Sagarin Ratings are Twisted

[HoustonHawkeye]

I usually give some credence to the Sagarin Ratings, but am baffled by the current rankings. I do not understand how the computer ranks Iowa at 14th with a strength of schedule of 36 and ranks Michigan at 13th (they are 5-2) with a strength of schedule of 30th. I would think 2 losses would push the computer ranking of Michigan down to 20th or worse. http://www.usatoday.com/sports/ncaaf/sagarin/ . To add to the oddity, USC is one spot behind us with a 3-3 record.

I don't trust human voters bias, but I don't understand this computer ranking.

 

[thirdandpunting]

ELO rankings have been removed from the calcs. wins last year were valued more.

This ranking system is mostly based on point differential

 

[Joes Place]

I usually give some credence to the Sagarin Ratings, but am baffled by the current rankings. I do not understand how the computer ranks Iowa at 14th with a strength of schedule of 36 and ranks Michigan at 13th (they are 5-2) with a strength of schedule of 30th. I would think 2 losses would push the computer ranking of Michigan down to 20th or worse. http://www.usatoday.com/sports/ncaaf/sagarin/ . To add to the oddity, USC is one spot behind us with a 3-3 record.

I don't trust human voters bias, but I don't understand this computer ranking.

I thought Sagarin (and a few others) used an 'initial condition' like preseason polls to 'set' their programs, generally based on past performances, and at some point in the season, they eliminated that starting point bias (like halfway thru).

If that is still how they are doing it, then you are seeing 'hangover' from the preseason expectations. And there are quite a LOT of teams who were expected to be Top 25 based on last year, which are pretty average this year.

 

[swihawkeye]

I usually give some credence to the Sagarin Ratings,

I don't. They are fun to look at and compare, but they are not accurate.

 

[artradley]

I usually give some credence to the Sagarin Ratings, but am baffled by the current rankings. I do not understand how the computer ranks Iowa at 14th with a strength of schedule of 36 and ranks Michigan at 13th (they are 5-2) with a strength of schedule of 30th. I would think 2 losses would push the computer ranking of Michigan down to 20th or worse. http://www.usatoday.com/sports/ncaaf/sagarin/ . To add to the oddity, USC is one spot behind us with a 3-3 record.

I don't trust human voters bias, but I don't understand this computer ranking.

It's a little odd, but not bad, especially if you look at it closely.

First, they're less than a point apart, which is pretty darn close. I think it would be hard to argue that Iowa/Michigan isn't a toss-up, so while you might find one part of the analysis odd it's hard to argue with the final result. Bear in mind, the two teams are one bizarre punt play and one 50+ field goal away from having the same record.

Also, Iowa is ahead of Michigan in the "Recent" column, where the hangover effect from last season is erased. As already mentioned, even using a computer you have to have a starting point, which has to be last season. The further into the season we go the less that starting point matters, but at seven games it still has an effect.

 

[thirdandpunting]

I don't. They are fun to look at and compare, but they are not accurate.

Agreed, math is stupid