2007 Pythagorean Wins
If you are familiar with my writings, you know that one of the cornerstones of my statistical research involves the application of the concept of Pythagorean Wins to SEC football. As we've come to realize, Pythagorean Wins is about the best indicator of team quality we have, and it comes with a long history of being a consistent predictor of future performance as well.
So let's look at things in regard to the 2007 season. After crunching the numbers, we have two overachievers with Tennessee and Mississippi State, and one underachiever, the Ole Miss Rebels. Click here for the entire calculations. Let's start discussing, shall we? We'll start with the overachievers, then look at the lone underachiever, and then wrap things up with some interesting observations from the calculations of a few teams. Here goes:
Tennessee
The 2007 Tennessee Volunteers won the SEC East, and will forever be remembered as SEC Eastern Division Champions. They should also be remembered, however, as one of the biggest overachievers in the last decade. Though the Vols went 6-2 and won the East, they had only 3.94 Pythagorean Wins. For the record, that is the fewest Pythagorean Wins ever for a champion of either the SEC West or the SEC East since the conference went to its current divisional format in 1992. In fact, Tennessee actually posted more Pythagorean Wins (4.30) in their disastrous 5-6 campaign in 2005.
It's really pretty simple. Three of Tennessee's six wins were very close -- edging 6-6 South Carolina in overtime at home, beating Kentucky in four overtimes, and narrowly slipping by 5-7 Vanderbilt after rallying late -- and the two losses were nothing short of absolutely embarrassing. Facing Alabama and Florida, two teams that combined for ten losses on the year, the Vols were annihilated by a whopping 63 points. Being quite frank about it, Tennessee played one great game all year (Georgia), and were nothing better than mediocre from there on out. All told, the Vols actually gave up more points than they scored.
Even more damning regarding Tennessee's performance is the schedule luck that they had. Thanks to the luck of the rotating conference schedule, Tennessee avoided having to face the two best teams from the SEC West in the regular season -- LSU and Auburn. Given that schedule luck allowed the Vols to miss out on West's two best teams, it is nothing short of inexcusable that they posted a mere 3.94 Pythagorean Wins.
For 2008, things don't look overly promising. A good bit of starters return, but there are a few key losses, such as Erik Ainge, Jerod Mayo, Jonathon Hefney, and offensive coordinator David Cutcliffe. It also doesn't help that that punter Britton Colquitt, possibly the best in the conference, has been suspended for almost half of the season.
In the previous ten years, only three other teams have overachieved by more than two games: 1999 Alabama, 2001 Auburn, and 2004 Tennessee. Two of those teams absolutely imploded, and only 2001 Auburn improved -- and it should be pointed out that they were laughably young all across the board in 2001, something Tennessee most certainly was not in 2007, and actually their conference record stayed the same in both years (5-3). The Vols likely won't implode in 2008, but you can probably assume safely that they will regress. Auburn returns to the schedule this season, Alabama will be even better, the same goes for Florida, and you know Georgia will be looking for revenge in Athens. Beyond that, South Carolina looks to be improved, and they will give Tennessee even more trouble. Even if the Vols can beat UCLA in the season opener, they may do no better than 8-4, and 4-4 in conference play.
Regression is almost certainly going to make a stop in Knoxville in 2008.
Mississippi State
Desplite all of the positive publicity generated from the 8-5 campaign in 2007, the Pythagorean projections are tough on the ol' Bulldogs.
Many believe that Mississippi State really turned the corner in 2007, but once you look closer it's far less impressive. They literally won every single close game they were in, and the losses were almost entirely lopsided blowouts. While they eaked out extremely close wins over Auburn, Alabama, Ole Miss, and Central Florida, they were annihilated by both LSU and West Virginia, and lost pretty big to Tennessee, South Carolina, and Arkansas. For all of the hype that they received, the Bulldogs were actually far closer to going 4-9 than 9-4. All told, Mississippi State allowed almost sixty more points than they scored, and as a whole they are simply a textbook example of an overachiever in terms of Pythagorean Wins.
And the losses for Mississippi State are major going into the 2008 season. Offensively, the team was terrible last year, and with having to replace five starters -- including the best lineman, left tackle Michael Brown, who was dismissed from the team after gun charges were brought against him earlier this Spring -- it's not going to be any better in 2008. Defensively, the majority of the unit returns intact, but they do suffer two major losses -- star defensive end Titus Brown, and defensive coordinator Ellis Johnson -- and though they won't regress dramatically, it's hard to see them improving in 2008.
No one should be expecting a complete collapse from the Bulldogs similar to the six seasons prior to the 2007 -- a decent defense and a decent running game should at least keep them somewhat competitive -- but on the other hand MSU fans should not be expecting a return to the postseason. MSU gets a huge break in scheduling with three creampuffs on the non-conference schedule, plus the rotating conference schedule spares them both Georgia and Florida, but the harsh truth remains that the team is still near the bottom of the conference in terms of both talent and depth -- only once has Croom hauled in a top 30 recruiting class, and even that was a class only ranked 27th, which was still only eighth in the conference. You can almost certainly bet on Alabama, Auburn, Tennessee, and LSU being losses, and the four game stretch that includes Kentucky, Arkansas, Ole Miss, and Georgia Tech will likely include a couple of losses as well.
Again, MSU won't implode to the levels set from 2001-2006, but it would take another fantastic coaching job by Sylvester Croom, plus a lot of luck, to get this squad north of 6-6.
Ole Miss
The only underachiever in the 2007 season was the Ole Miss Rebels, which honestly isn't too surprising because it's almost impossible to go 0-8 in conference play without underachieving to a degree. However, making no mistake about it, the Rebels were that bad. They finished dead last in the SEC, and considering they scored the fewest points in the conference while allowing the most, it was very much a well-deserved last place finish. However, they nevertheless should have won a game or two and were generally fairly competitive, sans the Arkansas annihilation. The Rebels went 0-8, but actually had a Pythagorean Projection of 1.40 wins.
The good news for Colonel Reb, though, is that 2008 should be a good bit better, and hopefully for him he can put away that tear-soaked confederate handkerchief. It's a simple formula for the Rebs: the coaching will be better, and the production from both the offense and the defense should improve as well.
The offensive line, anchored by sure-fire 2009 NFL Draft first round pick Michael Oher, will be pretty strong, and Jevon Snead will easily give the Rebs the best quarterback play they've had since Eli Manning was in Oxford. The wide receiving corps is actually pretty stout, and the ground game should be good as well. Cordera Eason, a highly-touted tailback coming out of Meridian two years ago, looks primed for SEC play, and five-star recruit Enrique Davis will play immediately as well. All told, the Rebs won't have the best offense in the conference, but it will be far better than what we saw a year ago, and there is little reason to believe it will not be the best since Eli left following the 2003 season.
Defensively, things cannot get any worse. Ole Miss finished dead last in points allowed last year, so there is really nowhere to go but up. Thankfully for the Rebels, things should be better. As I mentioned, effectively everyone returns, and the defensive line should continue to be very strong even without Ed Orgeron and the recent transfer of Chris Strong. Moreover, Houston Nutt always fielded good defensive teams at Arkansas, so things should improve.
All in all, things should be much better for Johnny Reb in 2008. They have a lot of changes, all of which have been positive for them, and considering they were easily the worst team in the league last season, the room for improvement is huge. Pythagorean projections can occasionally be wrong, but no one should expect this one to turn sour; you can safely bet the farm that the Rebels will be better in 2008.
Quick Hitters
Alabama: The Crimson Tide was no underachiever, end of story. Just like the "close" argument had no validity in 2006, it does not have any validity in 2007 either. Our Pythagorean Projection was for roughly four wins (4.52 to be precise), and that is exactly what we had... four wins. The improvement from 2006 to 2007 was very big, to be sure, and that is a testament to what Coach Saban and company did -- more on that later -- but nevertheless we did not underachieve in 2007. We were effectively projected to be a 4-4 team, and we were a 4-4 team.
Vanderbilt: The Commodores finished the 2007 season with a mere 2.57 Pythagorean Wins. But then again, this is Vanderbilt we are talking about. With that in mind, you really have to tip your hat to Bobby Johnson for the job he has done. Vanderbilt has posted over two Pythagorean Wins in each of the past three seasons, and that is the best average of any three-year stretch for the 'Dores since the conference expanded in 1992. It's not a lot of success in a raw sense -- no winning seasons or bowl games -- but when you compare what Johnson has done to the long-term historical average for Vanderbilt, it's impressive.
Florida: This is a bit of an aside from Pythagorean projections, but nevertheless here goes: The Gators scored 305 points in eight conference games in 2007, and that is the most prolific offense that the SEC has seen since the 2001 Florida Gators roamed the field under Steve Spurrier. They didn't, however, have a very good Pythagorean Projection ("just" 5.40 wins) because the defense allowed 244 points -- seventh in the conference. That's not exactly ground breaking news, but it does go to further reinforce the point that if the Gators young defense can ever figure out how to stop somebody, and Tebow can stay healthy while they do it, that team is going to be almost impossible to beat.
LSU: Finally, we have LSU. All told, the Bayou Bengals posted only 5.47, and did not overachieve or underachieve. But that's not the point here. As you can well remember from a year ago, LSU looked like gangbusters in the season opener, thumping MSU 45-0. From there, though, it was generally nothing but nailbiters, as LSU lost twice and on three more occasions had to rally from behind in the final 180 seconds to claim victory. Interestingly enough, if you factor out the MSU game and analyze them on only seven games, suddenly LSU becomes a pretty big overachiever. So, can you actually do that and still have the data predict with any degree of validity? I don't know, I haven't been able to do the entire amount of research just yet, but it's something to keep in mind.
0 recs |
8 comments
Comments
Great Stuff
OTL,
Always one of my favorite things you do. It is great to take an objective look at the teams without using the anecdotal evidence of what happened to them when Alabama played them (i.e. MSU).
The one thing that I have always been interested in is how does the correlation to pythagorean wins work out in college football. This is a statistic that has been used for year in MLB with great success. It works really well there due to the large amount of data (162 games) and relative short amount of turnover from year-to-year. How successful do you when judging college football and what are the limitations of the statistical data?
by Kenny483 on Jun 3, 2008 12:14 PM CDT reply actions 0 recs
Sample size...
Sample size is definitely an issue, of course, and the greater the sample size the more effective the metric would become. However, eight games works pretty dang well. Pythagorean Wins was first invented as a baseball statistic, but its basic tenets have been applied to just about every other sport as well (NBA, NFL, college football, college baseball, etc.), and it holds up very well. The ideal exponent used in the calculation changes based on the sport, but it’s still an effective metric almost regardless of which sport you apply it to.
Roster turnover is really one of the big things that you have to keep your eye out for, because it can occasionally make a fairly big difference. However, it’s not always a given, and even so it’s something you have to keep in mind with MLB too because you do occasionally have a lot of roster turnover there too.
Generally, though, it’s something you should keep your eye on. The biggest underachiever in the past ten years was the 2006 Arkansas Razorbacks and, with 18 starters returning for 2006, they suddenly won 10 games and the SEC West. Much along the same lines, 2001 Auburn was one of the biggest overachievers ever, but thanks to nearly everyone returning, they didn’t regress. On the other hand, returning starters doesn’t always save you. We were a big overachiever in 1999 with a ton of starters returning, but we still imploded to 3-8.
All in all, as I said in the previous post, it’s about the best indicator of team quality that we have, and it is an effective predictor of future performance as well. It’s not perfect, but it’s a damn good metric. You always want to take into consideration other developments when looking at future performance - coaching turnover, returning starters, schedule strength disparities, etc. - but you would be hard pressed to find anyone who can consistently predict future performance in a subjective manner on par with what we can do objectively with Pythagorean Wins.
by outsidethesidelines on Jun 3, 2008 3:36 PM CDT up reply actions 0 recs
By far...
... my favorite analysis every year. Comparing it to MLB, you can only be 1st place for so long leading the league in errors & strikouts, along with the on of the lowest on-base pct & highest ERA (ex: last year’s Mets, this year’s Marlins).
by DJackson on Jun 3, 2008 12:43 PM CDT reply actions 0 recs
Kinda
I agree that there are some very glaring metrics which point to regression for Tennessee. The close games being #1 The fact that Tennessee was bad statistically in SEC play is #2. However, those two metrics aren’t the only that you should look at. Over at my blog in the next few weeks, we’ll outline every conference looking at many metrics. Two where Tennessee look good are returning starters and that they were only +7 in turnovers, which is close to being lucky, but not quite.
Still, UT is probably 3rd in the East.
by The Power T on Jun 5, 2008 1:53 PM CDT reply actions 0 recs
I think...
I think we are more or less in agreement here on the Vols. It’s hard to overstate just how big of overachievers the Vols were, but as you said they do return a lot of starters. That should help some things a bit, but regression is going to take its toll, and the few losses that the Vols do suffer are quite big.
At the moment, and I’m on the low end here, I have the Vols 4th in the East. I certainly do not see them competing with either Florida or Georgia, and at the moment I have them neck-and-neck with South Carolina for fourth. Given, however, how well South Carolina played in Knoxville last year and how it took so much luck - fumble and otherwise - for the Vols to pull that one out, I expect the Gamecocks to beat Tennesee and edge out for 4th in the East. I’m thinking both are going to be in the 4-4 range, so the head-to-head tiebreaker will decide things.
by outsidethesidelines on Jun 5, 2008 3:39 PM CDT up reply actions 0 recs
Good stuff
This is the first time I’ve ran across Pythagorean wins being applied to football; interesting stuff, outsidethesidelines.
Sunday Morning QB has also done some work on Mississippi State, and he basically came to the same conclusion you did about how MSU’s record masked the real nature of the team. It also doesn’t help that he is trying to run the West Coast offense, something that doesn’t really work in college.
For all the points Florida scored, it gets my goat to think that had the Gators just broken 20 against Auburn they would have been in Atlanta instead of the obviously overachieving Vols.
by Year2 on Jun 5, 2008 2:55 PM CDT reply actions 0 recs
Thanks...
Mississippi State is just an easy analysis, honestly… you don’t need any advanced metrics to figure that one out. They literally won every close game they were in - and there were quite a few of them - and the losses were all blowouts of some variety. You simply cannot do that on a continuing basis regardless of who you are, but you certainly cannot expect to do it when you have a major deficiency in terms of both top-end talent and quality depth.
Speaking of the West Coast offense, that has been a catastrophe in Starkville. Interestingly enough, it wasn’t until Croom was stuck with a true freshman quarterback that he ditched much of the West Coast and went to a run-heavy strategy. It was only then that MSU began to experience a bit of success. If Croom would have continued trying to run the WC, or continues to do so, it would have been a disaster. They would have just killed themselves with turnovers.
As for the Gators, it’s hard to complain. Auburn has done a very good job of slowing the spread option, but still, the offense got them where they were. Had that offense not been so prolific, Florida would have honestly been a 3-5 or 4-4 team in conference play, and might would have ended up in Memphis. It was the defense that killed them all year long, I’m afraid.
by outsidethesidelines on Jun 5, 2008 3:51 PM CDT up reply actions 0 recs
No doubt about Florida. The offense was the thing keeping us afloat, but the four losses were by a combined 25 points. I mean, only the Georgia loss was by double digits, and it was down to a 35-30 Georgia lead midway through the fourth quarter despite Tebow being hurt.
It makes me real excited for next year to think that with the about worst defense since the 0-10-1 season in ‘79, Florida came that close to greatness. You figure the defense has to be better, the offense won’t miss a beat, and the schedule trades spread-buster Auburn for question mark Arkansas. Add to that a bye before Tennessee, LSU coming to Gainesville, and Georgia going to Baton Rouge a week before the Cocktail Party, and the schedule sets up very nicely for Florida.
by Year2 on Jun 7, 2008 12:00 PM CDT up reply actions 0 recs

by 




















