The realistic-worst-case scenario for the Romney campaign that I laid out on February 8th began:

He wins Arizona, which is winner-take-all, while Michigan is a split decision, with Romney winning his natural demographic and Santorum winning his natural demographic.

Depending on how the delegate split winds up looking in Michigan, and what the final margin is, that might be a reasonable description of what just happened.

The rest of that scenario involved Romney losing most the next twenty contests over the month of March: Washington, Alaska, Georgia, North Dakota, Oklahoma, Tennessee, Kansas, Alabama, Mississippi, Missouri and Louisiana, and possibly Ohio.

Could that still happen? Could Romney really go on that big a losing streak, winning only Wyoming, Idaho, Massachusetts, Vermont, Virginia (where Paul is his only opposition), Hawaii, Puerto Rico and (his one big win) Illinois?

Sure. It would require five things to happen:

  • There is no “rally around the obvious nominee” effect – GOP primary voters remain unconvinced that they have to give Romney the prize, and decide the subsequent contests based on their preference for who should win, not their impressions of who will win.
  • Angry conservatives out to bloody Romney’s nose again dominate the Washington caucus, and give Santorum a nice compensating victory.
  • The demographic factors that drove the vote in the February contests also drive the vote in the other March contests (which would mean Romney loses the Southern and border states by huge margins, loses the high plains states solidly, and I think loses Ohio narrowly).
  • Santorum raises the money to compete in those contests where he needs to spend to win (Ohio most prominently).
  • Gingrich doesn’t prove a spoiler, enabling Romney to win contests that would otherwise be out of reach by splitting the anti-Romney vote.

The first of these is the one that is the most uncertain, it seems to me. The media narrative tomorrow is going to be very positive for Romney. A three point win is close, but it’s clear, and unless the Romney campaign is totally incompetent (which, to be fair, they may well be) they should be able to spin this powerfully as Romney’s “Comeback kid” moment. That could move polls significantly – certainly in Ohio, possibly even in the border and high plains states, very likely in Washington. And movement in the polls will be self-reinforcing.

That rally effect usually happens eventually, because most partisan voters really want to vote for their party’s candidate, not have to choose the candidate. But it’s had a couple of chances to happen already and it hasn’t yet. It never happened in the 2008 Democratic race. Maybe it won’t, this time, ever?

If it doesn’t, the rest of the list looks pretty plausible. Even after winning Michigan, Romney could still wind up running a brutal gauntlet in March, and looking a lot more like a loser than a winner at the end.

But even in that realistic-worst-case, he’ll still be well ahead in delegates. And he’ll still be the only one with a plausible path to the nomination. And after California, New Jersey and Utah vote, he’ll have it. Assuming he still even wants it by then.