Opinion | Jon Kimball, David Rees: PredictIt is gambling, not political science

(Washington Post illustration; images by iStockphoto/Getty Images)
(Washington Post illustration; images by iStockphoto/Getty Images)


Jon Kimball and David Rees host the Election Profit Makers podcast.

As two more or less normal people who don’t know much about the government regulation of finance, we never expected to be emotionally invested in the repeal of a no-action letter by the Commodity Futures Trading Commission (CFTC). However, when the CFTC announced this month that they were withdrawing their 2014 agreement allowing the operation of PredictIt.org, the online political prediction market, we were shaken. We owe a lot to PredictIt: It rekindled our friendship, allowed us to forge a community of fellow enthusiasts and made us thousands of dollars.

Nevertheless, we are glad it is dying, and we thank the CFTC for saving us from certain doom.

Like many other high-minded online experiments, PredictIt’s descent into madness seems inevitable in retrospect. A project of Aristotle, a data company, and the Victoria University of Wellington, it launched in 2014. It was pitched as a not-for-profit scientific study about the wisdom of crowds, which enabled it to skirt online gambling laws. The theory was that people’s political predictions would be more accurate if they were financially invested in them. In practice, this means you could buy into markets such as “Which party will win the N.C. Senate race” at anywhere from 1 to 99 cents, depending on how it is currently favored by your fellow gamblers — ahem, investors — and then ride the market, if you want, all the way to its conclusion, when you get a payout or lose your stake.

At a time of unprecedented capital in politics, you can see a sort of populist appeal to being able to intervene in elections another way, for as little as a few cents. PredictIt was half investment vehicle, half wishing well, combining the social cachet of playing the market with the exhilaration of acting on one’s political prejudices. As such, it proved irresistible to know-it-alls, haters and (mostly) males, which we are.

We’ve been using PredictIt since the summer of 2016, when Jon started betting against die-hard Trump supporters who were convinced Hillary Clinton would go to jail. Jon won big shorting Donald Trump in the Iowa Caucus and basically felt like he had discovered a money-printing machine. David, his friend since the seventh grade, suggested we start a PredictIt-themed podcast, because he knew it was the only way he could get Jon to answer his calls. We believed that discussing politics via our PredictIt portfolios would prove our integrity: Unlike most pundits, we were literally putting our money where our mouths were.

PredictIt had markets in everything from which candidate would win a given primary to which words would be uttered during the presidential debates. At its most frenzied and debased, there were actually markets in how many times Trump would tweet in a given week. (You could argue this was where PredictIt started to go off the rails as scientific research, since that was one step up from guessing the number of jelly beans in a jar.)

The markets that nearly broke us, however, were about politicians’ approval ratings. PredictIt used polling averages from the RealClearPolitics site for these markets. Jon would spend hours trying to figure out which polls would be included in the RCP average. He would set his alarm for early morning hours when new polls were released, sometimes maxing out his investment at the PredictIt-mandated limit of $850, buying as high as 90 cents to squeeze out an 8 percent return (after fees), only to have everything come crashing down when an unscheduled poll was released.

This speaks to one of PredictIt’s shortcomings as a study of political wisdom: lots of profitable trades were simply knee-jerk reactions to breaking news. One sign of PredictIt addiction is feeling your political event horizon shrink down to the few seconds before you hit refresh on Twitter for the 10 thousandth time since breakfast. At the height of his mania, Jon was spending 16 hours a day reading Twitter while trading on PredictIt, powered only by gas-station energy drinks and avarice.

PredictIt rewarded not just avarice, but also schadenfreude. Because you’re betting against other traders, every time we made money it came at the expense of those idiots on the other side of the political aisle, poor slobs who lacked our decency, our humanity and our genius. We loved taking their money.

Turns out they loved taking our money, too. On election night 2016, when Jon’s all-or-nothing investment in a Clinton victory evaporated in a matter of hours, MAGA PredictIt traders delighted in our trauma — financial and otherwise. As we wept into our microphones, they laughed all the way to the bank. We recorded what we thought was the last episode of our podcast the morning after Trump’s upset victory. We were bitter. We were broken. We were broke. We swore we would never log on to PredictIt again.

But then the 2020 election cycle geared up. The old itch came back. We turned to the better angels of our nature, and heard only a demonic chorus screaming “Revenge!” Sure enough, Jon won enough money betting on Joe Biden that he was able to lease a new car. David calls it the PredictItMobile, a literal investment vehicle.

Even so, we have to admit that the CFTC is right to call PredictIt’s bluff. We’d probably feel sadder about PredictIt’s demise if we still took it seriously as an academic endeavor, or if we were some of the “whales” who made hundreds of thousands of dollars trading on it. As it is, we’re grateful for the hundreds of hours of podcast conversation it inspired, as well as what it taught us about ourselves: Namely, that we should never, ever try actual gambling.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top
%d bloggers like this: