Until the final hours of the Netherlands’ October 29 election, Polymarket traders were confident that Geert Wilders’ nationalist Partij voor de Vrijheid (Party for Freedom) would cruise to victory.
The market barely budged as Rob Jetten’s social liberal Democrats 66 led in every major poll. Then, within minutes of the first exit poll, D66 odds increased from 5% to 100%, wiping out millions of overconfident PVV longs.
Reuters reported on Thursday that with 98% of the votes counted, D66 and the PVV were both projected to take 26 seats in the 150-seat lower house of parliament. This is a loss of 11 seats for the PVV.
Kalshi was not much better, with traders overvaluing Wilder’s PVV by election day.
Data from Polymarket Analytics shows that the market has become a test of conviction rather than foresight as traders are losing PVV bets out of confidence. Many held steady for weeks, while a small group of data-driven participants quietly benefited from the late D66 surge.
During the recent US presidential election, all kinds of theories emerged as to why Polymarket was paying a premium to current President Donald Trump. Perhaps it was the involvement of crypto holders, who lean to the right.
One theory was that foreign money was trying to influence the vote by distorting the markets. This theory was further boosted when a French citizen using the handle “Theo” spread pro-Trump and pro-Republican posts on multiple accounts.
It turned out that Theo had no political agenda, as he told the Wall Street Journal. Instead, the self-described wealthy banker determined that there were gaps in national polling and instead commissioned his own, in which pollsters asked respondents how they thought their neighbors would vote.
The survey confirmed his thesis that the polls were wrong about Trump’s chances of victory, and he was confident enough to put up $30 million.
But there was no Theo for the Dutch election. There were many conviction traders who acted as effective counterparties with exit liquidity.
Accounts such as “WhiteLivesMatter” – whose usernames reflect their political views – drove thousands of people into PPV “yes” contracts and never backed down, even as the Ipsos and Peel.nl polls shifted decisively towards D66.
According to Polymarket Analytics, the situation remained unchanged for weeks. It was not the lack of information that doomed them, but the refusal to act on it.
That stablecoin was a sharp contrast to traders like “Visser” and “Siro2,” who pounced early on late polling data and made six-figure profits from the same volatility that crushed PVV loyalists. These participants used the market as a rational actor attempting to make money with trades, not a scoreboard for ideology.
Ultimately, prediction markets acted as mirrors rather than forecasters, reflecting the biases of their users. While Theo used voting to challenge the consensus, some traders ignored it altogether.
In a market with thin liquidity, the result was a real-time experiment in how markets can be rational in theory but irrational in practice, especially when conviction outweighs curiosity.
Read more: Polymarket is 90% accurate in predicting world events: research