Let's test your "oh crap" post Brian
October 16th, 2008 at 1:05 AM ^
Probably still favored over ND, Minnesota, Purdue, Northwestern, and EMU. Don't think they would be more than TD dogs against MSU either.
October 16th, 2008 at 7:50 AM ^
Because I must confess that the statistical analysis he performed in that post both impressed and baffled me. I really don't know how he came up with the numbers he did, otherwise I would try this one myself.
October 16th, 2008 at 8:31 AM ^
I mean, it was made pretty clear in the post that Chris Stassen did the numbers and sent that to Brian in an email. If you didn't read it closely enough to pick that up I'd wager that might have something to do with being baffled by the numbers.
October 16th, 2008 at 11:19 AM ^
I certainly am not trying to critisize the post, I just wanted to see how accurate a predictor they would be. Brian did say that Vegas lines are better predictors. I didn't catch that the math was done off site and given to Brian. My bad.
October 16th, 2008 at 12:35 PM ^
we'd test more than one season, but I like your idea (for all the shit it's been given).
October 16th, 2008 at 9:22 AM ^
I'm pretty sure we'd need to know what the Jeff Sagarin ratings were after week 2 of last year, and I don't know where those can be found.
I do know that we were 7.5 point favorites over ND immediately after the Oregon game.
We were also favorites in all other games except for PSU, OSU, and Florida, but those were the lines during each game week, which isn't the same thing.
October 16th, 2008 at 10:29 AM ^
The whole post uses a pretty flawed method of predicting favorites and their point spreads. The point spreads are all way too big, as the Penn State spread shows. That makes the win percentages too low. Does anyone here genuinely think that Minnesota will be a 16-point favorite against Michigan.
October 16th, 2008 at 10:58 AM ^
Wouldn't really work. First of all, you're trying to compare the results it would give you after 6 games to the results after 2. Even if you simmed last season, you would still expect thise year's results would still be more accurate, because its a much larger sample. Also running the experiment on one team isn't enough to prove anything. What you would have to do would be to run this for hundreds of teams based on their sagarin ratings after Week 6. But, I don't know where to find archived weekly sagarins, and doubt anyone cares enough to do this.
October 16th, 2008 at 8:46 PM ^
suggest anyone that hasn't followed it before, start checking out the scores against the line-I'm sure most of you will be shocked. I'd love to know the factors involved-and how many guys work on it. I'm willing to bet(there I go again), that getting those numbers is a sick process to say the least.