Donald Trump is now, after South Carolina, clearly leading the Republican primaries. Donald Trump has also been unable to get much beyond 35% of the vote in any primary or caucus. He has the largest share of support, but that share is only about a third of the Republican vote, and sometimes less. None of the other candidates can beat Trump, but that doesn't mean he's going to win.

There are four scenarios left for the Republican nomination, two likely and two not. The unlikely scenarios are the ones getting the most attention, because they resemble the normal election cycle. What most people would expect, because it's what usually happens, is that the front-runner (Trump) would expand and consolidate his support to 50%+ (or at least 45%+), or else that the rest of the field will consolidate around a single opponent who will go on to beat him (most people suggest Marquito Rubio). Those events would represent a reversion to the usual pattern, but I suspect they're not going to happen. They should be happening already, and they haven't been.

Trump is winning, but his support is not expanding. And he's not modulating the key; getting into a public beef with the Pope is not how you expand beyond your base, and the fight with the Pope was only one strange-but-typical day in that campaign, which traffics in that level of vitriol and weirdness all the time. And even with Jeb (;) Bush gone, the field isn't consolidating around Marco Rubio much. If Trump can't get to 40%, Marquito hasn't even gotten to 25% so far. Chris Christie dropping out didn't move Rubio past his previous ceiling. He's not suddenly going to get 40% of the vote because Bush is gone. And some other people, including Cruz, aren't going anywhere for a while.

The two most likely scenarios are a minority victory for Trump, where he gets more than 50% of the delegates and the nomination with only a smallish plurality of the actual votes, and a brokered convention in Cleveland, where Trump falls short of 50% of the delegates, while Rubio, Cruz, et al each net some smaller number, and the convention becomes about, ah, the art of the deal.

How could Trump win with only one-third of the actual votes? A lot comes down to delegate math. Each state doles out delegates by its own rules: some split them up proportionately between a number of vote-getters, while others give most or all of their delegates to whoever's first past the post. As we get further into spring, more and more states will be winner-takes-all or winner-takes-most. In a typical primary year, this works out pretty much right: the early, proportional states divide the delegates between a number of candidates, and later the winner-take-all states help decide between the top two. But this year, the Republican field won't be narrowed to two leaders by the time the winner-take-alls begin on March 15; the field may not be down to two by the convention. So those states, some of them big states, may end up giving all of their delegates to a candidates who only gets 30-something-percent of the vote. If the field stays broken, Trump may never need to get to forty percent. Remember, he just got ALL of South Carolina's delegates with less than one-third of the votes.

The other scenario also involves a broken field, with three different candidates, and maybe a fourth, getting significant shares of the delegates, so that no Republican has the required 1237 delegates before the convention. For this to work, though, it's not enough for Cruz and Marquito to keep showing "strong seconds" and "strong thirds." They have to win states. And in the later proportional states (because even late in the game a few are proportional), they have to put up strong numbers, at least a third of the total vote in that state, or else to run very strongly in particular sections of the state (because lots of state allocate by Congressional district). If that happens, there is going to be some back-room dealing in Cleveland that would make Marcus Hanna proud.

There are two things to keep in mind as things go forward. First, there is the twenty-percent rule. A candidate has to get at least twenty percent of the vote in a state (or sometimes in a Congressional district) to get ANY delegates. A beats-expectations 17% may get pundits chattering about your momentum, but it nets you exactly bupkis. I would point out that both Rubio and Cruz only polled in the low twenties in South Carolina, nearly tied between 22% and 23%, and (because of South Carolina's delegate rules) neither of them got any delegates at all. And Rubio hasn't gotten out of the twenties in any race so far. Rubio and Cruz are just above the cut-off point where they get nothing. They have to make stronger showings even to play spoiler.

This rule means that a candidate like Ben Carson may not get even one more delegate, not even if he stays in the campaign until June. If he picks up a dozen more, he will have done fairly well. But that doesn't mean that Carson doesn't have an effect on the race. A vote for Carson is a vote that the front-runner in that particular state doesn't need to get to win. It reduces the percentage of votes needed to carry the state. And it's a vote none of the challengers get; it may help keep the other candidates from hitting the number they need to get a delegate. A vote for Carson isn't a wasted vote; starting next week, it's effectively a vote for Trump.

The second thing to remember is the effect of favorite sons on state-wide races. Candidates carrying their home states can pick up big stashes of delegates. John Kasich is clearly banking on winning Ohio's winner-take-all primary on March 15. Marquito Rubio is planning to win Florida's winner-take-all primary the same day. Cruz hopes to rack up a large number of the delegates at play in Texas. If Rubio, Cruz, and maybe Kasich put together decent overall showings (with Kasich as essentially a regional candidate in the Northeast and industrial Midwest) added to some of this home cooking, we're moving closer to a scenario where Trump misses clinching the nomination and the others go to Cleveland with delegates to trade. On the other hand, if these guys lose their home states to Trump, we're not going to be in suspense very long.

The brokered-convention probably requires Little Marco to win a few states by the end of March, and for Cruz to roll up good numbers of delegates in the proportional states (because Cruz's best states turn out to be proportional and not winner-take-all). If Kasich wins Ohio and another state or two (Michigan, say, or Connecticut), then the game gets harder for any one candidate to win before the convention. If, on the other hand, you see Trump roll up everything on Super Tuesday, or if you see Trump starting to win states with a 45% share, then you can start forgetting who Kasich and Rubio are early.