How one trader’s $26m Trump bet exposed Polymarket’s prediction problem
AI summary of the linked articleA French trader behind four Polymarket accounts placed a $26 million bet this autumn that Donald Trump would win the US election, pushing the prediction market's odds for Trump far above national polls. Polymarket later said the accounts, including Fredi9999, the largest player, were controlled by one trader, according to a New York Times report. Harry Crane, a Rutgers statistics professor, said the market was either moving irrationally because of a whale or rationally on a smart bettor's information. Douglas Campbell of Insight Prediction said the price is not necessarily accurate, while Polymarket's thin liquidity leaves it open to distortion. |
Recommended by 1 curator |
1 comment  timSubmitter There’s a $244 million market, for instance, that asks: Who will win the Champions League this year? According to Polymarket, Manchester City, the English football club, has a 22% chance of winning. This means one “yes” share costs 22 cents. If they win the European title, that share will immediately shoot up to $1. If they don’t, that share drops to $0. The theory goes that if enough people bet on the market, especially clever sports analysts, the odds are a useful forecasting tool. People need to stop equating forecasting who wins a football match with forecasting who wins an election. An election is won because many people are, statistically speaking, swung to one side of a political spectrum. There's pretty much a persistence in this. Trump's group of supporters doesn't mobilize over night, this group has existed for a decade or more. It has been mobilized through labour. Same with Democrats. At the same time some sport events are decided merely by chance, some guy getting a good night's sleep, or whatever. You can effectively sample a large enough percentage of the population and you can actually predict the outcome of an election. That is as there is basically no way 300M people can coordinate who they're going to vote in as president. But in god-mode, as a statistican, you can totally build models that'll forcast the results. And please don't come at me with statistics on how public polls are right or wrong. Public polls are just there to manipulate the public, they aren't there to tell you the truth. Now, can you go and sample a sports team so much that you can predict who's going to win? I guess you could measure EVERYTHING, blood pressure, sleep, diet, etc. Yet I do think the law of big numbers will give you such an advantage over forecasting an election over forecasting who's going to win a football tournament. People need to stop equating sports betting with the election forecasting prediction market. These things aren't the same. And if you think they are, you're going to make other bad related decisions. |
More from dlnews.com on Kiwi News
|