Thursday, September 10, 2015

Week 2 College Football Math

Anyone care to guess how likely the model had WSU to beat Portland State in week 1? 98%. My model had WSU as favored to win by (on average) 32.65 points and they lost by 7, an underperformance of almost 40 points. I guess it's never too early to coug it.

So the model missed on WSU (didn't we all), but overall it performed well in week 1. The model correctly picked the winner in 76/85 games, outperforming Vegas lines, which correctly picked winners in a mere 74 games. So that's exciting!

The model also outperformed Vegas lines on my more standard measure of performance: cumulative points miss. Take the WSU game as an example. The model expected them to win by 32.65 and they lost by 7, a miss of 39.65. Vegas lines expected WSU to win by 28 and they lost by 7, a miss a 35 points. When you add up all those misses for each of the 85 week 1 games, my model missed by a total of 1,022.6 points. Which is a lot! But when you do the it's less than the Vegas total miss: 1037.5 points. Math is hard!

On to week 2, which is sneaking up on us (the first game is tonight! It features the Hilltoppers of Western Kentucky). 

Week 2 features what is expected to be one of the best games of the year in Oregon vs. Michigan State. Those teams are rated #7 and #5 by the AP, and my model has them even higher at #3 and #6 respectively. Vegas has Michigan State favored by 3.5, I have it close: Michigan State favored by 1. We likely won't get another top 10 matchup until at least week 5 when Bama plays Georgia, so enjoy!






  • A team is shaded green according to their chance to win (darker = better chance)
  • I'm experimenting with new names for the Watchability index. I really like the idea behind that number, but the name has never sat very well. I'm currently testing out Expected Game Quality, which I like for its simplicity. Don't hesitate to share comments on the new name or suggestions for other names
  • ExpectedGame Quality (formerly Watchability) is a combined measure of how good the teams are and how likely the game is to be close. Put another way: it's an estimate of how likely the game is to be a close, well-contested game
  • This post has more detail on the math behind Game Quality

No comments:

Post a Comment