I think the polls are seriously overestimating Democrat turnout in their models, by between 2-8 points per state. The GOP base is more motivated to vote than they've been in 25 years. While I don't think the map will look like it was back then (80, 84, 88) because of demographic changes, it damn well won't look like 2008 either. Then-Candidate Obama was blessed by a convergence of a handful of factors that led to his big win last time. But there are lots of disappointed and disaffected people who voted for him who won't again, and almost no one who didn't vote for him last time who has been so convinced by his political prowess that they will now. His 2008 victory is his absolute ceiling. The tide has turned. 2010 was either a foreshadowing or the high point.
On the off-chance that even I am underestimating GOP turnout, we could wake up to a massacre of this scale:
But back down to earth (or below it, depending on your perspective)... If the polls are actually, miraculously right (they can't all be, since they vary wildly), Obama will win, and it will look like this. A smaller margin than 2008 (ironic given most polls give his party a bigger share of the electorate), but still a decisive victory:
If some of the more mid-line polls are right, but GOP turnout still improves, Romney might eek it out with a painfully small margin (for both legal challenges and political capital post-inauguration).
I hope I'm right and we wake up on Wednesday to at least a 300 electoral vote win, so President Romney can get started in correcting our nation's course without controversy. Anything less and we're likely to see Florida 2000 in at least one state, probably more. The psychological damage to our republic would be more devastating than the last four years.