Allan Lichtman, a professor of history at American University in Washington, D.C, has made a name for himself through one unusual talent: He has correctly predicted the outcome of nine elections out of the ten most recent presidential races, using a model he invented known as the 13 Keys.

The only outlier was the 2000 Bush v. Gore race, in which Lichtman successfully applied his measures to predict the race to the White House since Ronald Reagan’s 1984 re-election.

During a recent phone interview with USA Today, the professor—and apparent presidential prognosticator—was asked about his methodology for accurately predicting the winner over the past forty years.

Sign up for Your Vote: Text USA TODAY reporters and the elections team by joining our SMS service.

Q: How did your knack for predicting presidential elections develop?

Normally, historians don’t make predictions. We write about dead people, and that keeps us safe. I’d love to say I came up with my prediction system by ruining my eyes in the archives, by deep and long contemplation. But to quote the late, not-so-great Richard Nixon, “That would be wrong.”

Like so many other good ideas, I came across the keys serendipitously when I was a visiting distinguished scholar at the California Institute of Technology in southern California in 1981.

Ther,e I met the world’s leading authority on earthquake prediction, Vladimir Keilis-Borok, the head of the Institute of Pattern Recognition and Earthquake Prediction in Mosco. It was his idea to collaborate. And, of course, being brilliant and foresightful, I said, “No! Earthquakes may be a big deal in southern California, but I have to go back to Washington, D.C; no one cares about earthquakes there.”

And he said, “No, no, I already solved earthquakes.” Get this: In 1963, he was a member of the Soviet Scientific Delegation that came to Washington D.C. under JFK and negotiated the most important treaty by far in the history of the world: The Nuclear Test-Ban Treaty. That stopped us from poisoning our atmosphere, our oceans, and our soil.

More: Professor Allan Lichtman’s 2024 presidential pick

Keilis-Borok said, “In Washington, I fell in love with politics and always wanted to use the methods of earthquake prediction to predict elections.” But, he said, “I live in the Soviet Union – elections? Forget it. It’s the supreme leader or off with your head. But you, you’re an expert in American history and politics, together we can solve the problem.”

So, we became “The Odd Couple” of political research, and like any out-of-the-box analysis, we reconceptualized the problem. We reconceptualized presidential elections—remember, this is 1981, that is Carter v. Reagan, liberal v. conservative, Republican v. Democrat.

But in earthquake terms, as stability – the party holding the White House keeps the White House, and earthquake – the party is turned out. And with that in mind, we looked at every American presidential election from the horse and buggy days of politics, the election of Lincoln in 1860 to the election of Reagan in 1980, using Keilis-Borok’s mathematical method of pattern recognition.

It was that retrospective investigation that led to the 13 Keys to the White House: Simple true false questions that probe whether or not there’s going to be stability or earthquake, and our six-key decision rule, or the six keys that go against the White House party.

This article originally appeared on Staunton News Leader: Allan Lichtman shares how he developed his election prediction model