Interesting twist to computer polling

by psuboy on Nov 5, 2008 9:04 AM EST

Here is some interesting info...I am a pretty big number cruncher at heart and like ot look at numbers. I have been a big fan of computer polls, Jeff Sagarin in particular, for a long time. I looked at Jeff Sagarin's ratings and noticed that only he ELO rating is used in the BCS. ELO is based only on wins. The score does not matter. You could play horrible and barely win and that is the same as a dominant trouncing. Jeff also does a rating that only looks at the score. So, a blow out is much better than a close win. In his true, more accurate, rating of a team he merges the two factors together. When he does this Penn State comes out as number one.

Thanks for the fanpost! Please do not post any content from a premium site that requires a subscription. Also, if you just want to share a link, quote, or video, please consider using fanshots instead. Thanks.

Tweet Comment 7 comments | 0 recs |

Do you like this story?

Comments

Yeah

They don’t allow margin of victory in the BCS because someone might run up the score and hurt another teams feelings…..

JoePa in '09

by JGuiher on Nov 5, 2008 9:19 AM EST reply actions

not like...

usc doesnt run up the score whenever they get the chance anyhow?

Meadville, PA born and raised.

by mikeissurreal on Nov 5, 2008 10:20 AM EST up reply actions

To be fair ...

It’s not like Washington’s starters would stand a chance against USC’s third string. It’s almost impossible not to run the score up on certain teams – it just so happens that at least two of them are in the Pac-10.

by SpartanDan on Nov 5, 2008 11:02 AM EST up reply actions

Bingo

Steve Spurrier and Bobby Bowden were notorious for running up the score in the 90’s to impress the computers. They are the reason the BCS banned victory margins from counting toward rankings.

Mike
Black Shoe Diaries

by BSD on Nov 5, 2008 10:34 AM EST up reply actions

No

They don’t allow margin of victory in the BCS because
1) there’s no unbiasable measure of victory in football (unlike, for instance, baseball, where the bias is minor)
2) margin-of-victory measurements are predictive, rather than descriptive. In a margin-of-victory ranking, as the score gets close, the value of a win goes to zero.

by Bleed Blue 'n White on Nov 5, 2008 4:35 PM EST up reply actions

the reason I take umbrage with that though...

is that the media always says that team’s need “style points” for teams to shine through. what are style points, if not running up the score. they exclude it in the computers, yet make tacit intimations encouraging coaches to run up the score and increase their shine. obviously, that relates to the entirely biased, fickle and easily-persuaded human polls, but it still seems like a double standard to me.

hopes Maybin wipes his ass with some TP

by conquering lion in the 215 on Nov 6, 2008 11:56 AM EST up reply actions