Here comes the campaign's last offensive.
It is a massive outpouring of manpower and money, canvassing and calling, designed to get every last supporter to the polls on Election Day. Four years ago a similar effort increased turnout by 8.3 percent in the 17 states regarded as battlegrounds. The fact that Republican gains were greater than Democratic gains contributed to President Bush's re-election.
Predicting turnout is only slightly less foolish than predicting the Dow Jones Industrial Average, but it's likely the voting rate will be around what it was in 2004, when 60.7 percent of eligible Americans went to the polls, the highest percentage since 1968, when turnout was 61.9 percent.
How an election turns out depends in large measure on who turns out.
That said, beware the groups - there will be dozens - who claim they are responsible for the election of the next president. It's never that simple. Catholics, who voted overwhelmingly for George W. Bush against the Catholic John Kerry in 2004, did not elect Bush. Neither did evangelicals, or white men who own guns or college graduates, all of whom gave Bush majorities.
"There is no single constituency that makes a difference," says Curtis Gans, director of the Center for the Study of the American Electorate. "It is always a combination of things."
So here is a menu of turnout considerations. Put them together or mix and match them to elect the next president:
About 85 percent of the black vote customarily goes to the Democratic candidate. Bush had unusually low levels of African-American voters, winning only 8 percent in 2000 and 11 percent four years later. The first black presidential nominee likely can count on 90 percent of the black vote as a floor, not a ceiling.
A surge in black voting in certain key states - Virginia, where blacks represent almost 20 percent of the population, and even Indiana, where blacks are only 8 percent of the population - could turn the tide for Barack Obama. Other Obama targets may include North Carolina (21 percent black) and Georgia (29 percent black), where the Democratic ticket faces an uphill but perhaps not insurmountable path.
These voters turned out heavily in 2004, with an even higher turnout among educated young people. Indeed, voters aged 18 to 24 increased their participation to the highest level since 1992 - an increase bigger than any other group.
In six of the last eight elections, the Democrats have won the youth vote. Obama may not need to take it by a larger cushion than the 54-45 margin John Kerry won in 2004, as long as he does well in narrowly defined pockets in swing states.
In short, it doesn't matter whether Obama does well among students at NYU or UCLA; he will win New York and California in any case. But if Obama's get-out-the-vote efforts in Charlottesville, the Research Triangle, Boulder, Madison, and Hanover and Durham, N.H., are strong, he could be better positioned to win Virginia, North Carolina, Colorado, Wisconsin and New Hampshire.
These voters, traditionally Democratic but drawn into the GOP by Ronald Reagan's toughness on national security issues and his impatience with social liberalism, are harder to predict. But while their economic interests in 2008 may tug them toward the Democrats, they may be skittish of Obama and his air of elitism.
Although the definition of a "Reagan Democrat" is elusive, there likely are more men than women among them. One slice of these voters is whites who don't hold college degrees. McCain's lead among them has dropped by about half between September and early this month. How Reagan Democrats break may tell us a lot about how Pennsylvania, Ohio, Illinois and Missouri break next month.
Each swing state has its own peculiar demographic and geographic composition. Black voters, for example, aren't the only group whose turnout may make a big difference in Virginia. Turnout in the northern part of the state, principally the Washington suburbs, is critical, but then again so is turnout among military families near the massive naval installations on the coast and among voters in southwestern Virginia, both of which will likely come in strongly for McCain.
One especially peculiar battleground for turnout and for support is the independent vote in New Hampshire. Nationally, the movement of independent voters to Obama is strong; the swing in the Illinois Democrat's direction was 17 points in a two-week period, according to the Journal/NBC survey.
In New Hampshire, the situation is far more complex. Many independent voters sided with McCain when he won the 2000 and 2008 Republican primaries but helped contribute to the Democrats' general election victory in the Granite State in 2004. They could swing either way this year.
In its last weeks, this campaign has become a ground game.