Menu
Final Sagarin rankings have ND at 7–behind Michigan at 6
(no message)
Link: http://sagarin.com/sports/cfsend.htm
Replies to: Final Sagarin rankings have ND at 7–behind Michigan at 6
Thread Level: 2
Another broken model. That's absurd...
(no message)
Thread Level: 3
Surprised he doesn't still have Bama ahead of Clemson...wonder how betting agencies who already paid
out for Bama winning feel right now.
Thread Level: 4
Oh, yeah . . . I kinda forgot about that early payout thing.
Pretty funny. I guess they got the attention they were seeking, so they got what they wanted anyway.
Thread Level: 5
Exactly. It was a publicity calculation and they only paid out the pre-season bettors.
(no message)
Thread Level: 3
8-5 Mississippi State checks in at #8.
Iowa rounds out the top 10. Sagarin needs to rethink things.
Thread Level: 4
...been saying that for a long time. His ""ranking"" is complete garbage unless its used for what
its intended, which is a guide for betting against sports lines.
Thread Level: 5
Is it not "garbage" as a betting data base?
Seems to be garbage either way.
Thread Level: 6
Actually not a badmodel
In this instance the early ND "close wins" count against them and hold them back.
If you look at the right hand column, which provides greater weight to games late in the season ND is 4 and Michigan is 22.
To me, seems pretty fair.
Go Irish
Thread Level: 7
That makes it seem even dumber.
You can’t program common sense.
Thread Level: 8
It is just the math... guess we disagree
You could look at the "body of work" and essentially rate teams based on the entire year performance with some math/model
OR
You could give greater weight to the most recent games... again with some math/model
OR
use some other data set to assess performance... or just look at the games… no math/ "eye test"
each tries to assess the relative performance of teams and I am okay that different views will give different results.
and if I look at the Recent rankings, I think ND at 4 seems pretty fair.
Thread Level: 9
If it has UM ahead of ND at the end, it's retarded.
Sorry, but it is. Math is great. But like I said, you can't program common sense. There is absolutely no way to reasonably find that UM is ahead of ND in a ranking, especially if it is as a predictor. And that is how that model shakes out at the end. It's laughable.
Thread Level: 10
^^^This^^^
(no message)
Thread Level: 7
Re: Actually not a bad model
(no message)