Tip Top 25 in helmets, smaller

Fixing the Final 2006 AP College Football Poll

1) Florida 13-1
2) Ohio State 12-1
3) Louisiana State 11-2
4) Southern Cal 11-2
5) Boise State 13-0
6) Louisville 12-1
7) Wisconsin 12-1
8) Michigan 11-2
9) Auburn 11-2
10) West Virginia 11-2
11) Oklahoma 11-3
12) Rutgers 11-2
13) Texas 10-3
14) California 10-3
15) Arkansas 10-4
16) Brigham Young 11-2
17) Notre Dame 10-3
18) Wake Forest 11-3
19) Virginia Tech 10-3
20) Boston College 10-3
21) Oregon State 10-4
22) Texas Christian 11-2
23) Georgia 9-4
24) Penn State 9-4
25) Tennessee 9-4

To the left is the final 2006 AP college football top 25. The fixed final AP top 25 follows the article below. 

This was the toughest top 25 to fix yet. It involved lengthy bouts of pondering and a whole lot of squinting. There are always a few gray areas: potential fixes that are somewhere between being logically necessary and not. But this top 25 had more gray areas than the others so far.

However, it also had some of the easiest no-brainer fixes yet. Like Arkansas. The teams they lost to were all ranked #7 and higher. They defeated #9 and #25. So where should they be ranked? Not #15, I can assure you that.

Then there's Michigan, who lost only to #2 and #4, and defeated #7. Yet they are ranked #8. Sound right to you? No, not to me either. I mean, it's possible that that could be a logically valid ranking. But very, very unlikely.

Jarvis Moss blocks South Carolina field goal attempt 2006

My painting of the 2006 play of the yearFlorida's Jarvis Moss blocks South Carolina's last-second field goal attempt to preserve the 17-16 win and save us all from watching Ohio State replay Michigan in the national championship game. Looks a whole lot like the 2009 play of the year...


The Michigan-Wisconsin rating fiasco is the first example I gave in the Head-to-Head Results section of my How to Rate Teams guide. Details are to be found there. Suffice it to say, our first fix is moving #8 Michigan ahead of #7 Wisconsin, whom they defeated 27-13. But wait, there's more...

Louisville (#6, 12-1) has a better straight record than 11-2 Michigan, but their loss came to a lower-rated opponent, whereas both of Michigan's losses came to teams rated higher than both Louisville and Michigan. And Michigan defeated more top 25 teams than did Louisville. So move the Wolverines ahead of Louisville too. But wait...

Boise State (#5, 13-0) really has a better straight record than Michigan. But since Michigan's only losses came to teams rated higher than Boise State, for comparison purposes, it is effectively as though Michigan were 11-0. And as with Louisville (and Wisconsin), Michigan defeated more top 25 opponents than did Boise State. Michigan also defeated a top ten opponent, and by 2 touchdowns. While Boise State's biggest win came in overtime against #21 Oklahoma-- yes, that is #21. That's where they'll end up after I've fixed this top 25. Sorry for the spoiler. And that will be the only top 25 opponent Boise ends up with.

Then there is performance. Boise State had 3 close wins (touchdown or less) over unrated opponents, and Michigan had but one. Frankly, nothing here points to Boise State. So move Big Blue ahead of Little Blue. But wait...

Michigan vs. Southern Cal

Sure, Southern Cal laid the smack down 32-18 on Michigan in the Rose Bowl, but USC also lost to two lower-rated teams. Michigan was effectively two games ahead of USC coming in, and even the Rose Bowl is, after all, just one game. So do we move Michigan ahead of USC? "Not so fast, my friend!" I think we all agree that late season games in general, and bowl games in particular, should be weighted more heavily than other games. The question is: How much more? Is a bowl game worth two regular season games?

This we do not all agree on, but for myself, I would say in general, no. A bowl game should be weighted more heavily, but should not as a rule be counted twice. But a convincing win in a bowl game, like the Rose Bowl was, comes pretty close to being worth 2 games. Add in the fact that USC defeated more top 25 teams, and beat all of them by at least 2 touchdowns, and things look good for the Trojans. Then there's the fact that one of USC's upset losses came in a rivalry game (UCLA). Now, a loss is a loss, but an upset loss to a rival is not so damaging as an upset loss to a non-rival.

Which all goes to say that Southern Cal can stay ahead of Michigan. So: Michigan #5, Boise State #6, Louisville #7, and Wisconsin #8.


I discussed the Arkansas 2006 rating injustice throughout the Record section of my How to Rate Teams guide. So let's jump right to the real question: should 10-4 Arkansas (#15) be moved up ahead of 11-2 Auburn (#9) just because Arkansas beat them 27-10? The answer is no. Oh, they should be moved ahead of Auburn, but not "just because" they beat them. After all, 7-6 UCLA beat 11-2 Southern Cal in 2006, but no one would suggest moving UCLA ahead of the Trojans.

Auburn 2006 was one of those strange cases of a team that could be the best in the country at one time, then completely mediocre another. They defeated #1 Florida and #3 Louisiana State, which is pretty amazing. But that's about all they did. Those wins are balanced out by Auburn's 37-15 upset loss to Georgia and that 27-10 loss to Arkansas. Arkansas lost to Florida and LSU, but other than beating Auburn, they also beat Tennessee 31-14 (and Tennessee beat Georgia 51-33). Arkansas and Auburn both had 2 SEC losses, but Arkansas won the West division on their head-to-head win.

Arkansas had 2 more losses than Auburn because they played #4 Southern Cal in their opener and #7 Wisconsin in their bowl game, whereas Auburn played no rated nonconference opponent at all. Performance-wise, Arkansas was terrible in September, but consistently played like a top ten team through their last 10 games, even in their losses. Auburn, on the other hand, played like a #25-30 type team down the stretch. In their last 5 games, they lost by 22 to lower-rated Georgia, and they won 3 games over unrated opponents by a touchdown or less.

The two teams' bowl games are good examples of their power level at season's end. Arkansas lost their bowl and Auburn won theirs, but don't be fooled by a mere L or W. Arkansas lost by 3 points to #7 Wisconsin. That is the performance of a team that is nearly as good as a #7 team. Auburn won their bowl game by 3 points over unrated Nebraska. That is the performance of a team that is barely better than an unrated opponent. So for that one game, Arkansas played like #10, and Auburn played like #25.

Auburn has a better straight record, but the two teams have the same relevant record, with Arkansas getting the edge due to their head-to-head win. And it was a decisive win: 27-10 at Auburn. Arkansas performed better, and more importantly, did it down the stretch: in the bowl game and last month of the season. So we'll move Arkansas up from #15 to #9, and drop Auburn to #10. The teams that were #10-14 all drop a spot too, of course.

Notre Dame, Penn State, and Tennessee

Twelve of the teams originally ranked #10-#25 lost to at least one unranked opponent. 10-3 Notre Dame (#17), 9-4 Penn State (#24), and 9-4 Tennessee (#25) are 3 of the 4 teams who did not lose such a game, the other being Arkansas. Notre Dame beat PSU 41-17, and PSU beat Tennessee 20-10 in their bowl game. And Tennessee beat two teams the AP poll has ranked above them: 10-3 California (originally #14) by a score of 35-18 and 9-4 Georgia (#23) 51-33. Those are pretty decisive outcomes.

Cal also lost to unranked Arizona 24-20. And when this top 25 is fixed, they will not have a single win over a ranked opponent. So I see no justification for their being rated ahead of Tennessee. And if we move Notre Dame, Penn State, and Tennessee up ahead of California, all three teams still have no losses to lower-rated opponents. As such, the move makes perfect sense, so that is what we'll do.

Notre Dame moves to #15, PSU to #16, and Tennessee to #17. Cal drops to #18, and everyone else behind them also drops to make room for the shift.

Georgia and Virginia Tech

Georgia (9-4, originally #23) beat Virginia Tech (10-3, #19) 31-24 in their bowl game. Virginia Tech has a better straight record, but they do not have a better relevant record. Their schedule was very weak. This is actually a much clearer case than USC-Michigan. How can you dismiss a relevant head-to-head bowl result, AP voters? Move Georgia up one spot ahead of Virginia Tech.

Furthermore, Virginia Tech had been ranked one spot behind 11-3 Wake Forest, despite winning at Wake 27-6. Wake Forest was the ACC champion, but they had 2 ACC losses, the same as VT. Their straight records and relevant records are equal, and VT's win over them was very decisive, so Wake should drop back behind VT.

There's been a lot of shifting around, and at this point we have Georgia #20, VT #21, WF #22, Boston College #23, Oregon State #24, and Texas Christian #25.

South Carolina and Kentucky

Except for their game against each other, 8-5 South Carolina and 8-5 Kentucky did not lose to any unrated opponents. South Carolina beat Kentucky, and Kentucky beat now-#20 Georgia. So both of these teams will be moving into the top 25.

At this point, you may have noticed that most of the fixes thus far have involved moving SEC teams up. There's a reason for that. The SEC went 47-10 against nonconference teams in 2006, which was the 3rd best such record for any conference since WWII ended in 1945. The AP poll likes to spread the wealth, and that is nice, very sharing, but sometimes one conference really is just that much better than the others. SEC teams have more losses than some of the non-SEC teams around them in the top 25, but obviously that is because they were playing in the SEC. Of course they had more losses.

South Carolina's five losses all came to top 25 SEC teams. The only one that wasn't close was their 18-0 loss to Georgia in their second game. They lost to Auburn (now #10) by 7, Tennessee (#17) by 7, Arkansas (#9) by 6, and Florida (#1) by 1 point. Those last four may be losses, but they are all top 25-worthy performances nevertheless. Their last loss, by 1 point at national champion Florida, was particularly impressive. They only lost because they had an extra point blocked, then a field goal attempt blocked with 8 seconds left in the game. SC beat Kentucky, 8-5 Clemson, and 10-3 Houston.

Kentucky's performance was much more erratic, but with no upset losses, and an upset win over Georgia, Kentucky is effectively 3 games ahead of Virginia Tech, for example, despite VT being 2 games better in straight record. That is because Virginia Tech has 2 losses to lower-rated teams, and no upset wins.

So we'll move SC and Kentucky in between Georgia and Virginia Tech. SC comes in at #21, Kentucky at #22, VT drops to #23, WF to #24, BC to #25, and Oregon State and TCU drop out of the top 25.

Brigham Young

Why am I always picking on Brigham Young? Originally ranked #16, BYU was 11-2, but their 2 losses were to unranked Arizona and now-#25 Boston College, and now that TCU has dropped out, they have no wins against a top 25 opponent. TCU was 11-2, but with no wins of any real value, they never should have been ranked in the first place. So should BYU be dropped back to #25, behind BC?

Not necessarily. That loss to BC came in Boston, in the third game, in overtime. Not decisive. In fact, one could reasonably conclude that the game would have gone differently on a neutral field. Ranking BYU behind BC would be a valid choice, but you could go either way on it.

BYU's other loss came in their first game. They won their last 10 games, all but one by huge scores. That one was a 33-31 win over Utah, but it being a rivalry game, the poor performance is forgivable. They beat TCU by 2 touchdowns, and though TCU is no longer in the top 25, they are nearly rated. They romped on 7-6 Oregon 38-8 in their bowl game, and though Oregon is not close to being rated, the point is that BYU stomped them like a top 25 team should.

Oklahoma (originally #11), on the other hand, lost to Oregon. And Texas (#13) beat an even worse Iowa team in the Alamo Bowl 26-24. But I'll be getting to those two overrated miscreants next.

BYU's schedule was weak, but their performance was very strong. This makes them a difficult team to rate, and I have spent a lot of time considering how much leeway I should give AP poll voters here. My first inclination was to place them behind 8-5 Kentucky and ahead of 10-3 Virginia Tech. Kentucky had no losses to unrated opponents, but Virginia Tech, like BYU, had 2 losses to lower-rated opponents (and both lost to BC, though VT's loss was much worse: 22-3).

Brigham Young vs. Kentucky

Kentucky's wins over Georgia and 8-5 Clemson are also worth more than BYU's wins over TCU and 7-6 Oregon. However. BYU's upset losses did come early, and both were very close road games. And their performance was vastly better than Kentucky's. Kentucky got stomped in 3 of their losses, and struggled to beat a couple of bad opponents. They beat Mississippi State by 3, and worse, barely won against Louisiana-Monroe 42-40 in their 11th game!

When I'm fixing these AP polls, I try my best to give AP poll voters what they want. Within reason. Sometimes, though, I think of them as toddlers crying for cookies. And if you are a parent, you know you cannot just give your toddler a cookie every time he cries for one. You probably know that even if you are not a parent. But you also know that sometimes it is nice to go ahead and hand him a cookie.

So I will compromise with the AP voters here and place BYU into a tie with Kentucky. That is the best that I can do, fellas. And that moves Georgia to #19, SC to #20, and Kentucky now shares #21 with BYU.

Oklahoma and Texas

Now for the "overrated miscreants." 11-3 Oklahoma (originally #11) and 10-3 Texas (#13) are also very hard to rate, largely because, like everyone else, I know they were better than they played. Yet for all that talent, Oklahoma beat Alabama-Birmingham by only 7, lost to unrated Oregon, lost to lower-rated Texas by 18 points, beat a couple of unrated opponents by 1 and 6 down the stretch, and in the end, they don't have even one win over a top 25 opponent. Does that sound like a #11 team to you? Yeah, me either.

Then there's Texas. Through 10 games they were 9-1 and looked like a very solid top 10 team, with a big win over Oklahoma and their only loss coming to #2 Ohio State. They did have a couple of poor road performances in close wins over unrated teams, but in most of their games they won by huge scores. Then, out of nowhere, they collapsed down the stretch. First they lost to unrated Kansas State 45-42. Then they lost to unrated Texas A&M 12-7, knocking the Longhorns out of the Big 12 title game. Then they struggled to beat a bad 6-7 Iowa team in the Alamo Bowl 26-24.

The last result is the worst. With the extra time to prepare and heal up, I expected far better than that. One can only conclude that in the end, they just weren't very good. Texas' only win over a rated opponent is their 28-10 win against overrated Oklahoma. These teams should drop. But how far?

Oklahoma vs. Texas

Before we get to how far they drop, perhaps you're wondering if Texas should be rated ahead of Oklahoma, having beaten them. That is an option, since their records (both straight and relevant) are equal. Texas had one more Big 12 loss, but Oklahoma made up for that by losing to Oregon. I myself would rate Texas higher. But Oklahoma being higher is an option too, due to Texas' collapse down the stretch. Oklahoma, on the other hand, performed much better in their last month.

Take the bowl results. An overtime loss to a top ten team (as in Oklahoma's famous Fiesta Bowl loss to Boise State) is a far better performance than a 2 point win over a #55 type team. Then there's the fact that Oklahoma's loss to Oregon came only because of the worst replay call in the history of college football's use of instant replay. The replay official was suspended afterward.

So the AP voters can keep Oklahoma rated higher. But as I said, they both need to drop...

But How Far?

So Oklahoma has 2 losses to unranked teams, and no wins over ranked teams. The teams behind them have much better records than that. The first team behind them with a comparable record is BYU, currently tied with Kentucky at #21. I can't see putting Oklahoma ahead of Kentucky, who had no upset losses and a win over now-#19 Georgia. But what about BYU? Should I take the AP poll voters' cookie away?

Their relevant records are the same. Oklahoma may have played a tougher schedule, but it wasn't much tougher. The Big 12 was lousy in 2006, going 36-20 against nonconference opponents, and worse, 5-14 against winning opponents and 0-11 against rated opponents. Yikes. And BYU's performance was much better, with only one of their wins being close. And it's hard not to look at BYU beating Oregon 38-8, and Oklahoma losing to the same team. Obviously BYU did not have to worry about a bad replay call.

So drop Oklahoma in behind Kentucky and Brigham Young, and Texas just behind Oklahoma, and ahead of Virginia Tech (who also had 2 losses to lower-rated opponents).

Fixed AP Top 25

That was a tough one. My brain is plain tuckered out, my eyes bleary and bloodshot. Was it worth it? Taste the proof in this pudding: the original top 25 teams in the AP poll had 31 losses to lower-ranked teams, while the fixed top 25 has only 20 losses to lower-ranked teams. A pretty huge difference. We could cut that down more by dropping BYU behind BC and Oklahoma behind Texas, but those AP choices are valid. Reducing the number of losses to lower-ranked teams is not the only goal of fixing these AP polls. It's just a good general measuring stick.

My own top 25 would look different. For one thing, I'd drop Boise State at least 5 places, and probably more than that. Boise State was a tough case when fixing the AP poll, but ultimately, I decided the voters had an argument for them, however weak. So the Broncos fell just one spot.

Oregon State and Texas Christian drop out, and South Carolina and Kentucky come in. The old teams had 1 win against a rated opponent, but 3 losses to unrated opponents. The new teams also have 1 win against a rated opponent, but they have no losses to unrated opponents. I am going to shake my own hand on a job well done now.

1) Florida 13-1--
2) Ohio State 12-1--
3) Louisiana State 11-2--
4) Southern Cal 11-2--
5) Michigan 11-2+3
6) Boise State 13-0-1
7) Louisville 12-1-1
8) Wisconsin 12-1-1
9) Arkansas 10-4+6
10) Auburn 11-2-1
11) West Virginia 11-2-1
12) Rutgers 11-2--
13) Notre Dame 10-3+4
14) Penn State 9-4+10
15) Tennessee 9-4+10
16) California 10-3-2
17) Georgia 9-4+6
18) South Carolina 8-5IN
19) Kentucky 8-5
       Brigham Young 11-2
21) Oklahoma 11-3-10
22) Texas 10-3-9
23) Virginia Tech 10-3-4
24) Wake Forest 11-3-6
25) Boston College 10-3-5

OUT: #21 Oregon State 10-4
#22 Texas Christian 11-2

Fixed AP Polls