[Justin7]

I'm doing general research now, and was a bit surprised by some of the things i'm finding.

I'm using 10 years of games, and the average home field advantage was 6.3 points, or 39.3 yards per game. Any anyone else affirm or reject those numbers? They seem kind of high.

[Justin7]

Another random thought about curve fitting...

The default method for curve fitting is to use a delta squared type fit. This means the blow-out type games dominate the curve. Wouldn't it make more sense to use a delta square-root function? This would weight the games that were close to the spread (and where your handicapping matters) more heavily.

[tomcowley]

How are you calculating HFA? If your method is too simple, you could be getting fooled by good teams scheduling several extra home games each year against crappy teams.

[Justin7]

HFA is the sum of all points scores at home, minus all points scored on the road.

Your observation makes sense. I'll have to think about how to remove that bias.

[Rufus]

looking only at conference games should be a quick and easy way to do it

[Justin7]

Originally Posted by modelman

looking only at conference games should be a quick and easy way to do it

That's a great idea! Thanks.

[BuddyBear]

You might want to consider developing a HF advantage based on conference affiliation. Some conferences HF seems to be worth more (Big Ten, SEC, & WAC for example) while other conferences it doesn't seem to be all that big (i.e. Sun Belt, MAC, & MWC).

If you can somehow obtain average attendance for each team (I think Phil Steele has that in his magazine), you can treat attendance as both a primary variable of interest as well as a control in your analysis Justin.

Good luck...

[pico]

you should put inverse weights according to the team's historical power ranking or some kind of strength indicator.