HRS College Football Computer Rankings: Prediction Results

Week 3: 7/20
Week 4: 7/18
Total: 14/38

Improvement at least? I would expect these to get better each week so hopefully this continues.

HRS Rankings Explanation

A few notes about how the ranking system works:

-A team's ranking is a sum of ranking points accumulated over its games. The elements that factor into a team's points for an individual game include: margin of victory (capped at +/-35), location (home, neutral or away), opponent winning percentage, and quality of opponent. This last element allows the entire system to iterate through and adjust instantly. For instance, a MAC conference game will have an effect on teams that played a MAC team in non-conference, teams that played teams that played a MAC team in non-conference, etc.

-Additionally, I have placed set values on games with non-D1 teams. I believe teams should be rewarded for taking on Division 1 opponents in their non-conference schedule, and the points awarded reflect that. Teams that lose to non-D1 opponents suffer dramatically.

-Because of the iterative nature of the system, the rankings will only improve as we get further into the season and have more data to input. As such, some of these rankings should be taken with a grain of salt early in the season. This is the reason that the BCS does not put out results at the beginning of the season. Unlike the BCS, I am not concerned with credibility (because I have none).

-You will notice that there is no dependence on outside rankings in my formula. I don't believe that Alabama should be #1 solely because they started the year that way in the AP Poll and they haven't lost yet.

That's pretty much it, check back regularly for updates and feel free to post any comments on the rankings or ideas for further analysis!

HRS College Football Computer Rankings

Monday, September 27, 2010

Prediction Results

No comments:

Post a Comment

BCS Computer Rankings

Followers

Blog Archive

HRS Rankings Explanation