Nate Silver’s final (?) projection is Obama +2.7, 50.9-48.2. That would be a little bit better than Bush did in 2004, and probably produce a decisive Electoral-College result that even the Ohio Secretary of State and the Florida Governor between them couldn’t reverse.
But I’ve always been the greedy sort. Silver has not adjusted his figures for any sort of systematic polling bias (in the technical, rather than the pejorative, sense of that term. So if in fact the specialty Latino polling showing a 50% Obama margin in that group is more accurate than the 30% margin that shows up in most national polls, and if Latinos are estimated to be 7% of the voters, that would add 1.4% to the margin, making for a convincing popular-vote margin of 4.1%, 51.6-47.5. If in addition the pollsters’ likely voter screens fail to fully account for increased Latino enthusiasm for voting in the wake of the Arizona law and Latinos accounted for 8% of the vote rather than 7%, that would add another half-point to the margin. Add another half-point for the Obama ground game, and we’re at a 5.1% win. That still wouldn’t be as impressive as the 7.2% margin of 2008, but then Sarah Palin isn’t running this year.
More to the point, the final tally – assuming Silver is right that 0.9% goes to third parties – would give Barack Obama 52.1% of the popular vote, and Mitt Romney … 47%.
If that happens I might have to re-evaluate my position on the existence of an omnipotent and benevolent Ruler of the Universe.