In the course of an article arguing that a large vote for Obama is not a vote for his policies (and, equally curiously, that the total and utter failure of conservative policies is not in and of itself a reason to try something else), Gerard Baker, who is to Thomas Friedman as Ricky Valance was to Richie Valens, says:
What, in these circumstances, would a scientific model predict as the winning margin for the Democratic presidential candidate: 10, 15, 20 percentage points? In fact, as of yesterday, Mr Obama seemed to have a solid but by no means overwhelming advantage of between 5 and 6 percentage points.
In fact, the Ray Fair model, with default values, predicts four points.
It is actually quite easy to look these things up you know.