Why can a prediction market get it wrong?

A prediction market can get it wrong because it is only as good as its crowd, its trading and its rules. The price is a summary of what traders do, and traders can be thin on the ground, ill-informed, outweighed by one big bettor, or nudged by someone who wants the price to move.

The crowd itself can fail. Some questions require specialized knowledge most people do not have, so the crowd can sometimes be very wrong. And crowds can be skewed by peer pressure, panic and bias when opinions lack diversity. Then there is liquidity, the ability to buy or sell quickly without causing a drastic change in the price. In a thin market, one large order can move the price a long way.

Manipulation is the dramatic case. It occurs when someone intentionally alters supply or demand to influence a price. In the 2004 Tradesports presidential markets, an anonymous trader sold so many contracts on one candidate that the price was driven to zero. If it was a deliberate manipulation effort, it failed, because the price rebounded rapidly. A 2005 experiment even found that manipulation attempts can end up increasing accuracy, because they give others more profit incentive to bet against the manipulator.

In October 2024, the Wall Street Journal reported that moves in Polymarket's election odds might be a mirage created by four bettors with about $30 million of wagers on one candidate, though the bets were not necessarily nefarious. Polymarket later said the four accounts were controlled by one French trader and that it found no evidence of efforts at market manipulation.