Thursday, December 28, 2006

Sagarin watch update

Well, we said we were going to watch Sagarin’s “Predictor” model in our post "Ranking Jeff Sagarin", and we have been.

For the two games played since the post, the results have been (with Sagarin ranking, his prediction, and actual) –

62nd Central Michigan v 84th Northern Illinois – CMU by 4
Actual – CMU by17 . Missed by 13 points.

20th UCLA v 50th FSU – UCLA by 5
Actual – FSU by 17. Missed by 22 points.

With these two fairly large misses, his average miss is now 17.25 points. The closest game his Predictor has estimated was the 9 point miss in the New Mexico – San Jose State contest.

His average miss on Pac Ten teams is a rather astounding 20 points through 3 games. In other words, he has over estimated the strength of Oregon, Arizona State and UCLA by nearly 3 touchdowns per game.

Sagarin’s calls for today’s games -

49th Oklahoma State v 45th Alabama – Ok St by 4

7th Cal v 34th Texas A&M – Cal by 8

16th Rutgers v 57th Kansas State – Rutgers by 10

We will update after the games.


Henry Gomez said...

Forgetting the margin of victory for a second, how is he doing straight up?

Mergz said...

He is 5-3 straight up I believe.

However, if you had been using his predictor to bet the line, I think you would be 0-8.

Henry Gomez said...

Answered my own question. He's 5-3 straight up.

I'm 6-3 straight up in the ESPN Bowl Mania contest and I don't have a fancy computer algorithm.

P.S. You are missing the South Florida, East Carolina Game.

Henry Gomez said...

USF Predictor: 76.84
E. Carolina Precictor: 71.5
Difference: USF by 5.34
Actual score USF 24, E. Carolina 7
Actual Difference: 17

So he's 6-3 straight up. Same as me.

SMQ said...

Wouldn't the difference in the predictor and the actual final scores be overestimating the PAC Ten teams' competitiveness, rather than underestimating them? Isn't Sagarin's peceived overestimation of the PAC Ten (by ranking their slates as the ten toughest schedules in the country) the inspiration for this series?

And I will say without actually checking myself here that I've been picking the scores of games all season with no aid from computation, and I come closer than 17 points on the margins of probably 75 percent of them. I'd have to look at my 'Friday Morning Quarterback' archives to be certain about that, though.


Mergz said...


You are exactly right. I meant to say "overestimated", rather than "underestmated".

Too many bowls games watched already, with perhaps two many beers tossed in.

I will correct the post - thanks!

Henry Gomez said...

Sagarin is now 9-3 straight up.

I'm 7-5 straight up. My SEC Bias caught up with me in the Bama game and my anti-Pac 10 Bias in the Cal game. Still plenty of opportunity to catch him.

KG said...

If you assume the number he generates are his "spread" how is he doing? For example if he predicts a team to win by 5 then they have to win by at least 5 to cover (or 5 or more).

Henry Gomez said...


If you do it that way he's 8-4.

It's the 3 he lost straight up plus the Oklahoma State game in which they won but didn't cover Sagarin's 4 point spread.

I'd like to see the analysis vs. the vegas spread. I have brain lock as to how you;d do it. Let's say Sagarin says a team should win by 5 and Vegas has a 3 point spread I guess you would take the favorite. I wonder where I could get the vegas spreads for the games already played.