PWR has a number of things about it that make it ugly (to me).  One is
called TUC, or teams under consideration.  A TUC is one with an RPI at
0.500 or above.  Even a tiny change (win or loss by a anyone) can
trigger someone's RPI to slip under or over the magic 0.500 mark ... and
that cascades into significant changes all over the place.  A trivial
example is if Maine had a win over a 0.500 team and that team went under
0.500, the absence of that win against a TUC might trigger a change in
one or several comparisons and result in a lower PWR.   See USCHO.com in
the PWR pages under Polls and Rankings or as a complete explanation as
you want.  If you want to do the research yourself, take a look at John
Whelan's page that allows you to see the detail of PWR with or without
any particular game:
http://slack.net/~whelan/tbrw/tbrw.cgi?2005/rankings.diy.shtml

cheers, wayne

William wrote, in part:

>>Can someone please explain the following to me:
>>Prior to tonight's Northern Michigan-Michigan Tech game, Maine was tied
>>for 16th in the PWR. Northern Michigan, I believe, was either 16th or
>>15th, as well. Now, Northern and Tech tied tonight. Michigan Tech is
>>8-19-3 on the season - including wins against Denver and Minnesota
>>(twice). So, after seeing the result, I go check the rankings again (it's
>>been a slow night) and, to my surprise, Maine DROPS to 17th in the PWR
>>while 'Cats jump to 14th. How is this possible? How does a tie against a
>>lowly team (record-wise) like Michigan Tech help you? Furthermore, why
>>would Maine drop a ranking? It doesn't make sense, but maybe I'm missing
>>something here.
>>
>>