Thursday, October 16, 2008

Wanna Bet?

Lately I have been working on various ways to predict scores in football for my fantasy football league. But since I am trying to predict something that doesn't really happen (i.e. fantasy football games) I can always claim that the models that I predict are right. So I decided to give a whirl at predicting some NFL football games versus the spread. So lets see how this goes.

Game M1 Conf. M2 Conf.
SD@Buf Buf .36 Buf .73
NO@Car NO .25 NO .66
Min@Chi Min .28 Min .52
Pit@Cin Cin .78 Pit .09
Ten@KC KC .20 KC .75
Bal@Mia Mia .12 Mia .95
SF@NYG SF .44 SF 1.00
Dal@StL Dal 1.00 Dal .30
Det@Hou Det .31 Det .32
Ind@GB Ind .45 GB .45
NYJ@Oak Oak .47 NYJ .39
Cle@Was Was .74 Cle .09
Sea@TB TB .89 Sea .70
Den@NE NE .79 NE .20

M1 and M2 are based on the same model, one just includes blocking on team defense. Conf. is the scaled confidence each pick has for the method.

I wonder if I can pick better than chance doing this.

Update: If I was going to bet $100 multiplied by my confidence, I would have ended up this week with the totals of: -$322 from the M1 system, and -$199 from the M2 system. No so good. I'll give it another try this week.

No comments: