Home

Fixing the Final 2015 AP College Football Poll

1) Alabama 14-1
2) Clemson 14-1
3) Stanford 12-2
4) Ohio State 12-1
5) Oklahoma 11-2
6) Michigan State 12-2
7) Texas Christian 11-2
8) Houston 13-1
9) Iowa 12-2
10) Mississippi 10-3
11) Notre Dame 10-3
12) Michigan 10-3
13) Baylor 10-3
14) Florida State 10-3
15) North Carolina 11-3
16) Louisiana State 9-3
17) Utah 10-3
18) Navy 11-2
19) Oregon 9-4
20) Oklahoma State 10-3
21) Wisconsin 10-3
22) Tennessee 9-4
23) Northwestern 10-3
24) Western Kentucky 12-2
25) Florida 10-4

To the left is the final AP poll college football top 25 for the 2015 season. The fixed final AP poll top 25 follows the article below.

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 be inclined to rank Notre Dame ahead of Iowa, and I would rate Florida ahead of Oregon, just as a couple of examples. But those choices the AP poll made are logically viable. This is about fixing only those AP poll choices that are not logically valid or fair.

This was not a good year for the AP poll, far worse than the previous 2 seasons. I'm wondering if this is an effect of the college football playoffs, the voters somehow thinking that the AP poll no longer really matters, and so ranking teams more lazily. But the purpose of the poll was never really to name a national champion, it was to properly rank the top 25 (or 20, or 10 in years long gone).

And that's something they've largely failed at this season.

Alabama tight end O. J. Howard scoring a touchdown against Clemson in the national championship game for the 2015 season

My painting of Alabama tight end O. J. Howard scoring a touchdown in a 45-40 win over Clemson in the national championship game. He had 33 catches for 394 yards and no touchdowns in the previous 14 games, but came out of nowhere to catch 5 passes for 208 yards and a pair of touchdowns in Alabama's biggest game.

Michigan State > Ohio State

I predicted and addressed a lot of the problems with this AP poll last week, and this is the first issue I addressed in those comments. I know that for your average sportswriter, November 20 is like 60 years ago to you or I, but let me remind AP poll voters what happened that day: Michigan State (12-2) went to Columbus and defeated Ohio State (12-1). That is why MSU went to the Big 10 title game and OSU did not, and that is why, after MSU defeated then-unbeaten Iowa, MSU was Big 10 champion and OSU was not, and that is also why MSU was in the playoff and OSU was not.

It really doesn't matter how badly Michigan State lost to Alabama, because Alabama was the national champion, ranked ahead of Ohio State anyway. MSU had already earned a higher rating than OSU by virtue of beating them and thereby going on to win the Big 10. It should also be noted that Michigan State defeated 4 ranked opponents, Ohio State just 2.

The average of the poll points collected by #4 OSU and #6 MSU was 1284.5, and #5 Oklahoma had 1270, so we'll place both teams ahead of the Sooners.

Michigan State #4, Ohio State #5, and Oklahoma #6.

Stanford

I've been saying for weeks that 12-2 Stanford is overrated, and now it's finally time to fix it. Stanford defeated the #9 and #11 teams, but they took 2 upset losses, to teams ranked #19 and #23, and that gave them a relevant record worse than that of several teams ranked behind them. 12-2 Michigan State and 11-2 Oklahoma each took just 1 upset loss, and both defeated 4 ranked opponents, and 12-1 Ohio State took no upset losses. Stanford should be rated behind all three.

#7 Texas Christian (11-2) also had a better relevant record than Stanford did, as TCU took just 1 upset loss, and I would be inclined to rate TCU ahead of Stanford myself, but I think I can give the AP poll voters this much. The reason is that TCU only defeated Oregon in their bowl game (in triple overtime, and thanks to the biggest comeback in bowl history) because Oregon's quarterback was knocked out of the game. Oregon had been running away with the game prior to that. Stanford, on the other hand, faced Oregon's quarterback for the entirety of their 2-point loss to the Ducks. Stanford also outperformed TCU on the season: TCU had 6 close wins, 4 over unranked opponents, while Stanford only had 2 close wins.

Michigan State #3, Ohio State #4, Oklahoma #5, and Stanford #6.

Iowa

11-2 Iowa (#9), whose losses came to top 6 teams, had a better relevant record than did 11-2 Texas Christian (#7) and 13-1 Houston (#8), each of whom took an upset loss. Therefore, let's move the Hawkeyes ahead of both teams. Iowa also had a better relevant record than Stanford did, but Stanford can remain ranked ahead of the Hawkeyes given that they destroyed Iowa in the Rose Bowl, they performed much better for the season, and their big wins were much more impressive than Iowa's.

Iowa #7, Texas Christian #8, and Houston #9.

Notre Dame

Like Iowa, 10-3 Notre Dame (#11) took all of their losses to teams ranked in the top 6, and so they also had a better relevant record than did TCU and Houston, each of whom, again, took an upset loss. As for #10 Mississippi (10-3), while they had a huge win over #1 Alabama, they also took 3 upset losses, and that Alabama win only makes up for one of them.

As we did with Iowa, we'll move Notre Dame ahead of TCU and Houston. I would move Notre Dame up further, ahead of Iowa, myself, but I'll give this one to the AP voters. It's tenuous, but Iowa's big wins were more impressive than Notre Dame's. It should be noted that Notre Dame, whose loss at Stanford was razor-close and on the road, could viably be rated ahead of Stanford, who had a worse relevant record than did Notre Dame despite the head-to-head result.

Notre Dame #8, Texas Christian #9, Houston #10, and Mississippi #11.

Michigan

10-3 Michigan (#12) is also underrated, but not so clearly as Iowa and Notre Dame are, since Michigan did take an upset loss, and those teams did not. Still, as noted, 10-3 Mississippi took 3 upset losses, and their upset win over Alabama only makes up for one of them, so Michigan should be rated higher than them. 13-1 Houston and 11-2 TCU are closer cases, as they also took one upset loss, but Michigan clearly outperformed both of them on the season.

Let's cut to the chase and compare Michigan to 11-2 TCU. Michigan had 2 close wins, while TCU had 6, 4 of them against unrated teams, and that is a huge difference. Furthermore, Michigan's upset loss came in their season opener, by 7 points, at #17 Utah. TCU's upset loss came late in the season, by 20 points, at home against #20 Oklahoma State. Now, when we fix this top 25, Oklahoma State will be ranked significantly higher than Utah, but there is still a big difference here in when the games were played, the sites, and the margins of victory. And Utah is a clear case of a team that was stronger in the first half of the season (6-0, and a 62-20 win over #19 Oregon to go with the big win over Michigan) than they were the second half (4-3, all 3 losses to unranked teams).

Michigan #9, Texas Christian #10, Houston #11, and Mississippi #12.

Baylor

10-3 Baylor (#13) is yet another underrated team, like Michigan taking one upset loss. And like Michigan, they should move past 10-3 Mississippi and their 3 upset losses. Baylor vs. Houston is a tougher case, very close, but I have to move Baylor up ahead of the Cougars because I can't quite see any good reason for Houston to be rated higher. Both teams took their one upset loss late in the season, to unranked, losing opponents, and both defeated 2 rated teams, so they have more in common than not. The difference is that Houston had 3 close wins, all over unrated opponents, and Baylor had just 1. Baylor did also take 2 more losses, to Oklahoma and Texas Christian, but since both were rated ahead of Houston, those games are irrelevant when comparing Baylor to Houston.

Baylor #11, Houston #12, and Mississippi #13.

Florida State

10-3 Florida State (#14) is the last of the parade of teams that moves up past 10-3 Mississippi here, and again, it's because FSU took one upset loss, Mississippi 3. Note that FSU also romped 27-2 over Florida, who romped 38-10 over Ole Miss.

Florida State #13, Mississippi #14.

Oklahoma State

10-3 Oklahoma State (#20) took their losses to teams that are all currently ranked in the top 14, and they notched an upset win over now-#10 TCU (by 20 points), so you know they were almost certainly underrated at #20. They had a better relevant record than all the teams ranked ahead of them up to Baylor, so they could be rated as highly as #12, but that's an extreme, and they're better placed at #15, behind Mississippi, who crushed them 48-20 in the Sugar Bowl.

Let's compare them to #16 Louisiana State (#15 North Carolina will be dropping quite a bit anyway, as covered below, so we'll ignore them for now). 10-3 Oklahoma State defeated 1 rated opponent, while 9-3 LSU defeated 2, but OSU's big win came by 20 against #10 TCU, and LSU's came by 7 against #25 Florida. And LSU's other win against a "rated" opponent came against 12-2 Western Kentucky, who beat no one and will not be ranked in this top 25 when it is fixed. So really, LSU defeated 1 rated opponent, and a lowly rated one at that, while OSU beat a higher-ranked team.

Furthermore, while OSU's losses all came to teams now ranked in the top 14, LSU took an upset loss to unranked 8-5 Arkansas. So OSU's upset win and LSU's upset loss combine to give OSU an effective 2 game lead over LSU for the season, and that is a slam dunk. OSU should be rated higher.

10-3 Mississippi can remain ranked ahead of Oklahoma State, as they beat them badly in the Sugar Bowl, OSU performed poorly down the stretch, and Ole Miss was very hot over their last 3 games (big-margin wins over 9-3 LSU, 9-4 Mississippi State, and OSU).

Move Oklahoma State up to #15, and drop all the teams that had been ranked between Ole Miss and OSU 1 spot each.

Florida > Tennessee > Northwestern > Wisconsin

The AP poll ignored 2 parts of this head-to-head chain. See the "Michigan State > Ohio State" topic above on the inability of sportswriters to remember what occurred on November 20. In this case, they cannot seem to recall that 10-3 Northwestern defeated 10-3 Wisconsin on the road that day. On top of that, Northwestern also had a big upset win over now-#6 Stanford to their credit, while Wisconsin's only win of any note at all came by 2 points over unranked 8-6 Southern Cal (whom Stanford routed twice, by the way). There is simply no question that Northwestern should be rated higher than Wisconsin.

As for 10-4 Florida, they performed very poorly over their final 6 games, and their win over 9-4 Tennessee was far from decisive, 28-27 at home, but Florida had a big 38-10 upset win over now-#14 Mississippi, and Tennessee took an upset loss to unranked 8-5 Arkansas (all 4 of Florida's losses came to teams ranked in the top 16), so Florida was effectively 3 games better than Tennessee on the season. And an effective 3 game lead is too big a difference to ignore on the basis of performance, especially when 1 of those 3 games is a head-to-head decision.

So how about 2 games? #20 Oregon (9-4) also had 1 upset win (Florida has 1 when moved ahead of Tennessee), but Oregon took an upset loss to unranked 9-4 Washington State, giving Florida an effective 1 game lead on Oregon for the season. But I did just say "2 games": Oregon also lost to 10-3 Utah, is currently rated #18. However, Utah is highly overrated, and will be dropped back to #25 in the fixed poll (as detailed next), and so they actually make a 2nd upset loss for Oregon.

I would rate Florida higher than Oregon myself, but I think I can rationalize just enough of an argument for Oregon that the AP poll can keep the Ducks rated higher. Florida's performance was miserable from the end of October on, including a 2-point win over 4-8 Vanderbilt and an overtime decision over 3-9 Florida Atlantic, both games at home, and a 27-2 home loss to Florida State and 41-7 loss to Michigan in the Citrus Bowl. Those are the results of a #50 team at best. Oregon, meanwhile, was red hot over the exact same time period, beating 8-5 Cal 44-28, #6 Stanford 38-36 on the road, and 8-6 Southern Cal 48-28 (USC will be rated in the fixed AP poll, as covered next). And Oregon lost in triple overtime to now-#10 TCU in the Alamo Bowl, the result of a #11 team.

Furthermore, Oregon's loss to Utah came early, to a very different, much-stronger Utah team than the Utes were to become down the stretch (as covered in the "Michigan" section above). It's probable that a loss to the early-season Utah was not an upset loss at all (indeed, that Utah team was likely top 10).

Move Florida to #21, Tennessee to #22, Northwestern to #23, Wisconsin to #24, and Western Kentucky to #25.

Southern Cal, Utah, and Western Kentucky

Here's an unusual case that AP poll voters may be forgiven for missing: 8-6 Southern Cal (way back at #32 in the "Others Receiving Votes" section) should be rated ahead of 10-3 Utah (originally rated #17). As you can see, there is a huge difference in their straight records, but this is entirely an illusion fueled by vastly differing strengths of schedule. USC played 6 ranked opponents, Utah 2, and that is an even greater difference than the number of losses each took.

USC and Utah both went 6-3 in PAC 12 play in the regular season, but USC defeated Utah 42-24, a very decisive result, and that's why USC won their division over Utah and went to the PAC 12 title game. The reason USC took 3 more losses than Utah is because they played PAC 12 champion Stanford twice, and Utah did not play them once, and USC lost by 2 points to a rated team (Wisconsin) in their bowl game, while Utah defeated an unrated team by 7. Those bowl results do not represent a relevant difference between these 2 teams, and USC's games with now-#6 Stanford are irrelevant when comparing USC to Utah, and that is why USC should be rated ahead of Utah.

Furthermore, Utah was simply terrible down the stretch, beginning with their 42-24 loss at USC in game 7. They followed that up with 2 losses to unranked teams (Arizona and UCLA), and in their last 2 games they barely beat another pair of unrated teams (Colorado and BYU).

The average rating of Utah and USC (#24.5) would place them behind the original rating for Western Kentucky. However, both USC and Utah accomplished far more than did Western Kentucky. WKU played 3 major conference teams this season, edging 4-8 Vanderbilt by 2 points and losing 38-35 at 6-7 Indiana and 48-20 to #16 LSU. Those are the results of a team that is not top 25, needless to say, and the rest of their CUSA schedule was worthless. WKU not only failed to beat a ranked team, but they did not even beat a team in the "Others Receiving Votes" section, and WKU definitely should not be ranked.

Utah took 2 losses to unranked teams, but both were major conference teams, and they also notched upset wins over now-#9 Michigan and now-#20 Oregon to make up for those losses. Utah also beat 8-5 Cal and 9-4 Brigham Young, so their accomplishments were so far beyond those of Western Kentucky that there is no legitimate comparison to be made here. In fact, Western Kentucky should not even be rated ahead of 10-3 Georgia, who was sitting just outside the original AP poll at #26, because unlike WKU, Georgia took no upset losses. Georgia also won 31-14 at Vanderbilt, whom WKU beat by only 2 points the previous week. So we'll drop WKU back behind Georgia as well, which will matter as we make our last fix to this poll (the North Carolina section below).

Drop Utah back to #24, move Southern Cal up to #25, and punt Western Kentucky out of the ratings. All the rated teams that had been behind Utah move up one spot (except, needless to say, Western Kentucky).

Navy

11-2 Navy (#18) beat no one (no rated teams, no near-rated teams), their 2 losses, both at rated opponents (Notre Dame and Houston), were routs, and as such they accomplished nothing and are overrated. How far should they drop? Well, I can see Navy being rated ahead of now-#24 Southern Cal (8-6), as USC only beat just one ranked (or near-ranked) team (Utah), and they took an upset loss to unranked Washington. Navy took no upset losses, and furthermore, they had only one close win, against archrival Army, while USC posted 3 close wins, all over unranked opponents.

But Navy does not compare so well to 10-3 Wisconsin (#23). Like USC, Wisconsin only beat one good team (Southern Cal), but unlike USC, Wisconsin took no upset losses. Furthermore, Wisconsin's loss to now-#7 Iowa was very close, whereas Navy was blown out by #8 Notre Dame and #12 Houston. Wisconsin had 2 close wins to Navy's 1, but their better loss evens up the performance race here, and all we are left with is the fact that Wisconsin beat a rated team and Navy did not (even at their original rating of #32, USC is still a better-rated team than anyone Navy beat). There is no good reason to rate Navy higher than Wisconsin, so we won't.

Drop Navy down to #23. Everyone that had been ranked between Navy and Southern Cal moves up a spot.

North Carolina

Finally we have 11-3 North Carolina, whose ratings I have been complaining about for weeks. Originally ranked #15, they are the most overrated team of the 2015 season. North Carolina is another team that hasn't beaten a rated team, and the only "near-rated" team they have beaten is 8-5 Duke, themselves vastly overrated even way back at #35 in the "Others Receiving Votes" section (Duke has 3 losses to teams that are not listed, and they have not beaten any team that is listed).

UNC is a lot like Navy in this regard. However, unlike Navy, UNC took an upset loss, to 3-9 South Carolina, and as such, I can detect no logical reason for UNC to be rated higher than Navy. But then, I can detect no reason for UNC to be rated at all. Let's compare them to 10-3 Georgia, who currently sits at #26 (Western Kentucky having been dropped from any top 25 consideration above).

Georgia, like UNC, beat no one of value. However, unlike UNC, all of Georgia's losses came to rated teams (#1 Alabama, now-#18 Florida, and #19 Tennessee). That makes Georgia effectively 1 game better than UNC for the season. Furthermore, the team that upset UNC, 3-9 South Carolina, was smacked around 52-20 by Georgia 2 weeks later. So we'll drop UNC behind Georgia, which is to say, out of the top 25. But on the bright side, I see no reason UNC cannot remain ranked ahead of Western Kentucky.

Drop North Carolina out of the top 25. All the teams that were rated behind them move up a spot, and Georgia moves into the rankings at #25.

Fixed AP Poll Top 25

11-3 North Carolina (#15) and 12-2 Western Kentucky (#24) fall out of the fixed AP poll top 25, replaced by 8-6 Southern Cal and 10-3 Georgia. UNC and WKU totaled 2 losses to unranked teams, and they beat no ranked opponents. USC and Georgia outdid them in both categories, totaling 1 loss to unranked opponents, and 1 win over AP-ranked teams.

And here it is, your fixed final AP poll for 2015, now logically coherent.

1) Alabama 14-1	--
2) Clemson 14-1	--
3) Michigan State 12-2	+3
4) Ohio State 12-1	--
5) Oklahoma 11-2	--
6) Stanford 12-2	-3
7) Iowa 12-2	+2
8) Notre Dame 10-3	+3
9) Michigan 10-3	+3
10) Texas Christian 11-2	-3
11) Baylor 10-3	+2
12) Houston 13-1	-4
13) Florida State 10-3	+1
14) Mississippi 10-3	-4
15) Oklahoma State 10-3	+5
16) Louisiana State 9-3	--
17) Oregon 9-4	+2
18) Florida 10-4	+7
19) Tennessee 9-4	+3
20) Northwestern 10-3	+3
21) Wisconsin 10-3	--
22) Navy 11-2	-4
23) Southern Cal 8-6	IN
24) Utah 10-3	-7
25) Georgia 10-3	IN

OUT: #15 North Carolina 11-3
#24 Western Kentucky 12-2

Fixed AP Polls
Home