| FTBL Sagarin's Ratings

[HiTide]

Take a looksee.

http://www.usatoday.com/sports/sagarin/fbt08.htm?loc=interstitialskip

 

[musso]

can someone summarize how this poll works?

 

[Argo]

I'm not even sure Sagarin can.

 

[ghice]

I think he calculates the poll similar to how a Rube Goldberg machine pours a drink.

[youtube]http://www.youtube.com/watch?v=mvCqxbUncmw[/youtube]

 

[TerryP]

can someone summarize how this poll works?

Who a team has played and won/lost to combined with the opponents played and their W/L schedule.

I find it useful in the last part of the season as an index when it comes to looking at spreads, etc., during bowl season.

Other than that, it's not something I pay a great deal of attention to until we hit November and then it's the later part of that month considering how many teams play conference games during the last part of the season.

If you want example, look where Tx. Tech is ranked and look where TCU is currently ranked. TCU is up there because of their loss to OU.

 

[Guest]

Its a power rating system that considers strength of schedule as well as margin of victory.

Here's a neat article discussing Sagarin and his rankings.

http://rolandallen.com/000854.html

 

[alagator]

can someone summarize how this poll works?

It is statistics. Statistics are from Satan himself. Satan is bad. Ergo, this stuff is pure evil.

I did okay until I saw the word 'bayesian' in his explanation. I then had to go sit in the corner with my baseball bat out of fear and trepidation.

All these guys who claim their statistical models are objective and unbiased and just report results based on the actual results are full of bunk. Every statistical model has to include some subjective data for the starting point. That beginning bias somewhat is removed as the year progresses, but not totally and completely. And they are often self fulfilling - for instance USC gets points for playing an Oregon State team that gets credit for beating a highly rated USC team. These things are full of circular logic - just like human pollsters. But at least everyone knows, understands, and accepts the human factor.

 

[Superchief]

95% of statistics can be made up 50% of the time.

 

[HiTide]

Regardless of the methodology, I was thinking of how cool it is to have Bama on top.

 

[RussB]

Regardless of the methodology, I was thinking of how cool it is to have Bama on top.

Me too, till I saw that he had USC's strength of schedule at #7!

 

[Guest]

Regardless of the methodology, I was thinking of how cool it is to have Bama on top.

Me too, till I saw that he had USC's strength of schedule at #7!

That is just strength of schedule so far. Considering the remaining cream puffs they have, they probably won't be in the top 50 in SOS.

 

[RussB]

oh ok