On Friday, Marc Ambinder raised the specter of a "constitutional crisis" that would result from a situation in which Barack Obama were to win the popular vote but lose the Electoral College. Ambinder says unnamed strategists from both campaigns are "chewing over" what would happen in this scenario.
As I've written before, this is exceptionally unlikely to happen. But don't take my word for it--check out Nate's stats over at FiveThirtyEight.com, which is quickly becoming the authority in terms of Electoral College projections, for what they're worth. According to the compilation of polls on the site, the national popular vote is currently as tight as can be--splitting exactly 50/50. And yet, after running 10,000 computer simulations (see "Scenario Analysis" on the right side of the home page), Nate finds there's less than a three percent chance of Obama winning the popular vote and losing the Electoral College, despite the closeness of the race. If it ends up not being decided by a razor-thin margin, that likelihood would quickly approach zero.
In fact, Nate finds that according to current state-by-state polling, there's a slightly greater chance that McCain would win the popular vote and lose the Electoral College--about four and a half percent. (McCain, no less than Obama, currently seems to be at risk of "wasting" votes in states he isn't going to win.) There's precedent for this error: In 2000, all the speculation before the election was that it was Bush, not Gore, who was susceptible to winning the popular vote but losing the Electoral College. I say this not to defend the Electoral College, whose continued existence is hard to justify on substantive grounds in any case, but just to remind readers that for all the speculation you're going to hear over the next five months about an Electoral College–popular vote split, it's very improbable, and if it does occur, it's just as likely to work to Obama's benefit as to McCain's.