Posts Tagged ‘football’

In Defense of Preseason Rankings

In College Football, General LSU, Post-game, Rankings Commentary on September 4, 2016 at 12:25 PM

I never claimed projections, especially preseason ones, as my strong suit, but a lot of smart people who spent more than a couple of hours before they make their projections picked LSU and/or Oklahoma to make the national semifinal. Sometimes I project the final rankings better than the professionals, and sometimes I don’t. I think I’ve done a pretty decent job over the years for the focus that I give to it.

I’ve always had a greater motivation to give teams proper credit for what they’ve done, and I’ll strive to do that for the wins over LSU, Oklahoma, and Bowling Green. (Those are the three results where there is the most apparent discrepancy with my preseason rankings so far.)

I actually spent more time in the off-season looking at my formula (so that later this year I do give teams the rankings they deserve) than I spent looking at anything to do with preseason.

I expected LSU to have a close game against Wisconsin in the state of Wisconsin. I just didn’t expect them to get into winning-field-goal position and for the quarterback to inexplicably throw It to the wrong team. I also didn’t anticipate that Wisconsin would hit two long field goals, the second of which was to take the lead in the fourth quarter. If they miss the second one, there is little doubt that LSU wins.

The Wisconsin defense won't get many easier interceptions this season.

The Wisconsin defense won’t get many easier interceptions this season.

Although I did not rank Wisconsin at the end of last season, I ranked them at the beginning of this season because I believe they can play good teams, especially close to home, and have a chance to win if said good team chokes. I don’t think they’ll beat everyone, but I would be surprised if they didn’t beat another ranked team at some point. I moved Wisconsin up more spots than I moved LSU in the preseason. Also, I only ranked LSU 3 spots higher than the AP poll did. I didn’t do anything crazy there.

As for Houston, 18th is pretty high for a preseason ranking for a non-power-5 team. I was surprised that the Houston defense took control late rather than the Oklahoma offense. Usually you expect a good offense to wear down a suspect defense, but maybe the Cougars are going to be a more balanced team than in past seasons. Absolutely no one should be surprised Houston scored 33 points (although the special teams is responsible for one touchdown), but holding Oklahoma to 23 was a bit surprising.

Regarding Ohio St., which got the most immediate push-back, I did want to say a couple of things about not buying in to certain preseason considerations that are present in other polls and rankings. This is typical of my preseason outlook of only seeing a team as worthy of continuing in the top 10 if they have most of their key players back.

Although Florida St. had about twice as many returning starters last year as Ohio St. does this year, I still got criticism for dropping the Seminoles from #2 (where I had them at the end of 2014) to #12. They finished #14 in both polls (and even lower in my rankings, obviously), so I was actually conservative in demoting them. Like Ohio St., the Seminoles had won the national championship two years before. Unlike Ohio St., they had an undefeated regular season and made the national semifinal the prior year.

Even if they’d rallied to end the year in the top 10, my skepticism in preseason was still warranted by the level of play early on. I try to accurately reflect how tough a team it is now or at least in the near future more than I try to look into the crystal ball to predict what might happen in late November and afterward, but often they mean the same thing. If you have a lack of experience now, that will in most cases plague you throughout the year because for most teams the experienced players will get even better, so it’s hard to surpass them.

I also think the pushback is a function of Saban’s success at Alabama, though he’s generally had 11 or 12 returning starters, which is a a lot more than 6. Everyone thinks their team should be able to be really good yet again if they were good last year. Never mind how much the final top 10 changed from one year to the next. Urban Meyer is a great coach, don’t get me wrong, but even great coaches have years with a few losses. Half of his Florida teams had at least three losses and a third of them had at least four.

Urban Meyer's results at Florida.

Urban Meyer’s results at Florida.

Speaking of Florida, I dropped them to #13 in the 2013 preseason when they only had 10 returning starters after finishing the regular season with one loss the year before (they lost the bowl game, but not showing up for the bowl game the year before didn’t stop Alabama in 2009 or 2015 (it also didn’t stop the Tide from making the national semifinal in 2014). Anyway, I should have dropped that 2013 Florida team a lot more than that since they finished 4-8.

I’m not saying anything like what happened to Florida in 2013 will happen to Ohio St. this year. I’m only picking one Ohio St. team to have a few losses.

The only real argument I got in response was that Urban Meyer is a really good coach and recruiter. In 2010 (when they had 10 returning starters according to Phil Steele), even Alabama had 3 losses. That was Saban’s fourth year there, so there isn’t some other coach to blame for that. There is just only so much even the best coaches with even the best recruits can do with raw talent in the offseason.

Maybe Ohio St. better talent than some of these other examples, maybe they’ll be really good at avoiding injuries. Maybe they’ll have a couple injuries, lack depth, and start having trouble. All this is about is assessing likelihoods. With 128 teams playing 12 or so games apiece, a lot of unlikely things are going to happen.

The final score in the Bowling Green game doesn’t mean much to me. Urban Meyer always liked to run up the score. There is absolutely no reason in a game like that to score 42 points in the second half. I don’t think Louisville scoring 70 means they should be in the top 10 either, and I also didn’t think that when Boston College scored 76 in a game last year that they deserved a ranking.

I’m not making a prediction at this time, but I’ll be very interested in the outcome of Ohio St.@ Oklahoma in a couple of weeks.

I’m sure there will be examples where I made better picks than the experts and examples where I made worse picks. The decision not to rank UCLA, for instance, is looking pretty good at the moment. Also, I was 9-1 against the spread in SEC games. Of course I would have preferred to go 1-9 with LSU winning, but that’s life.

Addendum: Review of 2014 and 2015 Preseason Rankings

I didn’t do a blog after last season about it (I was busy writing the NFL blogs, and then I just got onto college basketball and didn’t think about it again).

I think I did a good job when I reviewed the results after 2014. I got 3 of the top 4 right in preseason that year. I also had two other top 10 teams who ended up in the top 10, Ohio St. and UCLA. I got the exact rank correct for UCLA.

No one (by no one I mean the other major preseason listings I compared) had Boise St., TCU, Marshall, or Ga. Tech. Some had Michigan St., but I had them pretty close to the top 10 myself.

I also had Ole Miss higher than anyone else did in 2014. Northern Illinois was a good pick no one else had.

The only one I ranked in preseason who didn’t make a bowl game was Michigan. Brady Hoke always surprised me by how much talent he was able to waste.

I know you think I always hate the Big Ten, but I’ve actually been too positive about some Big Ten teams over the years. I think I’ve picked Ohio St. #1 in preseason more than any other team, although it’s possible Alabama tied or passed them up recently.

I also had the wisdom not to pick North Carolina or Oregon St. in 2014. They were both worse than Michigan. Other good non-picks were Texas A&M (which resulted in some criticism here) and Washington.

In 2015, I just calculated this really quickly:

Semifinal teams
In my preseason top 4: 1
In AP preseason top 4: 1
In my preseason top 5: 2
In AP preseason top 5: 2
In my preseason top 10: 3
In AP preseason top 10: 3

*Ohio St. finished in the AP top 4 but did not make the semifinal. The AP and I both had them in the preseason top 4.

AP final top 10:
5 were in my preseason top 10
4 were in AP preseason top 10
6 were in my preseason top 11
4 were in AP preseason top 11
The AP and I both had 8 in preseason top 25 (we did not have Houston or Iowa)

Week 11 Top 25 and Commentary

In Rankings, Rankings Commentary on November 16, 2012 at 8:06 PM

I apologize for this being so late. I’ve been without home internet most of the week. Please check out my blogs about Vanderbilt‘s accomplishments in general and Georgia‘s lack thereof relative to Florida. The Georgia one was just published today, but I thought all of this should be published before games start tomorrow morning.

Top 25

rank / team / prior

1 Notre Dame 1
2 Kansas St. 4
3 Alabama 2
4 Florida 6
5 Oregon 5
6 Ohio St. 3
7 Georgia 7
8 S Carolina 10
9 LSU 14
10 Nebraska 12
11 Stanford 16
12 TX A&M 15
13 Clemson 11
14 Florida St. 13
15 Oregon St. 8
16 Texas 18
17 Oklahoma 19
18 La. Tech 20
19 Louisville 9
20 Rutgers 22
21 Toledo 17
22 Michigan —
23 SJSU 23
24 UCLA 24
25 Utah St. —

Out of rankings: (21) Northwestern, (25) Mississippi St.

Full 124 permalink

Prior rankings:
Week 1
Week 2
Week 3
Week 4
Week 5
Week 6
Week 7
Week 8
Week 9
Week 10


I’m not going to do a separate SEC entry this time since talking about the top 10 is for the most-part a discussion of SEC teams.
If only Alabama had choked with a chance to win last week.

But as it stands, the SEC’s best chance is still the Crimson Tide. I have them behind only Kansas St. and Notre Dame, although Oregon may close the gap and even surpass them in the next couple of weeks.

I have 5 SEC teams in the top 9 and 6 in the top 12 but shut out of the top 2, and as mentioned, the SEC may not even continue to have a team in the top 3.
Florida, who was #2 in the BCS before losing to Georgia, would have been a really good one-loss possibility, but due to the wonders of head-to-head (which disregards strength of schedule and benefits the team that loses to the lesser opponent), Georgia has clinched the SEC East (follow link for my thoughts on this).

It is in the realm of possibility for Georgia to end up in the top two, but that would require two teams among the BCS top 3 (Kansas St., Notre Dame, and Oregon) losing. I’m only slightly less sure that’s necessary with Alabama, but the best chances for a one-loss SEC champion would have been Florida or LSU. I would have enjoyed seeing that happen, by the way, but we’ll never know if this kind of strength at the top of the SEC would have been enough for either team.

Apart from Florida’s pending contest against Florida St., none of the four have any really high-quality out-of-conference games, but the key is the interdivisional games. Alabama played Missouri and Tennessee, Georgia played Ole Miss and Auburn, LSU played South Carolina and Florida, and Florida played Texas A&M and LSU. So Florida is 3-1 against the SEC top 6. I think that should merit being considered the best team in the SEC right now, and going into the SEC Championship game, there is a good chance my rankings will support that.

By the way, I think it’s outlandish that people are acting like Texas A&M is a better team than LSU when they lost to the Tigers at home. So much for head-to-head. Of course, if that were the only reason, that wouldn’t be good enough, but LSU also beat South Carolina. Texas A&M doesn’t have a strong interdivisional win (no, Missouri doesn’t qualify). As to the Alabama game, the Tide made the play it needed against LSU but not against A&M. I don’t see that as over-riding the A&M/LSU result just because LSU has lost more recently as a result of these events. Not to mention that LSU has been a more serious top team lately with a total of 3 losses in the last two seasons combined and 6 losses in the last three seasons combined. Also, LSU finally found a well-rounded offense against Alabama.

I read somewhere that if Oregon is in the BCS title game and the Rose Bowl opts not to pick another Pac-12 team (which seems like something they’d do, given that they picked Illinois of all teams a few years ago), they would pick Texas A&M to play Nebraska (for instance). Why would a potential Big XII title game a couple of years ago be more compelling than Florida/Nebraska or LSU/Nebraska? I don’t see it.

Anyway, speaking of the BCS top 3, I can definitely see Oregon moving into the top 2. USC, Wake Forest and a bye doesn’t stack up to Stanford, Oregon St., and a Pac-12 South champion (probably USC). Kansas St. may be stuck behind Notre Dame—both teams’ best win is over Oklahoma, Kansas St. also has a bye and a mediocre opponent (Baylor). Texas is better than USC, but not by much (especially since USC would improve somewhat by beating UCLA). USC can improve by upsetting Oregon or Notre Dame of course (and would improve at least somewhat by beating UCLA), but that would make this all a moot point. If I were Kansas St., I would be cheering for UCLA this weekend for sure. It might be counter-intuitive, but I also suspect it would be better for Oklahoma to lose Bedlam since Kansas St. beat Oklahoma St. and Notre Dame did not.

Too bad Texas isn’t playing Texas A&M this year. That could have been a contentious event for all involved in this discussion.

I can already see people asking me why Texas A&M is so low, and with the Aggies playing Sam Houston St. this weekend, I expect even more such questions next week. First of all, it doesn’t help much to play two FCS teams and SMU out of conference. Louisiana Tech is nice, but the Bulldogs have a very weak strength of schedule, so they count less as an opponent than they do in their own right. Also, the Aggies are going to play Missouri Thanksgiving weekend. This won’t help much either. Of course, my ratings don’t differentiate winning a big game last week from winning a big game (in hindsight) in Week 1. So how high would Texas A&M realistically be in that situation?

I think it’s pretty clear why Texas A&M is 6th in the SEC here if you think about it that way. And also, they followed the same script, go out to a big lead, then run into a wall. Alabama just wasn’t as competent against the A&M defense as they have been against defenses all year.

But it’s not as clear why Nebraska and Stanford are ahead. It’s having that good second or third win, I think. #2 and #3 for A&M are probably Louisiana Tech and Mississippi St. That’s not a top-10 resume. Granted no one outside of the top 9 really has one anyway, but Nebraska has good enough depth of wins: Wisconsin, Michigan St., Michigan, Northwestern, Penn St. The Big Ten has had better seasons, but Nebraska played all the desirable games against the other division, even though of course Ohio St. didn’t have a desirable result for the Huskers. The other loss was to UCLA, which of course is a much worse loss than any Texas A&M has suffered, but I think when you get toward the middle of the top 25, it’s fair to rank teams higher that have good (but not great competition) on a regular basis given that there may only be one win or no wins against the top 10 and perhaps only a couple against the top 25.

Stanford’s schedule lags behind Nebraska’s for the moment, but with the potential to steal the Pac-12 North from Oregon by beating the Ducks and Bruins, that’s going to change. Of course if they lose and Nebraska doesn’t, the Huskers will probably stay ahead though.

What can you say about the top of the ACC? Ball St., Auburn, and Furman don’t knock your socks off as non-conference opponents, but that just tells you how poor the Noles’ non-conference schedule is thus far that a team who played that slate and lost to the Seminoles is still ahead. But don’t forget that losing to Florida St. doesn’t hurt as much as losing to North Carolina St. does.
Oregon St. has fairly similar results as Stanford, but the Beavers didn’t play USC and Notre Dame. Wisconsin might make up for USC, but it doesn’t make up for both.

Texas also has two losses, but the ’Horns are being more and more weighed down by West Virginia. I mentioned Bedlam. A Cowboys win would also hurt Texas. Also, their best non-conference win is Ole Miss, which has a fair chance of losing to LSU and Mississippi St., although I wouldn’t bet the farm on the latter.

Oklahoma doesn’t even have a non-conference win as good as the one Texas has, but they should pass up the ’Horns due to playing West Virginia this week while Texas is idle. The tables might turn once again if Texas upsets Kansas St.

I don’t think Nos. 18-20 are likely to beat any of these teams (although La. Tech sure put a scare into A&M), but the important part is despite their respective records, they’re nowhere near national-championship consideration. I would much rather have the top 5 be in a satisfactory order than ensure the most deserving teams are in the top 20. But a three-loss team like Michigan being even farther away doesn’t bother me too much anyway. If Michigan shocks me and wins the Big Ten, they’ll be rewarded though.

Toledo is ahead of Michigan, but the MAC is strong enough this year to put a team with one conference loss in the top 25 when that team played a strong non-conference schedule, even though that schedule included a loss.

And then the WAC is strong enough to put a second and third team in the top 25. Utah St. is undefeated in conference with respectable losses to Wisconsin and BYU. San Jose St.’s only conference loss is to Utah St. and their non-conference loss is to Stanford.

I mentioned that UCLA has to play USC, the Bruins will also play Stanford, so two wins there would allow them to pass up several of these second-tier-conference schools. UCLA also has a shot at playing for the Pac-12 title. But (in Veruca Salt voice) why aren’t they higher now? For one thing, they could have picked a better team to lose to than Cal. Bad losses hurt, that’s done on purpose. If there are two teams in the top 5 with comparable wins, I want to knock out the one with a bad loss. Again, I’d rather make sure the teams 1-5 are the way I want them instead of teams 16-20 or 21-25. They also lost to Oregon St. and chose to play both sides of the Houston Derby. Nebraska was a really good win for them, but it’s not a cure-all.

#27 USC has the best chance of making it back into the top 25 with a win, but that would likely open up a spot for someone else. #26 Boise St. is idle and #28 Northwestern plays Illinois. #29 Texas Tech may jump both if it can get past Oklahoma St. The MAC had a big game this week between #30 Northern Illinois and Toledo, so they may essentially switch spots, although the Huskies may not get quite as high.

The next 5 out are Arizona, Washington, Cincinnati, Wisconsin, and Oklahoma St. This gives a little more context to the parts where I talk about these teams in respectful terms. I kind of left it assumed the people know LSU beat Washington and Washington beat Stanford, but I guess it’s worth mentioning more explicitly.

Georgia should not represent the SEC East

In College Football on November 16, 2012 at 2:10 PM

I was reading a blog by one of the few people who has bothered to comment on a wordpress blog of mine. He wrote a blog illogically denouncing the SEC and he has Georgia as the SEC’s best team. In reflecting upon how wrong this is, it occurred to me that Georgia should not even be considered the best team in the East, particularly if both Florida and Georgia finish with one loss apiece. I’ve now left two comments on his blog to discuss this.

Just to avoid any idea of bias against them, I actually like Georgia. After Kentucky and Vanderbilt, who I mostly favor as regular underdogs, Georgia has been my favorite team in the East. I can’t think of a recent instance where I didn’t cheer for them against Florida, Tennessee, or South Carolina. I guess I’m more neutral with Tennessee now, but I remember being very annoyed on the Bulldogs’ behalf that Tennessee won the SEC East in 2007, and I was definitely for Georgia before that. So I’m being consistent now that the team I like less deserves to play in the game and will not be able to.

As an aside, I also like newcomer Missouri—I remember cheering for them many times as underdogs against teams like Nebraska and Oklahoma—but I knew no matter how many losses there had been by A&M and Missouri, the first time one of the two got a breakthrough win, people were going to say, “Aha! The SEC defenses aren’t so good after all. Look at that Big XII team go!” And that’s exactly what happened when Texas A&M, despite falling flat offensively against LSU and Florida after going out to early leads, finally managed to win such a game against Alabama. So I’m not quite in solidarity with the new Tigers and the Aggies just yet.

There is a fan interest that has made me passionate about this issue, and that’s being a fan of LSU. As you may have noticed, Florida has been pretty good for about the last 25 years, and they’ve played LSU in all of those years. Every single one of them. Does LSU get any consideration if they tie for something as a result of this? Win or lose, the answer is no. So when I see that a team like Florida, who played LSU and Texas A&M, is being passed over by Georgia, who instead played Ole Miss and Auburn, I sympathize with Florida even though I like Georgia much better.

Georgia earned a win over Florida, don’t get me wrong, but don’t forget South Carolina beat Georgia handily. The only reason we’re not looking at a 3-way tie right now is that unlike Florida, South Carolina didn’t win its game against LSU.

Well, how did Georgia do against those top teams in the West, you may ask. 0-0! Florida was 2-0, having played both LSU and Texas A&M.

There are 6 divisional games now, two non-divisional games, but I don’t like that we treat those two non-divisional games as equal. I think they should be used as tie-breakers. That would give us that 3-way tie. Now if a team in the three-way tie had beaten both other teams, a 3-way tie is easily to resolve. Give the win to that team.

But instead, we have a circular outcome: South Carolina beat Georgia, who beat Florida, who beat South Carolina, who beat Georgia, etc.

If the SEC considered this a three-way tie, it would be resolved by looking at the BCS, where Georgia is ranked ahead, but it’s close. I would also note that Florida is ahead in the computers. The only reason Georgia is ahead is because it’s been longer since they’ve lost a game, but I think that’s an improper consideration. When head-to-head is used, no one cares when the game was (example, Georgia/Tennessee 2007).

I think one should look at who these inter-divisional these opponents were and how they fared in the West. LSU is currently 3-1, and Texas A&M is currently 5-1 against other teams in the SEC West. (I don’t think a team in the West should be looked down upon for losing to multiple teams in the tiebreaker; if LSU had lost to Florida and South Carolina, this shouldn’t be used to Georgia’s benefit). Georgia has played Ole Miss and Auburn, who have won a combined 2 conference games (both against the West, and one of which was when they played one another).

I do think you eliminate South Carolina at one point, but I don’t know if it’s merely for the loss to another SEC team or if any non-divisional losses should be considered: for instance, when Tennessee had an out-of-conference loss in 2007. Another way could be if you dropped the team with the worst inter-divisional schedule and started over. In this case, you’d drop Georgia and Florida would prevail over South Carolina by virtue of being 2-0 instead of 1-1. Or if South Carolina had beaten LSU, you could then go to head-to-head.

Although I can see how it’s the accepted norm in a two-way tie, I don’t like head-to-head anyway. Why should Georgia be rewarded tor losing to a worse team? It’s especially wrong in this situation. Florida has better wins and a better loss with the same record, but they’re shut out of the SEC title game. I haven’t even mentioned that the argument for Florida being the better team becomes even stronger if they can get past Florida St., a much better team than Georgia has played out of conference and, except for Florida, better than any team Georgia has beaten.

I also note that if you only look at divisional games, LSU, Texas A&M, and Alabama are also all tied. Certainly Alabama should prevail here, so whatever scenario one comes up with, that’s something to keep in mind, you wouldn’t want a 1-loss team falling behind two two-loss teams. But I do think that when it’s this close of a call between two teams like Georgia and Florida, especially if a third team has the same divisional record, there should be a way to look beyond head-to-head.