Tuesday, November 3, 2015

Week 10 College Football Math

A couple of things before I get started:

  • Today is election day. I know it's not a sexy election day the way 2012 was and 2016 will be, but it is election day and you should go vote. I did some math on a few local ballot measures
  • I've been working on a long post that gets into some details behind my college football model's most powerful idea. It's an idea I really like and I'm super excited to be writing about it. Look for that post later this week
  • Washington is favored over Utah, not sure what to make of that
  • Dashboards still aren't updated but will be tonight

CFP Prediction

This year's model includes an improved module for predicting the CFP committee's votes. Tonight we get the first vote. I plan on watching the committee's explanation for their rankings and continuing to tweak the model as appropriate, but here's its first predicted ranking. I'm not sure what to make of #5. Ohio State, or some of the other placements, but I'm sharing it here and opening myself up for comments/feedback/mockery.

The teams are graphed based on their CFP Predictive Score, which is a combination of power rating, accomplishments, loss penalties, and win bonuses. I've chosen not to show the y-axis because the number value itself isn't meaningful yet. For now, it's simply a way to  establish where teams should be ranked and how far apart they are.

Week 10


Week 9 Results

I'll update this post tonight

No comments:

Post a Comment