As read the article, and compare to results locally in IL, I don't see that they were as insulated from reality as the posts in this thread suggest. Miscalculation 2 is probably the key to the failure of the experts:
This directly caused the Dem turnout to be a larger fraction than 2008 when few pundits on either side expected a return to those historic levels. One problem in the polls could be that they weren't asking respondents how they identified themselves in 2008. If they had we might have seen more pre-election stories about the conversion of I to D. With that info I believe that political advisers would have treated their own polls differently as well.
