Save this storySave this storySave this storySave this story
The Web site Polymarket is something like a stock exchange for current events. The platform allows users to place bets on future outcomes ranging from who will win the Presidential election to how the Federal Reserve will shift interest rates to whether Taylor Swift and Travis Kelce will break up this year. (Regarding the last matter, “yes” is currently trading at seven cents, representing seven-per-cent odds—unlikely.) In early October, Polymarket had Donald Trump and Kamala Harris tied at just under fifty per cent each, with the remaining fraction of a point going to tiny extant bets on figures including Joe Biden and “Other Republican Politician.” As of this week, Trump is at sixty-four per cent. If you bet on Trump now and hold your shares through the election, then, in the case of a Trump victory, the bet will return a dollar for every sixty-four cents invested. Trump’s odds on Polymarket are significantly higher than they are in national polls and above those on other, smaller election-betting sites such as Kalshi and Predictit, which put Trump at sixty and fifty-eight per cent, respectively. Over all, Polymarket has accumulated more than two billion dollars of bets on the election, in part because, beginning around June, an account called Fredi9999 began plowing money into the Trump side. In August and September, several other accounts seemingly connected to Fredi9999 joined in, eventually amassing more than forty-three million dollars in pro-Trump and pro-Republican bets, according to Bloomberg, and driving the swell in Trump’s price.
The Trump bump on Polymarket, in turn, has become a point of discussion in the election itself. Outlets including Bloomberg and the Financial Times have habitually cited Polymarket numbers along with polling results in their analyses of which way the election is heading. On October 6th, Elon Musk posted on X that Polymarket is “more accurate than polls, as actual money is on the line.” Trump name-checked the platform during a rally speech in Michigan, on October 18th, saying, of his percentage there, “I don’t know what the hell it means, but it means we’re doing pretty well.” For those whose job it is to study the election, the question is whether the enthusiasm of a relatively small number of crypto-savvy bettors—Polymarket has roughly a hundred and fifty thousand active accounts as of October—is actually in any way indicative of reality, or whether the bets might constitute a form of manipulation, astroturfing momentum for Trump.
Shayne Coplan, the twenty-six-year-old founder of Polymarket, rarely gives interviews, but he recently spoke to me about the surge in pro-Trump bets. He told me that the activity of whales like Fredi9999 simply proves that the market is working. “There is nothing stopping people from going and buying something that they think is underpriced, but they’re not doing it right now,” he said, of the Harris side, adding, “because no one has enough conviction that it should be worth more.” Coplan is considered something of a wunderkind in the blockchain-technology world. He grew up in New York City, the child of two teachers, and became an early adopter of Ethereum as a teen-ager. He read academic papers about market theory and was particularly influenced by Robin Hanson’s 2013 article “Shall We Vote on Values, But Bet on Beliefs?,” which argues that a free market is a powerful tool for “inducing people to acquire info” and “collect that info into consensus prices that persuade wider audiences”—and, thus, that market systems should shape government decision-making and political policy as much as, say, the valuation of public companies on the stock market. Hanson coined a word, “Futarchy,” for policymaking driven by such predictive mechanisms. Polymarket—funded by venture capital from investors including Peter Thiel’s Founders Fund and Vitalik Buterin, the creator of Ethereum—is Coplan’s attempt to enact Hanson’s ideas, creating, in effect, a literal marketplace of ideas, or at least an intellectualized version of sports betting. He told me, “What we’re really trying to build is something that is a beacon of truth.”
We talked on a video call from Coplan’s glass-walled office in SoHo, where he manages a team of roughly thirty employees. He has a boyish mane of dark-blond curls and speaks with the terse impatience of someone who feels certain he has thought through every question before. Polymarket selects the markets that show up on its site, based on user suggestions, but, in keeping with crypto’s ethos of decentralization, allows them to function hands-off through blockchain tech. Each bettable subject is framed as a question; they include “Next James Bond actor?” (favoring Henry Cavill at eight per cent) and “Macron out as president of France in 2024?” (only one-per-cent odds of “yes”). Any given question “has to have an unambiguous resolution criteria,” Coplan said, which is a technical way of saying that there has to be a clear way of determining the answer. For the Presidential election, Polymarket’s criteria require that the Associated Press, Fox News, and NBC call the race for the same candidate. Once a question is live, bettors can invest in their chosen answer in a peer-to-peer market, meaning that every bet requires another user to take the opposing side. The percentage number represents, as Coplan put it, “the current market price based on all people who want to trade on either side.”
Polymarket runs on a cryptocurrency called U.S.D.C. The currency itself is not speculative in the way of bitcoin—it is a “stablecoin,” pegged to the price of the U.S. dollar—but as with other cryptocurrencies its existence on the blockchain means that transactions made in U.S.D.C. are transparent. Any and every bet on Polymarket is recorded and publicly traceable, which is why it was possible to identify Freddi9999’s outsized investments. When a question seems to have been resolved, Polymarket deploys another market-driven cryptocurrency tool called Uma, which provides a system for proposing and disputing outcomes. (According to Polymarket, despite Trump’s hearty disputes in real life, there was no trouble resolving the bet on the 2020 election.) Finally, blockchain smart contracts—essentially, self-executing, irreversible exchanges of cryptocurrency—automatically process the trades and distribute money to the winners.
Polymarket’s first tagline was “harnessing the power of free markets to demystify the real world events that matter most to you”; Coplan imagined his platform predicting when COVID vaccines would début and offering a rational alternative to the chaotic environment of social media. “There’s news and misinformation all day, and you have to figure out, does this matter? Is this material or is this just bullshit? What you see again and again is something will drop and the market won’t move, and the market’s telling you, ‘Hey, this is meaningless.’ ” Markets force participants to have skin in the game, Coplan’s logic goes; the market’s behavior, in turn, contains more wisdom than, say, pundits spouting opinions. Yet there are obvious constraints on the predictive powers of Polymarket’s self-selecting population of bettors. For one thing, using the platform has been technically illegal in the U.S. since 2022, when the Commodity Futures Trading Commission determined that the platform amounted to an unlicensed derivatives market; Polymarket coöperated with the government and settled for a $1.4-million fine as well as instituted a ban on U.S. users, though plenty of people avoid the ban by using proxy servers and other tools. (Kalshi and other election-betting platforms are facing similar existential lawsuits.) The platform also requires users to have some fluency in cryptocurrency wallets, even if there are now shortcuts available for placing bets using traditional bank accounts. As a result, Polymarket’s user base skews toward certain demographics: non-American, male, and extremely online.
One can imagine, then, that many of the users betting on the election know little about U.S. politics, and that their opinions might be shaped by a crypto-world bias toward Trump, who has courted the industry more directly than Harris, going so far as to release his own non-fungible tokens. The pro-Trump whale seemed to fit into this mold. As one participant in the election bet told me, “It was hard not to notice who he was; a lot of the whales talk to each other and know each other.” Polymarket power users saw the trades and traced them back to the same source; a user named Domer, one of the site’s most hyperactive bettors (he has money on Harris), briefly chatted with Fredi9999 on Discord and used ChatGPT analysis to hypothesize that the user was likely a Frenchman. Reuters later reported that the accounts were indeed owned by non-Americans.
I asked Coplan if all of these factors made Polymarket’s election bet less likely to be accurate. He insisted that knowledge of politics, at least, had nothing to do with it. “As price-discovery theory would go, a ton of clueless people trading would be more predictive and accurate than one expert,” he said. (In 2020, Polymarket bettors gave Trump a thirty-six-per-cent chance of winning.) The Polymarket bettor I spoke to had a different take. The booming pro-Trump market was driven only half by the wisdom of the crowd, he said, and half by “this giant whale.” Was it possible that manipulating the Polymarket odds was a way of trying to influence the election? Perhaps, as some observers theorized, the whale was attempting to provide a positive talking point for Trump and his supporters, which might then inspire more voters to join in. “No,” Coplan said when I brought this up. The bet “resolves to the outcome of the event—that doesn’t necessarily mean it has bearing on it.” His emphatic rationalism stood somewhat at odds with the fact that Trump’s followers have found it simple enough to manipulate plenty of other platforms involved in politics.
And yet if the market disagreed with one heavy-betting individual you would expect it to self-correct. So far, that’s not what has happened. Though Fredi9999’s bets have slowed or stopped (or else been reallocated to more hidden accounts), Trump’s odds on Polymarket are creeping ever higher. Coplan is unfazed. “Everyone's making a fuss, but it’s only sixty per cent,” he said. “It’s still nearly neck and neck.” ♦
Sourse: newyorker.com
