Thursday, May 22, 2008

Is there Something in the Water in Bristol?

Mark Schlabach has ranked the ten BCS Champions and magically puts the four SEC national champions in four of the bottom five spots. If Schlabach wasn't a former writer for the AJC, where he covered the UGA beat quite competently for several years, I'd be fulminating about his being biased against SEC teams. Instead, I simply think that he underrates the importance of strength of schedule and overrates rolling up big numbers in one or two games.

Take his treatment of 2004 USC and 2005 Texas, for example. The 2004 USC team was excellent, but they did have a number of close calls against pretty average opponents. The luster of that team comes from the beatdown they put on Oklahoma in the national title game, but in retrospect, is beating Oklahoma in a BCS game really that much of an accomplishment? In contrast, 2005 Texas was a more dominant team from start to finish. They blew out every opponent, save for an excellent Ohio State team (on the road, no less). They then beat much the same USC team (OK, USC's defense wasn't the same, but their offense was intact and even better than 2004) in Southern California. For my money, 2005 Texas is the only team in the past ten years that would give 2001 Miami a run for its money.

To test Schlabach's conclusions, I decided to take a look at Sagarin's rankings for the same time period and here's what the computer says:

1. 2001 Miami - 108.7, no losses, SOS of 27
2. 2005 Texas - 106.0, no losses, SOS of 13
3. 1999 Florida State - 102.1, no losses, SOS of 11
4. 2004 USC - 101.2, no losses, SOS of 7
5. 2000 Oklahoma - 99.35, no losses, SOS of 29
6. 1998 Tennessee - 98.5, no losses, SOS of 24
7. 2003 LSU - 96.3, one loss, SOS of 28
8. 2007 LSU - 92.4, two losses, SOS of 11
9. 2006 Florida - 91.9, one loss, SOS of 8
10. 2002 Ohio State - 87.8, no losses, SOS of 30

A few notes on the rankings: I used the Sagarin predictor, but Sagarin's site only explicitly lists out the predictor starting in 2001. I assume that the 1998-2000 rankings are the predictor because Sagarin only created his modified ranking at the behest of the BCS when he was told that he needed a ranking that did not take margin of victory into account. Also, the 2000 rankings were not updated after the bowl games, so Oklahoma is a little undervalued in the rankings.

And now, a few observations:

1. Schlabach isn't out on a limb when he puts the SEC national champions in the lower half. I thought that introducing an objective ranking set would reward SEC teams by putting a greater value on strength of schedule, but that was not the case. For instance, my initial thought was that Florida 2006 is analogous to USC 2004: two teams that played their best games in the national title game, covering for a number of close calls during the season. What I had not taken into account was how good 2004 USC's schedule was, as well as the fact that 2006 Florida had more close calls because of their weak offense. (2004 Auburn's SOS, in case you were wondering, was 60. That's why there wasn't more of an outcry at the Tigers not winning a share of the title.)

1a. Then again, the SEC has four national champs in the decade and no other conference has more than two, so screw everyone else.

2. The Big Ten is not good. It has only one national champion in the past decade and that champion is the lowest rated team on the list by a not-insignificant margin. The 2002 Bucks would be a 21-point underdog to 2001 Miami on a neutral field. Then again, the 2002 Bucks beat much of the 2001 Canes on a neutral field, so what do I know? Also, Sagarin ranks the '98 Buckeyes marginally ahead of Tennessee as the best team in the country, so Ohio State does have a saving grace.

3. If you're wondering where the best Georgia teams of the past decade would come in on the list, the 2002, 2003, and 2007 Dawgs all had Sagarin Predictor ratings of 89.

4. For comparison's sake, here are the ten teams as ranked by the power ratings on James Howell's site:

2004 USC - .963, SOS of .817
2001 Miami - .955, SOS of .688
2005 Texas - .951, SOS of .794
2000 Oklahoma - .934, SOS of .702
1998 Tennessee - .927, SOS of .716
2006 Florida - .924, SOS of .825
2002 Ohio State - .923, SOS of .730
1999 Florida State - .918, SOS of .724
2003 LSU - .896, SOS of .636
2007 LSU - .852, SOS of .775

For what it's worth, 1995 Nebraska scores a .941, which would put them 4th on this list. 1992 Alabama scores a .944, which would put them ahead of all of the SEC teams on this list. 1991 Washington, which is the one team I've seen since I started watching college football that would rank with the '95 Huskers, scores a .946. The last team to have a higher ranking than '04 USC according to Howell's rankings is '72 USC with a .964. '71 Nebraska scores a .985. '58 LSU scores a .976. I stopped looking once I got to World War II because I don't want to be an appeaser.

1 comment:

Mr.galle said...

Hi, I like this blog is so cool!
I already link to this page.
If you don't mind, Please link to my blog this URL
http://sportmultiple3.blogspot.com/