A huge caveat is in order here—those delegate estimates are based on a guess of how delegates will be awarded according to election results. That's not how many are selected, however. In Iowa, for example, there is a three-round process to select the final delegate slate. That's why the NY Times' delegate list is different, significantly so, than CNN's. The fact that no one can get an accurate read as to the state of the GOP race is yet another sign of how f'd up the GOP is. Add to all that chaos the fact that up to a quarter of those delegates are unbound, and can vote for anyone they want in the end, and things get that much more confusing.
But let's assume that the NY Times estimate is roughly accurate. That means that Romney has to win 729 of the 1,541 delegates left, including about 100 remaining super delegates. That would be fairly easy in a winner-take-all primary system, such as we saw in 2008 on the Republican side. Winner-take-all means that if win a state, you get all of its delegates. However, most states this year are apportioning their delegates proportionately.
Of the states that have yet to vote, only a handful are winner-take-all contests—Maryland, Washington DC, Wisconsin, Delaware,
California, New Jersey and Utah. Together, they add up to 400 228 delegates, but well over half of them—California (169) and New Jersey (50) doesn't vote until June 5. And Utah's 40 winner-take-all delegates won't be awarded to Romney until June 26.
The rest of the states will allocate their delegates proportionately, meaning that Romney has to reach his final goal in small increments, each one bleeding precious time and money. And for every Illinois, in which Romney should do well, there's a Texas, where he won't.
Much of the punditry talks about how impossible it is for the not-Romney's to get to 1,144. And they're pretty much right—Santorum needs 968 of the remaining 1,541 delegates. Gingrich needs 1,039 of them. Paul needs 1,097. Sure, it's mathematically possible for them to get to 1,144, but it just ain't gonna happen.
But what they can do is prevent Romney from getting the delegates he needs, which would then give us that mythical creature called the "brokered convention." But that's just fantasyland. It won't happen.
None of these candidates will give up their claim to the nomination to a white knight (and you know he'd be a "he" and "white," no matter how much Sarah Palin would desperately hope otherwise). If it happened, the surprise GOP nominee would emerge from a bitter convention with no campaign organization, no money, and facing a hundreds-million defining ad-blitz from the Obama campaign and its allies. Furthermore, as Republicans found out in 2008 with Sarah Palin, you kind of want your candidates properly vetted before thrusting them on the national stage.
If Romney can't close the deal at the California, New Jersey and Utah contests in June, he'll make a deal with one of the other candidates in order to cross the magic line, and it would happen before the RNC convention. If Ron Paul has enough delegates to push Romney over the line, he could trade them for VP nominee Rand Paul. Santorum could offer his delegates up in exchange for being the veep nominee himself. Gingrich could ... nevermind. He'll hoard his pittance of delegates and declare himself emperor in his own mind.
Twice Romney could've put an end to this clown show—in Michigan and in Ohio. By failing to win convincingly despite spending tens of millions against underfunded joke opposition, he ensured that this race will continue into spring, and possibly (hopefully!) beyond.
Update: Turns out California is proportionate, not winner-take-all, which makes Romney's task a little more difficult.