For the last few months, major mistakes by Mitt Romney have been accompanied by this warning from journalists, pundits and commentators — Romney could still win. The massive furor around Romney’s fundraiser remarks has been accompanied by patient reminders that the race is still anyone’s game.

This made sense before the conventions, when the candidates were positioning to the fall, and catering to a smaller group of voters. But now that the conventions are behind us, and most voters are paying attention to the campaigns, it’s harder to make this argument.

For starters, the post-convention period has made it clear that this election is not a toss-up. Even with his (small) bounce from the GOP convention, Romney was not able to overtake President Obama’s polling lead. In fact, Romney has never led in a polling average — he’s been behind by roughly 2 points since April, when he clinched the Republican nomination. And while it’s tempting to dismiss Obama’s convention bounce as a temporary spike, we’re at the point where it should have dissipated. If the race were going to revert to its pre-convention status quo, we would see it in the polls. As it stands, Obama’s bounce looks like a permanent bump; he’s now at 48.5% in the Real Clear Politics average and 49% in the Talking Points Memo average — within striking distance of 50.1%. Romney, on the other hand, is stuck at 45% support.

It’s possible that Romney could make up ground over the next 49 days, and win a critical number of undecideds and Obama supporters. But history isn’t on his side. As Robert Erikson and Christopher Wlezien point out in their book The Timeline of Presidential Elections, history tells us that voter preferences tend to harden in the post-convention period:

The vote margin coming out of the conventions is different than that going in. And if we measure the consequences a few weeks after the dust of the final convention settles, the result is a decisive bump in the polls for whoever had the best convention — not a fading bounce.

Yes, there are examples of candidates making rapid gains in September and October; after trailing in the weeks after the conventions, John Kerry caught up and tied George W. Bush in Gallup’s likely voter polling. If Romney could make a similar comeback, he’d have a real shot at victory. But as Nate Cohn points out, Kerry’s gains came from winning back Democratic voters who strayed from the flock. As Cohn puts it, Bush’s lead was inflated by Kerry voters “hanging out in the undecided column.”

The same isn’t true of Obama’s current lead. According to Cohn, Obama’s gains came from members of his coalition and traditionally Democratic groups — if anything, he lost support among Republicans. Barring catastrophe, Obama will hold on to his supporters. Overall, both candidates are winning the vast majority of partisans and partisan-leaners, and there’s simply not enough slack in the electorate for Romney to make rapid gains. He’ll have to win people who voted for Obama in 2008, and right now, that doesn’t seem likely.

Anything is possible, and so Romney could make a comeback. The collapse of the Eurozone or a major terrorist attack could shift the electoral balance toward Republicans. But in the absence of a dramatic event, Romney’s odds are shrinking. The debates could make a difference, but the evidence suggests otherwise — at best, debates can nudge a very small share of the public in close elections.

To put this another way, the last presidential candidate to reverse his fortunes at this stage of the game was Harry Truman, in 1948. Those are long odds, and frankly, Mitt Romney doesn’t seem like the man to beat them.

Jamelle Bouie is a staff writer at The American Prospect. You can find his blog here.