Chris Stirewalt has broken prospective VP candidates into three categories:

If Romney is leading, he will play it super safe and choose someone well-vetted and non-controversial who doesn’t make news, just as George W. Bush did in 2000.

If Romney is still in a dead heat, he will focus more on the electoral map and reinforcing his own positives. That’s what Bill Clinton did in 1992.

If Romney is running behind by the end of this summer, he will do what McCain did and take a chance on a less conventional pick in an effort to shake up the race – the Hail Sarah pass.

This seems backwards. If Romney is leading in the weeks before he announces his running mate, he might be able to take more of a political risk by naming a “less conventional” VP nominee, and if he’s trailing he won’t be able to afford to follow McCain’s example by making an “exciting” pick. Bush didn’t choose Cheney because he had a comfortable lead over Gore. He chose him because he had an enormous deficit of national security experience that Cheney was supposed to offset. The lesson the Romney campaign seems to have learned from observing the 2008 Palin experience is that this is something that a nominee should never do. Regardless of the polling, Romney has an incentive to choose a running mate who reinforces the qualities he wants to emphasize.

The lists Stirewalt makes don’t seem to fit the scenarios he has in mind. For instance, Romney wouldn’t entertain the idea of choosing Mike Huckabee or Condi Rice unless he believed he was in a position to risk alienating some people from his own coalition. Those choices definitely would alienate Republicans of different stripes. Politically, Huckabee or Rice would be disastrous picks to varying degrees. Economic conservatives would panic in reaction to a Huckabee selection just as they did when he was a presidential candidate (and his old ethics , and conservatives and Republicans from all sides would revolt against a Rice selection. She has the unique distinction of being disliked by pro-life conservatives, non-interventionists, most realists, and most neoconservatives all at the same time, and she is so closely identified in voters’ minds with George W. Bush that it would become virtually impossible to refute the charge that the next Republican administration would be Bush’s third term.

Virtually none of the prospective candidates that Stirewalt gives for the “dead heat” scenario matches Clinton-Gore example. He names Rubio, McDonnell, Toomey, Ayotte, and Thune, only one of whom has any executive experience, and all but one were elected to their current positions in 2009 and 2010. Relatively untested political novice with no executive experience is not what Romney wants to emphasize. (How does Thune keep ending up on these lists?) Portman and Daniels are much better-suited to play Gore to Romney’s Clinton, but they are grouped with Pawlenty among the “safe” selections. Pawlenty aside, the “safe” picks are probably the best choices Romney has available, and Portman or Daniels are so much more qualified for the position than any of the others, and Daniels evidently doesn’t want to face the media scrutiny that comes with a national campaign. The next few months of speculation about the VP selection seem as if they’re going to be a lot like the nominating contest: the most likely result is obvious from the beginning, but to keep ourselves entertained we’re going to pretend that the outcome is seriously in doubt.