I'm trying to play this out . . . this whole Republican contested convention thing. How does that work if, say, no one walks in the door without having in hand an adequate number of delegate votes for president.
Let's game it out; or at least, weigh the probabilities. Call it a mental exercise to clean out a few cobwebs.
The best odds now favor businessman Donald Trump coming close to the major number of 1,237 votes, but falling short maybe 50 to 100. (He may still get there, but after Wisconsin the probabilities are running just slightly againn.) Texas Senator Ted Cruz is maybe 250 short of him; that gap could be closer or narrower, either one, by the time contests end in early June. Cruz is running much the better campaign, but Trump still has strong polling advantages in both of the magilla states of New York and California.
There will presumably be some attempt, after the statewide votes are done as of June 7, at a limited non-aggression pact between them. There will be attempts by other interests - maybe hoping for a third contender to ride in as a dark horse - to change the rules, to allow for additional possibilities for the nominee other than those two. Trump and Cruz have a joint interest in defeating that, so they may try and shut down rule changes. Both candidates would be interested in adjusting the rules in their own individual interests as well, so maybe any non-aggression pact would be reliant on allowing no rule changes at all. They could pull that off as long as, between them, they keep effective control of most of the delegates at the convention.
Okay: Suppose we then go to the first ballot with no significant rule changes. Both candidates will have been hustling hard to scrounge additional delegates in an effort to reach 1,237. (Those could come from undeclared officials who are automatic delegates, and maybe some from the few other candidates, like Marco Rubio or John Kasich, won.) It's possible, probably not likely, Trump could be close enough in his pre-convention count to pull that off. Cruz' campaign has been good at collecting stray delegates and probably could add significantly to his count, but enough to reach 1,237 before the first ballot? For now, that seems unlikely.
So we get to the first ballot and delegate votes are cast. The probability is that Trump comes in first, but still a little short of the magic 1,237, and Cruz comes in somewhere around 100 to 150 behind him, with a small scattering of other votes unwilling to line up behind either.
Then it gets interesting, because many - not all - of the delegates bound to Trump and Cruz on the first ballot are "released" on the second to do as they choose. (Most of the rest are similarly "released" after the second.) Presumably (and this may be a hotly-challenged point), the second round of balloting also will feature Trump and Cruz, and no one else. Will others be added? The existing rules don't seem to allow it at this point, allowing (as Cruz has repeatedly pointed out) a threshold of support in at least eight states, but that could be a matter of interpretation. It could be that after the second ballot, a well-organized effort might be able to come up to that threshold.
On that second ballot, who picks up? Well, more likely, it's: Who drops? Some of Trump's delegates may be there only as party officials fulfilling a role and, once unbound, they may start pushing for someone else. Who?
That would seem to suggest a serious organizing effort well in advance of the convention for someone else lying in wait, to become the recipient of those stray Trump and Cruz votes once they're available. But who will that be? House Speaker Paul Ryan seems likely to disown any efforts of that sort. But someone ought to be ready, otherwise chaos - in the form of dozens of possibly contenders, including local favorites from various states - will swiftly surface. If that happened, and it isn't brought quickly under control, you could get into a series of sudden cascading conflicts that could lead to ballots 3, 4, 5, 6 and beyond. This could go on for days.
Part of this relates to how many votes would Trump and Cruz each lose once delegates become unbound, a question no one yet can answer. (Could they persuade a few crossovers?) By the time they walk into the convention, they probably will have maxed out - at least for the moment - on the number of delegates they can easily get. After the second ballot, their numbers may be smaller but their opponents may have multiplied, absent some kind of strong organizing effort up front.
If there is a major organizing effort up front, Trump and Cruz could use that as a lever to keep their troops in line, at least for a while. And they could re-up their non-aggression pact in opposition to some new third contender, because they surely would, between them, continue to command the loyalty of more than half of the total delegates.
And that's about as far as I can take this exercise for the moment: A cycle of failed nomination votes with a built-in dynamic that keeps anyone from breaking free with more than half the total number of delegates.
Eventually, sheer exhaustion may take over, some key player or more than one will drop out, and someone will emerge. How long will it take for that to happen? The longest national party convention, the Democratic in 1924, took about two weeks and 103 ballots. That nomination, by the way, went to the lesser-known John W. Davis, after the two top contenders at the start of balloting pulled out. Davis went on to lose in a landslide to Republican Calvin Coolidge.
And that was with a mighty incentive the delegates of 2016 won't have: They didn't have air conditioning back then.
How long might the Republicans stay in Cleveland? A long time unless someone, somewhere, beats the odds.