[Husker] Re: Polls

[Andrew Smith]

If you use BCS computer ratings to predict future outcomes, keep in mind 
that (1) the computer ratings were NOT chosen because of their degree of 
accuracy and (2) BCS forced the systems to be modified in ways that 
reduces their accuracy - for example, they do not consider margin of 
victory.

Billingsley is particularly bad since it rates games results based on 
the teams' ranking at the time of the game and uses the previous seasons 
final rankings as a starting point.  So beating a preseason favorite who 
eventually goes 4-7 counts for much more than a Cinderella team who goes 
9-2.

For more, see http://www.bcsknowhow.com/tag/explanation

Andy


On 10/31/2010 4:49 PM, Steve Reichenbach wrote:
> Last week, Sagarin had NU #9 (predictor #10) and Missouri was #2
> (predictor #5).  This week NU is #5 (#9 predictor) and Missouri is
> #4 (#6 predictor).
>
> Coffey has NU #6, Missouri #5.
>
> Claasen has NU #8, Missouri #5.
>
> Last week, 2 of the 6 BCS computers had Missouri #1, 3 had Missouri
> #2, 1 had Missouri #3, and 1 had Missouri #10.  NU's win should help
> with their computer ratings, even if the win was at home.
>
>
>> Wow that high, Steve what was Nebraska's rating last week ?
>>
>> ----- Original Message -----
>>
>>> Sagarin has NU #5, so I think the BCS computers will treat NU well.
> _______________________________________________
> husker site list
> husker at tssi.com
> http://romaine.tssi.com/mailman/listinfo/husker
>