Prof. Helmut Norpoth predicted in March that Donald Trump had an 87 percent chance of defeating Hillary Clinton for the presidency.

The Stony Brook University political scientist was right – with a catch.

Norpoth’s model predicts the popular vote winner. So while Trump pulled off a historic upset over his Democratic rival to win the election via electoral votes, Clinton appears set to take the popular vote by a narrow margin.

“I feel a little bit caught in a bind,” Norpoth said Thursday.

He referred to how he used his model to correctly predict five consecutive popular vote winners starting in 1996.

“This time it’s almost tragic for me to have to admit that I didn’t get that right,” he said.

advertisement | advertise on newsday

Big picture: he did get the election right. After all the expectation and buildup beforehand that Clinton would win, “vindication is a good word for it,” Norpoth said.

His formula, The Primary Model, takes into consideration a candidate’s performance in the presidential primaries and the broader electoral cycle to forecast who will win in November. His model picked the popular vote winner when applied retroactively to elections going back to 1912, except for 1960. He has called winners ahead of time beginning in 1996.

Norpoth said his good experience with the formula gave him confidence with his prediction this year.

As he watched the returns Tuesday night, it looked like Virginia would go to Trump, though it ultimately didn’t. But merely the prospect that Virginia would not be in Clinton’s camp “was a sign of things to come,” Norpoth said.

Seeing Trump leading in Michigan “was a real shock,” he said.

“When I saw Michigan, I thought that Trump would pull it out,” Norpoth said. "Just the fact that he was ahead by that much with a good portion of the votes counted was just a sign that he would pick up states somewhere else.”

The election reaffirms how difficult it is one for one party to win three terms in a row in the White House, Norpoth said. The one exception in recent history came in 1988 with George H.W. Bush.

In 2016, the pendulum swung back to the Republicans.

advertisement | advertise on newsday

One big story of the election, Norpoth said, is “a lot of polling has led people to have a certain expectation, and then that sort of colors news coverage of what’s going on.”

Norpoth had a busy day Wednesday, including an interview with WNBC/4, and a class he taught on election campaigns.

"They clapped. They gave me a round of applause,” Norpoth said of his students. "It was good. It was good.”

On Thursday, he put Clinton “in the same boat” as Al Gore in 2000 – and said the bottom line was what he predicted in March.

“I made a prediction that the chances were 87 percent that Donald Trump would be president. That’s something that I went on record with,” Norpoth said.