BCS Observations

The first BCS standings came out this week. The top 15 are as follows:

1. Ohio State
2. South Florida
3. Boston College
4. LSU
5. Oklahoma
6. South Carolina
7. Kentucky
8. Arizona State
9. West Virginia
10. Oregon
11. Virginia Tech
12. California
13. Kansas
14. Southern Cal
15. Florida

Two components of the BCS, the Coaches Poll and the Harris Poll, are public and have been widely published for weeks. The computer rankings, on the other hand, are not nearly as public. If one looks at the way the computers have the rankings, the top 15 is very different.

The Computers have the rankings as follows (omitting the highest and lowest ranking for each team of course):

1. South Florida 1.000
2. LSU .9300
3. South Carolina .8600
4. Kentucky .8500
t5. Ohio State .8300
t5. Arizona State .8300
t7. Kansas .6600
t7. Virginia Tech .6600
9. West Virginia .6400
10. Oklahoma .5500
11. Auburn .5300
12. Oregon .5000
t13. California .4800
t13. Virginia .4800
15. Missouri .4300

Where's Southern Cal? 22. at .0900

South Florida is #1 in at least 7 of the 8 computer polls. Pretty good for a team that was thought to be an also-ran in the Big East at the beginning of the year.

The SEC has spots 2-4 in the computers. Perhaps their reputation as the best conference is not simply media hype.

The Conferences break down this way SEC: 4; Pac-10: 3; the rest 2 each.

If the two human polls are taken as representative of reputation, preseason predictions and hype and the computer polls are taken as indicative of numbers, then who of the BCS top 15 is benefiting from reputation, preseason prediction, and hype and who is being hurt the most?

Benefiting:

1. Southern Cal +.5736
2. Oklahoma +.31845
3. Florida +.1878
4. California +.17995

Suffering:

1. Kentucky -.25005
2. Arizona State -.2203
3. Kansas -.21445
4. South Carolina -.1752

Thus, Southern Cal and Oklahoma are way overrated in the polls. Kentucky, Arizona State, and Kansas are way underrated.

Why are these teams overrated or underrated? Another way to look at it might be why are some teams getting so much credit in the polls and not from the computers? It could be bias; it could be reputation; it could be hype; it could be ignorance. I think it is something much simpler: the preseason poll. Take for instance the Coach's Preseason Poll. The overrated teams were ranked 1, 8, 3, 12, and 10. The underrated teams were ranked NR (1 vote), NR (1 vote), NR (0 votes), NR (90) votes.

It takes a long time to either climb the polls or to fall out of them. Too often (all the time) a team's rank in the poll is based significantly on where they were ranked last week, which was in turn based on the week before, so on and so forth until you get to preseason rankings. The preseason rankings are not based on football, but are based on the predictions of people who year in year out are proven wrong. Maybe we ought to consider no polls for at least a few weeks so that our rankings of football teams can be based, at least in part, on football.

(Numbers derived this way: In the BCS formula, each team is awarded a percentage score for their ranking in the polls. The percentage is determined by dividing the points (or votes) a team received by the total points possible. The number used by the BCS to represent the computer formulas is done in a similar manner, but each team's highest and lowest ranking is omitted. I simply subtracted the computer average from the poll average.)

Delenda est Clemson