We seem to be left with 3 salient possibilities:
- Case 1: Trump gets to 1,237 first before the beginning of the convention.
- Case 2: Cruz gets to 1,237 first before the beginning of the convention.
- Case 3: Nobody gets to 1,237 before the beginning of the convention.
The probability of Case 1 is high[a] but not certain. Here’s Harry Enten:
My own delegate estimate has Trump falling short of the 1,237 delegates he needs because he has done poorly in the west so far, and many of those states haven’t voted yet. It’s also possible that Kasich plays better than we might think among moderate voters in the remaining states to vote in New England and Mid-Atlantic.
Moreover, there are plenty of signs that Trump would have lost a majority of states that voted on Tuesday had Rubio not been in the race. I’m talking about Missouri and North Carolina, where Ted Cruz beat Trump in a one-on-one race in the exit polls.Trump may be rising, though he is still not getting close to a majority of the vote in most states. If the anti-Trump voters can find a better way to coordinate behind one candidate, they probably can beat Trump in a lot of upcoming contests.[b]
Trump has certainly fallen behind the pace that he needs to be on in order to get 1,237 delegates before the Republican convention. Sure, he’s only behind by 3%, but he was up by 8% a week ago[c], and that’s a significant slowdown. But I’m coming around to the view[d] that a lot will be riding on California and its 172 delegates[e]. Trump currently is polling in the low 30s there, with Cruz about 10 points behind[f][g], though it’s worth noting that FiveThirtyEight doesn’t think there’s enough good data available even to offer a forecast about who will win the golden state[h]. While I’m not as skeptical about Case 1 as Enten is, I do think that it has no more than a 2-out-of-3 likelihood of occurring.
The probability of Case 2 is very low. IIUC, Cruz has 423 delegates and would need to get more than 77% of the remaining delegates in order to arrive at 1,237 before the Republican convention[i]. I’ll assign Case 2 a probability of 1% just to make the nagging voices in my head go away. (What’s that you say, nagging voices? 9/11 was an inside job? No, sorry: that’s not even getting a 1% probability.)
As for Case 3, it might be hard to calculate its probability alone, but since Cases 1, 2, and 3 are, for all intents and purposes, mutually exhaustive, we can just assign it the remaining probability space. Hence:
- 67% – Case 1: Trump gets to 1,237 first before the beginning of the convention.
- 1% – Case 2: Cruz gets to 1,237 first before the beginning of the convention.
- 32% – Case 3: Nobody gets to 1,237 before the beginning of the convention.
So who gets the nomination in a contested convention? Let’s not pretend anyone really knows[j]. I sure as hell don’t. But my best guess at the possible outcomes is as follows:
- 40% – Case 3a: Trump wins.
- 15% – Case 3b: Cruz wins.
- 10% – Case 3c: Kasich wins.
- 35% – Case 3d: Someone else (a “white knight”) wins.
Let me explain: Initially, I thought Case 3a was pretty unlikely because Republican party elites wouldn’t let it happen. Well, these elites haven’t really shown us much reason to think they can get what they want this year. And there might well less opposition to Trump than I perviously thought[k]. If Trump has anything going for him, it’s the ability to negotiate a deal (and, perhaps, to bribe the hell out of the delegates). Trump doesn’t quite have an even-money chance here, but he’s close. Cruz would have the next strongest claim at a contested convention in virtue of his delegate totals. But he’s so loathed by many within the party[l] that I doubt he’ll receive much support. And, whatever the polls say at the moment[m], I don’t think he’s perceived as a good general election candidate. Kasich is likely to be a better matchup against HRC – at least that’s what he says[n]! He’s probably right about that too. However, I’m not sure how clearly this will be seen on the convention floor in Cleveland. Group polarization bias[p] is a major factor in situation like this one. So I think Cruz and Kasich each have a little by more than a 1-in-10 shot at a contested convention. That leaves a little less than a 50/50 chance for a white knight to pick up the nomination. Paul Ryan won’t remove himself from contention[r]. Neither will Mitt Romney[s]. However, I really can’t see Jeb Bush or Marco Rubio coming back into the fray as the both left under humiliating circumstances[t][u].
Putting all of this together we get a 0.79 (=0.67 + (0.32*0.40)) chance of winning, with Cruz at 0.06 (=0.01 + (0.32*0.15)) chance and None-of-the-above (i.e., Kasich or a white knight) at 0.14 (=0.32*(0.10+0.35) chance. I’m up to changing my mind about the specific probabilities that I’ve assigned here, but the general approach looks right to me.