I think Ezra gets this right. Obama’s primary campaign, in particular, was clearheaded, methodical and rational, focusing on delegates rather than “media cycles” and other mystical nonsense. With the Clinton campaign, the frightening thing was not merely their “voters/states that vote for us count more even if it’s a minority coalition” spin — when doomed campaigns are spinning, they have to by definition say things that aren’t true — but that they acted as if it was true.
None of this is to say, of course, that Obama’s win was inevitable. Resources and institutional advantanges matter; you can get away with hiring a Mark Penn or a Ned Colletti if your opponnent is a Bob Dole or a Brian Sabean. If Edwards had been Clinton’s major opponent, her old-school campaign/attractive candidate combination would have been enough. And Obama’s ability to get funds from online donors is a rare instance of the internet really having a major impact on a campaign. Even Billy Beane can’t win consistently with nothing to work with, and without the ability to tap enough small donors to make his campaign clearly viable the Obama’s vastly superior tactics wouldn’t have been enough. Same thing in the general — although I rarely say such things, I think McCain’s campaign really was abysmal, but under the right structural circumstances he could have won. (And conversely, under these structural circumstances he had virtually no chance; we can quibble about margins, but I don’t think there’s any serious question that Clinton/Penn would have also beaten them pretty badly.)
But, then, sabermetric analysis is always about probabilities, not certainties. Obama’s smart decisions increased his odds, and in both cases it was enough.