>Dave Hendrickson writes: Before anybody played any games, there were
>polls based on expectations. That is true, but games have been played
>and of course early season polls mean very little I think it is foolish
>to even have Harvard rated in any poll until they at least play a game.
>I always thought polls were based on the overall performance and record
>of a team not upcoming expectations.
>
>Brad Cuyler.  NMU
 
This only leads to second guessing. Are one team's games worth the same as
some other team's games, especially when they are a _different number_ of
games, played against _differing opponents_. The only way to remove
expectation is to have something like power ratings to equalize
schedules. But then you are talking about a different sort of poll.