[Husker] Interesting statistical analysis of Wisky

[Paul Dalen]

I don't know about his methodology....

Seems pretty off-the-cuff to me.  Perhaps valid, but it's hard to tell.

The accepted practice to approach something like this is to normalize the
data so you're comparing apples to apples.  One way to do that is to
compare how one team's offense performed against its opponents defenses in
terms of percent.  I think NU has put up numbers in the neighborhood of
120% of what our conf opponents are allowing.  Not sure about Wisky.

I'm also of the opinion that using arbitrary standards such as FBS opps or
blowouts to exclude portions of the data is a dangerous thing to do.  And I
believe that most of my peers would agree.

On Thu, Nov 29, 2012 at 9:58 AM, Steve Reichenbach <reich at inetnebr.com>wrote:

> WRT Wisconsin having better field position, it may be because their
> special teams play is better.  This is an area where NU hasn't
> excelled.
>
> A point-spread of 1.5 and a probability of winning at 53.4% doesn't seem
> like the basis for confidence.
>
> Most of the computer programs are even closer than the 3pt betting line.
>
> Nebraska can be much better than they have played in many games, esp.
> wrt turnovers.  So, like many championship games, the outcome of this
> game may hinge on turnovers and special teams.  If NU plays cleaner
> than it has and makes few mistakes, it could be a comfortable win.
> If NU gives the ball away as it has been doing, it may give away its
> best shot at a conference title since 1999.
>
> > There, as most have heard, three kinds of lies; lies, damned lies and
> > statistics. Bit ibalways find them interesting and this article talks
> > about a couple of sites for stats comparisons and gives reasons why
> > Nebraska may be pretty confident heading into the B1G championship.
> >
> > http://hailvarsity.com/2012/11/trimming-the-fat/
> >
> > Not saying I buy all of it but interesting none the less.
>
> _______________________________________________
> husker site list
> husker at tssi.com
> http://romaine.tssi.com/mailman/listinfo/husker
>