Donald Trump likes to say that he’ll surprise political observers in Tuesday’s election. “This is going to be Brexit plus. They don’t know what’s coming,” he said Sunday.

Brexit did catch many observers off guard. But this doesn’t make a Trump victory likely.

The Good Judgment forecasting project has been asking a group of its experienced forecasters to weigh in on Trump’s chances. Here’s the neat thing: 1,740 of the forecasters made a prediction about Brexit at that time, and then made a prediction about Trump in this election.

So what do these people think about Trump? Here’s the graph:

As of Nov. 3, those who thought Brexit was at least somewhat likely — only 215 of the 1,740 — give Trump a 30 percent chance of winning. Those who thought Brexit was at least somewhat unlikely give him only a 15 percent chance.

That’s a clear difference between 15 percent and 30 percent, of course. Indeed, the 30 percent number isn’t that different from the much-maligned 538 forecast, which gives Trump about a 35 percent chance.

But 30 percent is still more unlikely than likely. Moreover, as this summary at Good Judgment describes, the forecasters who “called” Brexit correctly were less accurate across a range of forecasting questions besides Brexit, compared with those who doubted Brexit. This could mean the 30 percent figure is even less noteworthy. Maybe we can call it “Brexit minus”?