After the first weekend of college bowls, here’s how my model stacks up against the actual results. 

As reported by Fox News and the London Telegraph, Simon Blackburn of London’s Royal Holloway College has developed a mathematical equation describing the geometry of parallel parking. Dr. Blackburn’s page (found here) has a link to his original paper.

I have developed the following college football bowl predictions based on my model of college football rankings. A few things to note: the most lopsided matchup seems to be Nevada vs SMU. The best matchups seem to be Oklahoma vs Stanford and South Carolina vs Connecticut.

Below I gave top 30 rankings based on a system I developed. Those rankings included such factors as home field advantage and point spread — factors the BCS forbids computer models from considering.

Below is another top 30 list. The rankings are based strictly on the Elo ranking system. For an excellent description of this system, see this article in Wikipedia. I take a different approach than straightforward Elo ranking, though. I first rank all games between FBS and FCS opponents. I then set each conference’s initial score to their respective division score. I then rank all out of conference games. I set each team’s initial score to their respective conference score. For the division and conference ranking I use K = 24. Next I rank each game, using K = 96. The teams are ordered by descending score (ie. highest score is ranked #1).

Here are the results through games played by December 5.

As usual the BCS bowl selections caused stirs throughout college football fandom. Many people hate the fact that computers running mathematical models have a major roll in determining those bowl bids. Well, get over it. The computers are imperfect, but they are unbiased. You cannot say the same thing about the human voters in the polls.

Not to be outdone by other geeks, I have developed my own math model and computerized it. Unlike the BCS models, I account for the strength of your conference schedule, the point spread in the game, home field advantage, and ties in regulation time. Boo hoo it all you want, but my model had Alabama over Florida before the game and predicted Stanford as a major factor in determining the PAC 10 champion.

My rankings:

Disclosure: I am a fan of the following (in priority order). 1 - BYU  2 - The Mountain West Conference  3 - Oregon