This isn't as gloomy for Democrats as it sounds:
Public Release: 17-Sep-2008
INFORMS Annual Meeting
Latest Electoral College forecast shows McCain ahead by as many as 27 votes
A new approach to determining who will win the most electoral votes in the US Presidential race factors in lessons learned from the 2004 election and uses sophisticated math modeling. The research will be presented at the annual meeting of the Institute for Operations Research and the Management Sciences (INFORMS). As of Sept. 16, the margin in electoral votes could be as high as 282.8 votes for McCain against 255.2 for Obama.
http://www.eurekalert.org/...
Go to http://election08.cs.uiuc.edu/ and you'll see a range of forecasts. To begin with, the website shows 60% probability of McCain winning -- rather lower than 100%.
And results will differ if any of these happen:
More undecided voters than expected vote for Obama or for McCain. Or many more undecided candidates go for one candidate. You can adjust the model for any of these alternate assumptions.
A higher than expected number of registered voters show up to vote, and vote differently from likely voters.
Maine, Nebraska, or both split their Electoral College votes. These states allocate EC votes by Congressional District, but the model doesn't take that into account.
Oh -- and the polls whose data they're using could be wrong.
Is this election website worth following? I don't know yet. I say it's worth a look.