Sagarin is out!

[Guest]

Part of the problem is that Sagarin's rating is part of the BCS... (the ELO-CHESS score)...

---

Hey, E-Nerd... any chance you could do a "statistical modelling for dummies" and explain how these things work??

[Economics_Nerd82]

Part of the problem is that Sagarin's rating is part of the BCS... (the ELO-CHESS score)...

---

Hey, E-Nerd... any chance you could do a "statistical modelling for dummies" and explain how these things work??

Tried to write out the "for dummies" thing. Can't really do it. But, the best way to think about it is he is taking all the data he collects (mainly, win/loss, home/away, and dummy variables for every team in Division 1) and finding that on average which team has the best probability of winning on any giving Saturday. So, here is an example. He finds (after all the math) that:

USC .790

Ohio .950

USF .820

UF .650

He would rank them by highest to lowest:

1. Ohio

2. USF

3. USC

4. UF

This ranking is for a bowl game (i.e. on a neutral field). That's why he says add 3.08 for home field. So, the rankings (probability of winning) will change after you put that into:

1. USF

2. Ohio

3. USC

4. UF

In essence, he is dissecting what goes into a win. Winning is the most important thing... WIN WIN WIN... either by 1 point or 100 points it doesn't matter. Either with 100 yards of penalties or 5 yards. Either sloppy or beautiful a win is a win and that is all that matters.