up the conference champ in the final poll. It is irrational, and it would be nice if voters could put an end to this bad habit.

Georgia Tech was ranked #9 and Virginia Tech #12 going into the bowls. Georgia Tech lost to #10 Iowa, while Virginia Tech defeated unranked 7-6 Tennessee. Neither of those results should logically push Virginia Tech past Georgia Tech. Iowa was ranked higher than Virginia Tech before the bowls and still is after the bowls. Obviously, the voters think Iowa is better than Virginia Tech as well as Georgia Tech.

Georgia Tech is the ACC champion. They defeated Virginia Tech on the field of play. Yet forever and always, Virginia Tech will sit three places ahead of Georgia Tech in the final 2009 AP poll. So that will be our first fix. Virginia Tech drops back behind Georgia Tech (Oregon now #10, BYU #11, GT #12, VT #13).

Texas Christian

Georgia Tech 2009 is why the AP poll asks its voters to pay attention to head-to-head results. Because they don't. Most of the fixes to most final AP polls involve accounting for head-to-head results. But for our next fix, let's look at a problem that does not involve ignoring a head-to-head result: 11-2 Ohio State being ranked ahead of 12-1 Texas Christian.

All you have to do is look at their losses. TCU lost only to #4 Boise State (ranked higher than Ohio State). Ohio State lost to a bad unranked team (Purdue) and to the AP poll's #22 team (Southern Cal). TCU is effectively 2 games ahead of Ohio State right there. To make up for that big a difference, Ohio State would have to have played a massively tougher schedule than did TCU. And Ohio State did defeat 4 ranked teams, which is a large amount. But TCU defeated 3.

I understand that the Big Ten had a strong bowl season, going 4-3 with the four wins coming against rated teams. But the Mountain West did well too, going 4-1. I also understand that Ohio State became much stronger in the second half of the season (after the Purdue loss), winning their last 6. But TCU was much stronger down the stretch too. In their last 7 regular season games, the closest anyone got to TCU was 27 points, and that stretch includes two top 25 opponents. They did lose their bowl game, but that was a close loss to an unbeaten team that is rated higher than Ohio State anyway.

This is a no-brainer. TCU moves to #5, Ohio State to #6.

Arizona

And for our next few fixes, it's back to head-to-head results. First up, unranked Arizona (8-5), who defeated #22 Southern Cal (9-4). Actually, Arizona defeated a whopping four teams that are ranked ahead of them in the AP poll (counting the "also receiving votes" teams). That is one harsh case of Lastgamitis-- a disease that causes poll voters to throw out an entire season and only look at a team's last game.

And Arizona's last game was indeed ugly, a 33-0 Holiday Bowl loss to #14 Nebraska. But it's not as if USC didn't have some ugly results of their own: a 47-20 loss to Oregon in game 8 and a 55-21 loss to unranked Stanford in game 10 (much uglier than Arizona's Holiday Bowl loss, given where Stanford and Nebraska are ranked). Arizona beat Stanford and lost to Oregon in overtime. The reason Arizona has one more loss than USC is because of that bowl loss to Nebraska, who is ranked 8 places higher than USC anyway, and because USC defeated an unranked Boston College in their bowl game.

Arizona is the second place team in the PAC 10, a placement they sealed when they beat USC on the road, in the last game of the regular season, 21-17. They were ranked ahead of USC before the bowls, and now they are essentially being punished for beating USC and thereby drawing a far more difficult bowl opponent than USC had to play.

So we'll fix that by moving Arizona into the poll at #22, dropping Southern Cal and the other teams behind them one place each. Incidentally, Arizona also defeated Central Michigan, whom the AP Poll had sitting right behind USC at #23, by a score of 19-6.

Oklahoma, Oklahoma State, and Texas Tech

Next up is the problem of 9-4 Texas Tech, whom the AP poll had at #21. This is a problem because they lost to 9-4 Oklahoma State (unranked, but #30 in the "also receiving votes" section). And Oklahoma State lost to 8-5 Oklahoma (unranked, just outside the poll at #26). To make the circle complete, Oklahoma lost to Texas Tech. So are they all equal?

Not quite. Texas Tech and Oklahoma State were both upset by Houston, who is unranked and did not get a poll point from a single ballot. Whereas Oklahoma's other four losses are all to teams who are already ranked higher than any of these three teams (Texas, Nebraska, BYU, and Miami: and none of those games were at home!). Look at it this way: if Oklahoma had played two patsies rather than BYU and Miami, they would be 10-3, and this would not even be an issue. This is another case of the AP poll punishing a team for playing a tough schedule.

Due to the two Houston upsets, Oklahoma is effectively a game ahead of both Texas Tech and Oklahoma State (it should be noted that Texas Tech did get a big win at #14 Nebraska, but that is canceled out by their getting drubbed at home by 6-7 Texas A&M), so move Oklahoma ahead of both. And that leaves Texas Tech and Oklahoma State with the same straight record (9-4) and the same relevant record, so the head-to-head tiebreaker gives Oklahoma State the nod.

Therefore... put Oklahoma in at #21, Oklahoma State at #22, and drop Texas Tech to #23 (and that drops Arizona to #24 and Southern Cal to #25). As I said earlier, I wouldn't personally put Texas Tech into my own top 25. But if the AP poll likes Texas Tech that much, it necessitates moving Oklahoma and Oklahoma State in as well. And I never even had to get to the fact that Texas Tech's best nonconference win came against 6-7 Michigan State, whereas Oklahoma beat 8-5 Stanford (AP #29) and Oklahoma State beat 8-5 Georgia (AP #33).

Mississippi

So now Oklahoma, with one upset loss (to #23 Texas Tech), sits just behind Mississippi, who has three upset losses to unranked teams and one upset win over #17 LSU. Effectively, that makes Oklahoma one game better than Ole Miss. Oklahoma's one upset loss was ugly (41-13), but again, Texas Tech was ranked, and Mississippi lost twice to unranked teams by more than a touchdown, including a 14 point loss to 5-7 Mississippi State in their regular season finale. And the rest of the games for both teams give no advantage to Mississippi to make up for their deficit. Ole Miss played perhaps the weakest nonconference slate in the FBS (Memphis, SE Louisiana, UAB, Northern Arizona).

So move Oklahoma ahead of Mississippi, to #20, and Mississippi down to #21.

Pittsburgh

I also have a problem with #15 Pittsburgh. They have a very impressive straight record of 10-3, but they lost to 5-7 North Carolina State and to West Virginia, who was #25 and has now fallen out. And Pitt doesn't have a single win over a rated opponent. My own inclination would be to drop Pitt below West Virginia, who beat them. But West Virginia did only beat them by 3 points at home. Not decisive.

And Wisconsin, who sits behind Pitt, may have only one upset loss (to Pitt's two), but they also only have one win that holds any value at all for a #16 team (their bowl win over Miami). Pittsburgh, on the other hand, has a big five wins over good-but-unrated opponents (10-4 Navy, 8-5 Connecticut, 9-4 Rutgers, 8-5 South Florida, and 8-5 North Carolina). Wisconsin also performed poorly, beating five mediocre-to-poor opponents by a touchdown or less.

So Pitt can stay. Similarly, there are other teams that seem to me to be overrated or underrated, but an argument can be made for them, so I'll let the AP poll voters keep such teams as is. I just wanted to use Pitt as an example of the many such cases I looked hard at.

Fixed AP Poll Top 25

And that wraps it up. Oklahoma, Oklahoma State, and Arizona are now in. Central Michigan, Clemson, and West Virginia fall out. And if you really look at those teams, it makes a lot of sense.

Central Michigan does not have a single win of any real value (and no, a 2-point win over 6-7 Michigan State and an overtime win over Troy do not hold value). And they were badly beaten by 8-5 Boston College in their 9th game-- I would rank them behind BC, well out of the top 25. West Virginia lost to 7-6 Florida State by 2 touchdowns in their bowl game, and got trounced by unranked South Florida and Auburn in the regular season. I don't know what they were doing in the top 25 in the first place. Clemson is the best of the three, with a win over Miami and close losses to Georgia Tech (twice) and TCU. But they also lost to 2-10 Maryland and by 17 points to 7-6 South Carolina in their regular season finale.

In all, the 3 teams that fall out originally had seven losses to unranked teams, whereas the 3 teams now in have only three losses to unranked teams. Plus the newly ranked teams are 1-0 against the old ones head-to-head (Arizona's 19-6 win over Central Michigan). So the AP poll has been greatly strengthened by these fixes, and not just because the teams that come in fared better overall than the teams that fell out. A ripple effect from these fixes strengthens other areas of the poll that had also been questionable.

Take #14 Nebraska (10-4). In the original poll, they look highly questionable, with a loss to unranked 7-6 Iowa State, another to #21 Texas Tech, and not a single win over a ranked opponent. But with Oklahoma and Arizona in, Nebraska has two such wins, and now they make more sense at #14.

Then there's #2 Texas, whose case against both Florida and Boise State was weak due to having just two wins over ranked opponents, and not very highly ranked at that. But with Oklahoma and Oklahoma State now in, Texas has four wins over ranked opponents, strengthening the AP's case for putting them at #2.

And 11-2 Iowa was ranked ahead of 12-1 Cincinnati even though Iowa has an upset loss (to unranked Northwestern) and Cincinnati does not. Iowa's schedule looked stronger at the top, if not overall. But West Virginia falling out leaves Cincinnati with only one win over a ranked opponent, strengthening the AP poll's case here too (though I myself would still rate Cincinnati higher due to their playing more good teams overall, and Iowa performing poorly in a number of games against bad opponents).

So you see, these few fixes have vastly increased the structural integrity of the entire AP top 25. And now, without further ado, I present to you--fully edited and logically coherent--your fixed 2009 AP college football top 25!

1) Alabama 14-0	--
2) Texas 13-1	--
3) Florida 13-1	--
4) Boise State 14-0	--
5) Texas Christian 12-1	+1
6) Ohio State 11-2	-1
7) Iowa 11-2	--
8) Cincinnati 12-1	--
9) Penn State 11-2	--
10) Oregon 10-3	+1
11) Brigham Young 11-2	+1
12) Georgia Tech 11-3	+1
13) Virginia Tech 10-3	-3
14) Nebraska 10-4	--
15) Pittsburgh 10-3	--
16) Wisconsin 10-3	--
17) Louisiana State 9-4	--
18) Utah 10-3	--
19) Miami-Florida 9-4	--
20) Oklahoma 8-5	IN
21) Mississippi 9-4	-1
22) Oklahoma State 9-4	IN
23) Texas Tech 9-4	-2
24) Arizona 8-5	IN
25) Southern Cal 9-4	-3

OUT: #23 Central Michigan 12-2
#24 Clemson 9-5
#25 West Virginia 9-4

Fixed AP Polls
Home