[log in to unmask] wrote:
>
> Why do we have this problem?  Mainly because no consideration is given
> (unlike say basketball) to how a team plays down the stretch.  If
> that came into play, almost everyone, including the most biases
> Gopher fans would have to admit that Minnesota is playing like
> garbage and in no way deserves a #1 seed.  The last 16 used to be a
> criteria, but it was taken out because it hurt teams in deeper
> conferences when they knocked each other around down the stretch and
> in the conference tournaments.  It seems that somewhere there needs
> to be a happy medium.

The problem more generally was that there was absolutely no accounting for
schedule strength, which could be highly variable even in the stronger
conferences (an example would be BC's end-of-season run in 1997-98).  That,
and the 16- or 20-game cut-off was completely arbitrary and would give full
credit for the 16th (or 20th) game and none for the 17th (or 21st), although
the current TUC criterion has similar problems.

So if you're looking for a happy medium, you'd presumably like to work in
some kind of strength of schedule -- how do you go about doing that while
maintaining consistency with the current PWR framework?  The most "correct"
way of adding SoS to the original L16 would be to do a mini-KRACH rating for
the object team using the full-season ratings of their opponents, I think...
although it doesn't address the arbitrary nature of the 16-game cut-off.
But that's not consistent with the current framework, because it introduces
KRACH to the calculation.  If you stick with RPI, how do you calculate it?
Use the full-season wpct and owpct for opponents, or only a later-season
window?

I think that the committee probably looked at all of the problems inherent
in evaluating the end-of-season trend and decided that the easiest way to
fix it was to remove it entirely.

Then, too, addressing the other point (which I've snipped) about Minnesota
locking up an unreachable seeding position with their early-season play, I'm
not sure how fair it is to downgrade that.  I don't know that the committee
wants early season games to mean less than late-season games, and I'm not
sure, historically, how much it matters how well a team is playing at the
end of the season.