Second, throw everything possible behind Cruz, and however close he can get by himself, then beg, buy or steal enough delegates to get him a majority. Not as easy to do as some commentators assume, but possible.

The advantages of choice two are that it eliminates Trump and arguably legitimates the nominee because Cruz will have won on the rules by ultimately getting the most delegates, being shrewd enough to steal some from Trump. After all, having fewer primary votes did not stop Gerald Ford from defeating Ronald Reagan in 1976. Stealing delegates is unseemly but traditional (see Eisenhower-Taft 1952). Cruz does have a base within the party, the ideological purists. Cruz has adopted every far-right position to be found, while Trump has not adopted them all. In a calmer year, a bad defeat for Cruz might have usefully disproved the right’s usual excuse for GOP defeats, that the candidate lost because he was not conservative enough to bring their supposed conservative majority to the polls. But the Trump intervention offers the right a fresh excuse for failure.

But choice two does have disadvantages. The first risk is the “riot” at the convention that Trump has cheerfully predicted/threatened (and which he might be unable to prevent if he wanted). This happens if Trump is denied for anyone. Cruz would almost certainly suffer a defeat about as bad as Trump’s, with equivalent collateral damage across the party as thousands of moderates and Trump loyalists stayed home. The depth of Cruz’ potential defeat is not yet obvious because he and his views are not as well-known as Trump’s. But they will be. Either Democrat would exploit that record. Shutting down the government is not widely considered a qualification to run it, and hard right policy on issues like abortion and same sex marriage has no relation to twenty-first century national opinion.