Author: Kyle Chayka
Translator: BitpushNews Yanan
Imagine a website where you can bet on future events, just like trading stocks - from who will win the presidency to the Federal Reserve's interest rate moves, even the relationship status of Taylor Swift and her boyfriend (currently, the odds of them breaking up are 7 cents, meaning the market thinks there's only a 7% chance of that happening). This is Polymarket, a platform for prediction markets that is both entertaining and challenging.
Back in early October, when everyone was debating whether Donald Trump or Kamala Harris would win, Polymarket's odds were close to 50%, with the remaining slim chances going to other candidates including Joe Biden and "other Republican politicians." But as of this week, the situation has changed, with the market seeing Trump's chances of winning at 64% - meaning that if you bet on Trump now and hold until the election is decided, you could turn every 64 cents invested into a $1 payout if he wins.
Interestingly, this Polymarket prediction deviates significantly from national opinion polls, and is also higher than the assessments of similar platforms like Kalshi and Predictit, which put Trump's odds at 60% and 58% respectively. As of now, Polymarket's total bets have surpassed $2 billion, and behind this impressive figure is an undeniable driving force. Starting in June, an account named Fredi9999 began making large bets in support of Trump. Then in August and September, several accounts seemingly associated with Fredi9999 joined the fray, collectively wagering over $43 million not only to back Trump but also to boost the Republican party's chances - a series of moves that directly pushed up Trump's odds of winning.
In turn, the dynamic predictions about Trump on Polymarket have become part of the broader election discussion. Authoritative media outlets like Bloomberg and the Financial Times have been citing Polymarket's data alongside traditional poll results when analyzing election trends. On October 6th, Elon Musk stated on X platform that Polymarket is "more trustworthy than polls, since real money is at stake here." And on October 18th, Trump himself mentioned the platform at a rally in Michigan, saying: "I'm not sure of the exact meaning, but it shows we're doing very well."
For those focused on election analysis, a key question arises: Do the bets of these relatively niche but crypto-savvy investors - Polymarket had around 150,000 active accounts as of October - genuinely reflect the real-world situation? Or could these betting behaviors be a behind-the-scenes force bolstering Trump's momentum? In a rare appearance, 26-year-old Shayne Coplan, who founded Polymarket, recently had an in-depth discussion with me, focusing on the remarkable surge in bets on Trump's support. He explicitly pointed out that the active participation of "whale" players like Fredi9999 is evidence of the market forces at work. "While the market allows people to freely buy what they believe to be undervalued assets, they have not taken such actions so far," he explained when the topic turned to Harris, adding, "because no one is confident that her value is severely underestimated and worth taking a position on."
Coplan, a prodigy in the blockchain technology world, grew up in a teacher's family in New York City and became an early Ethereum adopter as a teenager. He has extensively read the academic classics of market theory, with Robin Hanson's 2013 article "Shall We Vote on Values, But Bet on Beliefs?" having a profound impact on him. Hanson argued that free markets are a powerful driver for "motivating people to acquire information" and "aggregating that information into a consensus price that can persuade the public" - and that the market system should influence government decisions and policies just as it does the valuations of publicly traded companies. Hanson also coined the term "Futarchy" to describe a policy-making model driven by prediction mechanisms.
With the venture capital support of Peter Thiel's Founders Fund and investors like Ethereum co-founder Vitalik Buterin, Coplan is trying to translate Hanson's ideas into reality. His creation, Polymarket, is essentially a market for trading ideas, or more precisely, an AI-powered upgrade of sports betting. "Our true goal," he told me, "is to build a lighthouse that illuminates the truth."
In the glass-walled office in SoHo, I had a video call with Coplan, who leads a team of about 30 elite individuals. With his deep golden curls, he spoke with a sense of urgency, as if each question had long been pondered in his mind, waiting to be expressed.
Polymarket's market choices are based on users' enthusiastic suggestions, but to uphold the core decentralized principle of cryptocurrencies, the platform uses blockchain technology to achieve a "non-interference" approach to these markets. Each bet-able topic is carefully designed as a question. For example, "Who will be the next actor to play James Bond?" (Currently, Henry Cavill has an 8% chance of winning) or "Will Macron step down as French president in 2024?" (Only 1% believe "yes"). Coplan emphasizes that each question "must have a clear resolution criterion" - in other words, a definitive way to determine the answer.
Using the presidential election as an example, Polymarket has set strict standards: the race is only settled when the Associated Press, Fox News, and NBC all declare the same candidate the winner. Once a question is opened, bettors can wager on the answer they believe in within this peer-to-peer market. This means that each bet requires another user to take the opposing side, forming a battle of wits. As Coplan puts it, those percentage figures represent "the collective judgment of all those who want to trade on the current market."
Polymarket uses a cryptocurrency called USDC. Unlike the speculative Bitcoin, USDC is a "stablecoin" pegged to the US dollar. However, similar to other cryptocurrencies, every USDC transaction on the blockchain is transparent and traceable. This is why we can observe the extraordinary betting behavior of Fredi9999.
When the answer to a question seems to be settled, Polymarket deploys another crypto-powered market mechanism - UMA. UMA establishes a system for proposing and challenging outcomes, adding another layer of fairness to the market. (Polymarket says that despite the endless controversies surrounding the 2020 election in the real world, resolving related bets on the platform has not encountered too many complications.) Ultimately, the blockchain's smart contracts - the automated, irreversible exchange mechanisms in the crypto world - will efficiently process all transactions and precisely distribute the winnings to the victors.
The slogan of Polymarket when it was first founded was "Harness the power of free markets to unveil the mysteries of the real world events that matter most to you." Copeland envisioned that his platform could predict the launch time of the COVID-19 vaccine and provide a more rational alternative to the chaotic social media environment. "News and misinformation are flying all over the place, and you have to figure out: Is this important? Is this real information or just nonsense? You'll often find that some things are being hyped up, but the market is completely unmoved, telling you, 'Hey, this is worthless.'" Copeland's logic is simple: the market will force participants to delve deeper, and the overall market performance is often wiser than the so-called expert opinions. However, the self-selected bettors on Polymarket do not seem to have unlimited predictive abilities. First, after the U.S. Commodity Futures Trading Commission determined in 2022 that the platform was essentially an unlicensed derivatives market, using Polymarket in the U.S. became illegal. As a result, Polymarket chose to settle with the government, paying a $1.4 million fine and banning U.S. users. Nevertheless, many people still use VPNs and other tools to circumvent this ban. (Kalshi and other election betting platforms face similar survival challenges.) Furthermore, although there are now convenient ways to bet using traditional bank accounts, the platform still requires users to be proficient in using cryptocurrency wallets. Consequently, Polymarket's user base exhibits a certain bias: they are mostly non-U.S. citizens, male, and heavy internet users. Therefore, it is not difficult to imagine that many Polymarket users betting on the U.S. election may not be very familiar with U.S. politics, and their views may be influenced by the cryptocurrency community's preference for Trump. After all, Trump has directly courted the industry, even launching his own Non-Fungible Token, while Harris seems less direct in comparison. The "whales" supporting Trump appear to fit this pattern. As one user involved in the election betting told me, "It's hard not to notice who he is, and many 'whales' know each other and communicate with each other." Polymarket's super users have noticed these trades and successfully traced them back to the same source. One of the most active bettors on the site, a user named Domer (who bet in support of Harris), had a brief chat with Fredi9999 on Discord, and using ChatGPT analysis, speculated that the user was likely French. Reuters later reported that these accounts were indeed held by non-U.S. citizens. When I asked Copeland whether these factors could reduce the accuracy of Polymarket's election betting, he insisted that the level of political understanding is not directly related to this. "According to the price discovery theory, even a large group of people with limited political knowledge can make more accurate predictions than a single expert through trading." (In 2020, Polymarket's bettors believed Trump had a 36% chance of winning.) However, the Polymarket bettor I interviewed held a different view. He believes that the market's prosperity for Trump supporters is only half due to the wisdom of the crowd, and the other half is driven by these "whales." Could manipulating Polymarket's odds become a way to try to influence election results? Perhaps, as some observers have speculated, these "whales" are trying to provide a positive narrative for Trump and his supporters to motivate more voters. But when I mentioned this, Copeland stated, "No, the bets only reflect the outcome of the event, not a direct relationship with the event itself." His emphasis on rationalism seems somewhat at odds with the fact that Trump's followers have found it remarkably easy to manipulate many other politically-related platforms. However, if the market does not agree with a large bet, we should expect the market to self-correct. But so far, this has not happened. Although Fredi9999's bets have slowed or been reallocated to more hidden accounts, Trump's winning probability on Polymarket has continued to rise. Copeland is not concerned about this. "Everyone is making a big fuss, but it's only a 60% chance." He said, "It's still a close race."
