Historian's election prediction system is (almost) always correct. Here's how it works.

For those of you who know who Allan Lichtman is, you know his presidential predictions are something to take seriously. For those who don't know, Lichtman 77, has correctly predicted 9 out of the past 10 presidential elections dating back to 1984.

And this election he has predicted Kamala Harris will win the White House, according to a video first reported by The New York Times.

If a 90% track record isn't a good enough reason, perhaps Lichtman's credentials might also sway you . He:

• Has a Ph.D. specializing in modern American history and quantitative methods from Harvard.
• Earned the title distinguished professor from the American University in Washington, D.C
• Has written 13 books and hundreds of academic articles.
• Has lectured in the U.S. and around the world.
• Has served as an expert witness in civil and voting rights cases.

What are Allan Lichtman's 13 keys to pick the winner, and why did he choose Kamala Harris over Donald Trump?

Lichtman developed the metrics for his predictions with the help of seismologist and geophysicist Vladimir Keilis-Borok in 1981. The system uses 13 historical factors or “keys” to determine presidential races − four of those factors are based on politics, seven on performance, and two on the candidate’s personality. Essentially the metric is a true/false questionnaire from the incumbent party's perspective. If the incumbent party's answer to the question is true, a point is awarded. If the answer is false, the point goes to the challenger.

If six or more of the keys are against the party in the White House, it's predicted to lose. Otherwise, Lichtman forecasts the party in power will win again.

The video describing the reasons behind Lichtman's decisions can be found here .

How accurate has Allan Lichtman's prediction for the presidential election been?

Lichtman has correctly predicted the outcome of almost every election over the past half-century, except for the race in 2000, in which Republican George W. Bush defeated Democrat Al Gore.

Lichtman claims he correctly predicted 2000 election

Lichtman has maintained he was correct in 2000 and Al Gore won the election over George W. Bush.

According to a video posted to YouTube last month , Lichtman says his 2001 study found 180,000 ballots in Florida were rejected as invalid. Of the ballots rejected, Lichtman claims, there was racial disparity. He found that African Americans' ballots had a rejection rate of 1 for every 9 to 10 ballots and that 95% of African Americans had voted for Gore. Ballots of white voters who leaned toward Bush were rejected at a rate of 1 out of 50, he says. Lichtman claims that if even rejection rates were even, Gore would have won.

In the days after the election, a slew of recounts took place , eventually revealing overcounts − people who voted for more than one candidate − and no-counts among voters.Ballots design was at the root of the problem: Butterfly ballots caused voters who intended to vote for Gore to accidentally cast ballots for independent Pat Buchanan. “Hanging chads,” in which a punched piece of paper hung on at one corner, or a dimpled chad, in which a ballot was not punched but only indented, contributed to the debate.

Eventually the U.S. Supreme Court ruled that recounting some ballots but not others violated the Constitution's equal-protection clause, and Bush famously won the state by 537 votes − and the election.

Contributing: Marina Pitofsky, Sudiksha Kochi, Rachel Barber, Elizabeth Beyer and Joshua Bote

This article originally appeared on USA TODAY: Historian's election prediction system is (almost) always correct. Here's how it works.