With all of the rumblings and prognostications of impending doom for the Trump campaign following his drubbing in Wisconsin, I feel compelled to come out of my hibernation to weigh in again on the chances of an open Republican convention.

Your election junkie has been on a self-imposed sabbatical brought on by the lack of anything to say about the primaries which, in his opinion, were effectively sewn up six weeks ago on both sides of the fence.  Given that I had not expected Trump to win Wisconsin anyway, little has changed in my mind since then. But since sites like the well respected fivethirtyeight.com have Trump trending downward, I thought it might be worth revisiting my prediction on the R side of things.

I should note that, while we got there by different paths primary-to-primary, my prediction for Trump’s delegate total is slightly higher than their original prediction of 1178. Mine is +/- 1189. Their new figures are in the 1150s, which would put Trump roughly 80 delegates short of the 1237 needed for nomination.

Indeed, if 1150-something is all Trump can muster on the first ballot, it will be a very open convention (actually a “contested” convention) and may go to several ballots. He won’t be in the 1150s though. Here’s why.

First, let’s note that fivethirtyeight, as good as they are, have under-predicted in a number of races, with winners fairly frequently outperforming their predicted margins of victory. If this pattern holds, it will be reflected in higher delegate counts being awarded in many of the remaining primaries to Trump, who is favored in most of them.

Second, Cruz’s bright start to April in Wisconsin is about to turn very dreary. The primary path now swerves into the Northeast with New York’s primary this Tuesday, where Trump will do very, very well, as he might it; and then to Connecticut, Delaware, Pennsylvania, Rhode Island and Maryland the week after. These are all strong battlegrounds for Trump and the media will be buzzing about his re-found momentum going into Indiana on May 3rd and Nebraska a week later on the 10th.

The current preferences in these two states remain a mystery, since there is no polling published for Nebraska and the last and only poll in Indiana was last December (showing a Trump lead). Indiana’s Governor Mike Pence’s endorsement, if any, will certainly play a role, but so will Trump’s momentum coming out of the Northeast.  The two biggest prizes after that are New  Jersey and California. Trump will take all 51 of New Jersey’s delegates and should do well in California–somewhere in the neighborhood of 90 to 100 of the 172 available delegates (fivethirtyeight says 88-94; I say 103).

Third, let’s talk about what we mean by an open convention. In the most technical of meanings, one might say a convention is open if no candidate walks in to Cleveland with enough pledged delegates to win the nomination. The reality though is a bit different. The convention is not truly open, I would maintain, unless someone has failed to secure the votes necessary before the first ballot to win the nomination. And this is where unpledged delegates will play a huge role in the outcome and plot line.

Trump should win Pennsylvania handily, but for his efforts he will only get 17 pledged delegates. The other 54 delegates from that state are elected directly and are officially unpledged. North Dakota has 17 unpledged, American Samoa 9, Guam 5. Altogether there will be at least 90 delegates showing up in Cleveland that are effectively up for grabs.

So realistically, Trump only needs to be somewhere around 1175 in pledged delegates to get to a first ballot victory. (One can quibble about the exact number, but it’s certainly around there.) There will be tremendous pressure on Pennsylvania’s unpledged delegation to honor their state’s primary winner. This should get him in the 1220s. At that point, it’s a mop up game, not unlike Carter in 1976, where he just had to secure South Carolina’s delegation to win the nomination. If Trump gets into the 1220s with Pennsylvania and then can’t secure enough votes out of the remaining unpledged delegates to win, then he certainly doesn’t deserve to.

Don’t get me wrong, it will be a scramble. But the chances are still well less than 50/50 that this thing goes to a truly open convention. A pledged delegate total in the 1150s of course makes it much more probable, but by no means it a dead certainty.

There is another interesting tipping point to look for that will most likely determine just how open a convention it is, should in fact it become an open definition at all under my definition. That tipping point is the number of  delegates’ votes Trump gets on the first ballot less than 1237 that, below which, signals there will is likely to be multiple ballots to determine the nominee. That is to say, the point at which the convention moves from open to contested.

I would venture this number is somewhere around 1220, maybe a little higher. By this, I mean that if Trump, despite all of his efforts can only muster, say 1223 delegates on the first ballot, in all likelihood, there would still be a tendency to push toward consolidation that would get him to 1237 on the second. If, on the other hand, he gets only to, say 1215, well, the center won’t hold at that point and all hell breaks loose. No one would win on the second ballot and it may take several to get to a nominee who may, or may not, be one of those currently standing.

Again, I’m not sure exactly where that inflection point is, but it’s there, and it will be determinative of the level of chaos that ensues and perhaps of the future of the Republican party itself.