Polymarket accurately predicts the US election, but experts say the encrypted betting platform "has gone too far".

On November 14th, according to a blockchain analysis company and the trader himself, a mysterious trader made a profit of about $85 million by successfully betting on Trump winning the presidential election, which is tens of millions of dollars more than previously reported. Research by Chainalysis shows that the trader, known as the "Trump whale," has 11 accounts on Polymarket, more than the previously known four accounts. Polymarket is a cryptocurrency-based betting platform. In recent months, as more bettors flocked to the platform to bet on the US presidential election contracts, Polymarket's popularity has soared. Polymarket states that the "Trump whale" is a French national with extensive trading experience and a background in financial services. Polymarket is a cryptocurrency platform where anyone can bet on almost any event. It has proven to be the best platform for picking winners in this year's US election, as its betting data correctly predicted Trump's landslide victoryunlike what most pollsters predicted. However, Bloomberg columnist Lionel Laurent points out that as French and American regulatory authorities increase scrutiny on Polymarket including an investigation by the FBI into the platform, which is accused of having political motives it is necessary to curb the enthusiasm for what some supporters of prediction markets call a "new era" for the industry. It can be further predicted that if there is to be a shake-up in the polling industry in the future, it is more likely to come from artificial intelligence than from betting platforms. On November 14th, it was revealed by sources that the FBI had seized the phone of Polymarket co-founder and CEO Shayne Coplan. Polymarket was a popular betting platform during this year's US presidential election. Sources say that federal agents conducted a raid at Coplan's Manhattan residence early Wednesday morning. The reason for the FBI's search of Coplan is currently unclear. A company spokesperson stated, "Polymarket is a completely transparent prediction market that helps ordinary people better understand the things that matter most to them, including elections." The idea behind prediction markets is simple: the flow of funds from bettors provides intelligence more valuable than surveys from polling organizations, and the most accurate traders are rewarded financially. Like betting in financial markets, constantly changing odds can continue to be in effect, and almost anything can be bet on, as evidenced by Polymarket's series of wild bets, including whether nuclear weapons will be used in 2024. However, there are limitations, not only because election gambling is a legal gray area. For one, these are bets, not polls they require both privacy protection and liquidity, making them relatively narrow in measuring public opinion. Another reason is that markets can be wrong, just like polls. There is also the risk of manipulation; economic incentives do not only attract those seeking the truth. Recent examples include large bets on Mitt Romney in 2012, which made that year's election appear more tense than it actually was. US regulatory agencies have issued warnings about the risks to election integrity, and Polymarket was fined in 2022 for operating an illegal derivatives market. Polymarket's terms of service describe it as a Panamanian entity, not available in the US, but reports suggest that US traders are using the platform. However, all of this does not absolve the polling industry itself of its shortcomings: for the third consecutive election, experts seem to have underestimated support for Trump. Interestingly, market research firm Ipsos SA coined the term "nouveau nihilism" to describe 2024, based on the gradual disappearance of the dreams of economic independence and homeownership among the younger generation; this ultimately better summarized the final results than actual surveys. But we should not overstate the polling industry's failures. Author of "Lost in a Gallup," W. Joseph Campbell, says that polling performed better this year than in 2020 and 2016. Polling in 2020 was the worst in 40 years. He said that many models have been tested to more accurately reflect Trump's support, with a final error of 1-2 percentage points, which is a relatively commendable result. Nevertheless, it is clear that polling institutions once again underestimated Trump's support this year. The question now is, where will further improvements come from? Former chairman of YouGov Plc Peter Kellner points out that as landline usage declines, response rates in traditional polling surveys have significantly fallen, giving control of the sample to survey panels. According to Kellner, one possible answer may lie beyond prediction markets: in the field of artificial intelligence. He notes that AI has the ability to handle large amounts of data, which may reveal correlations that no one had thought of; by using non-demographic data such as credit card transactions to build more accurate models, and by analyzing existing polls to find discrepancies and errors. AI also has risks, such as the possibility of manipulating public opinion Microsoft issued warnings during the election season about the risks of false information, but it could also bring positive impacts.