In 2020, he predicted Joe Biden’s arrival at the White House. A year later, he predicted the name of the winner in the Senate. And his latest prediction has as its axis the presidential elections in the United States, between Kamalla Harris and Donald Trump. In doing so, he not only got the winners right, but also the difference in votes they would win. This is not Paul the octopus or any other fortune-telling animal. Nor is it a prophecy. According to Thomas Miller, director of the Data Science program at Northwestern University, there is a clear winner.
For Miller, Kamala Harris will win by a large margin in NovemberUsing his same complex numerical analysis models, Miller is now predicting not only that Harris will win against rival Donald Trump, but that she will do so with 449 of the 573 electoral college votes, according to his simulation website Virtual Tout.
The Northwestern data scientist’s approach is not based on polling likely voters, as traditional pollsters do, but on electoral betting markets. For Miller, Polls are a “snapshot of the recent past” and use sample sizes that are too small (usually between 500 and 1,500 people) to accurately reflect the electorate.
The betting markets are, According to the scientist, the “best at predicting the wisdom of the crowd”Rather than being based on political ideals, these prediction markets are based on cash, and as people place bets in the hope of making a profit, they are guessing who they expect to win rather than who they want to vote for.
In both the 2020 presidential race and this year’s election, Miller used data from Predictit, the country’s largest election betting siteBy breaking down betting counts into jurisdictions and tracking hourly changes, this expert feeds the information into his custom model that outputs daily odds.
As he explains on his website, the custom model correctly predicted the 2020 presidential race has been suggesting since September 10 (the day of the Harris-Trump debate) that the vice president would usher in a “Democratic landslide” in November.
“The electoral prospects of both parties changed dramatically on September 10,” Miller notes, Following two critical events: the debate and pop megastar Taylor Swift’s subsequent endorsement of Harris. Before that, it was virtually a tie.”
While only time will tell if Miller will be proven right again, he does appear to be correct in pointing out one of the key themes in this year’s race: that it went from a dramatic landslide in Trump’s direction to A drastic landslide in favor of Harris on Biden’s exit from the race: “We’ve never seen a reversal of fortune on this scale, but then again, we’ve never seen anything like this happen in recent political history.”