How one trader’s $26m Trump bet exposed Polymarket’s prediction problem

When a trader on Polymarket bet $26 million this autumn that Donald Trump will win the US election, crypto fans of the former president gushed with excitement.

Unlike traditional polls, which show that Vice President Kamala Harris had a narrow advantage over her Republican rival, the Polymarket wager has helped propel Trump 32 points ahead.

As Trump supporters revelled in the mojo and piggybacked on the whale-sized bet, political analysts scrambled to understand what was going on — did Polymarket bettors know something pollsters had missed?

Moving irrationally

Harry Crane, a statistics professor at Rutgers University who specialises in prediction markets, gives two possibilities:

“There’s a whale just buying up a bunch of shares, and the markets are moving irrationally,” he told DL News.

“Or there’s a smart bettor who is buying up a bunch of shares, and the markets are moving rationally based on that person’s information.”

In other words, who knows.

One thing does appear certain — the Trump whale has cast doubt on Polymarket’s efficacy as forecaster.

If one trader can skew a $2.6 billion election bet so out of whack with other models, how useful are its predictions?

“I don’t think the Poly price is necessarily accurate,” Douglas Campbell, an economics professor at the New Economic School and founder of Insight Prediction.

‘Betting markets will become more of the priority... not in four years, but in eight years.’

—  Harry Crane, Rutgers University

Polymarket did not respond to requests for comment.

The whale’s impact is the latest turn in Polymarket’s rise from obscurity to the heart of the 2024 presidential election’s narrative.

Ever since the crypto-powered prediction market surged to $4 billion in total cumulative volume, it’s been featured on Bloomberg’s trading terminals, CNN, and CNBC.

“Studies show prediction markets are more accurate than polls in predicting election results,” tweeted Will LeGate, a growth lead at Polymarket, on Friday. (He later deleted the tweet.)

Market wisdom

Until now, most soothsaying in the quadrennial presidential cycle has come from national polling firms and news organisations that have spent decades honing their methods for divining voters’ intentions.

Yet the notion that Polymarket is manifesting the wisdom of the marketplace — an idea known as the efficient market hypothesis — has galvanised the entire prediction market sector.

“Polling is still a huge piece of data that’s being used in these markets,” Crane said.

“That being said, prediction markets will become more of the headline. Betting markets will become more of the priority if not in four years, then definitely in eight years.”

Founded in 2020 by Shayne Coplan, Polymarket enables punters to buy “yes” or “no” shares for various outcomes of events like sports matches, weather crises, and elections.

The shares analogy is apt — as users buy up “yes” shares for an event’s outcome, their price rises and serves as de facto odds for the event.

There’s a $244 million market, for instance, that asks: Who will win the Champions League this year?

Forecasting tool

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.

‘The idea from sharp bettors is you should probably just fade this guy.’

—  Douglas Campbell, New Economic School

Polymarket’s use of stablecoins to facilitate transactions also catered to a ready-made audience: the tens of millions of holders of Bitcoin and other digital assets.

In October, Polymarket shook up the election story when four accounts — Fredi9999, Princess Caro, Michie, and Theo4 — started amassing huge “yes” votes on Trump.

Last week, Polymarket said the four accounts were actually controlled by one French trader, according a a report in The New York Times.

A ‘yes’ vote

Right now, one “yes” share for Trump to win costs 66 cents, giving the former president 66% odds of winning. Compared to various national polls, that’s a huge discrepancy.

Polling platform FiveThirtyEight gives Harris a 1.4% lead. Nate Silver, the founder of FiveThirtyEight and now an advisor to Polymarket, also gives Harris a 1.2% lead.

Of the four whale accounts, Fredi999 has been the biggest player and has steadily placed bets since October 7, according to blockchain data platform Polygon Scan. Polymarket is built on the Polygon network.

A big part of the problem at Polymarket is its lack of liquidity.

Polymarket may have hosted $4 billion in volume cumulatively, but that also includes closed trades. There is far less in daily open trades.

Just over $300 million in open orders exist across the entire platform, including markets unrelated to the elections. This is a lot, but when one trader makes up more than 10% of all those orders in one direction, it distorts the market.

99% chance of winning

This is because Polymarket operates as an order book exchange. People buy shares from sellers. If there are more buy orders than sell orders, the price of the shares will rise until the next available sell order is found.

Though there are more than 178,000 monthly active users, the vast majority of orders are for less than $500.

This dynamic can create extreme market moves, like reducing the wisdom of the crowds to just one French trader.

On October 25, for example, another Polymarket user bought $3 million worth of “yes” shares for Trump winning the election.

Because there were so few sell orders to match his purchase, the price of “yes” shares briefly spiked to 99 cents, pushing Trump’s chance of winning to 99%.

When asked if Polymarket was efficient at predicting odds, Campbell told DL News there are still issues. He says it’s enough to take his chances and bet against Polymarket’s mega whale.

“The idea from sharp bettors is you should probably just fade this guy. He doesn’t know what he’s doing,” Campbell said, referring to Fredi9999.

Better information

Crane disagrees.

There’s no reason someone with that much money shouldn’t have access to better information, he told DL News.

“When someone with big money makes a big move, it is often because they know something nobody else knows.”

Liam Kelly is DL News Berlin-based DeFi Correspondent. Got a tip? Email at liam@dlnews.com.