I wrote a post about what Soros knew about prediction markets. The thesis: crowds with skin in the game produce better forecasts than pundits. The price aggregates scattered knowledge into a live probability. The price shapes the outcome it measures. Soros shorted the pound and his bet helped make it fall. Reflexivity.
That post covered the loop. This one covers who’s inside it.
Prediction markets create odds on sports, politics, wars, release dates, life and death. Polymarket. Kalshi. The pitch is simple: real money, real stakes, real signal. You can see the odds and weigh the risk.
Here’s what the pitch leaves out. The people betting on these outcomes often have influence over the outcomes themselves.
A political donor sees their candidate at 30% on Polymarket. Contracts are cheap, so they buy a large position. Then they fund the campaign.
The bet and the donation reinforce each other. The odds shift. Pundits cite the new price. Volunteers show up.
The donor moved the signal AND the outcome. They profited from both.
This is the part the prediction market evangelists skip.
The markets are truth-seeking. They do aggregate real information from people with real stakes. The research on forecast accuracy is strong. All of that is true.
It’s also true that the people with the biggest positions often have direct influence over the outcomes.
- A whale on an election contract can fund ads
- An insider on a product launch contract knows the date
- A government official betting on policy outcomes can shape the policy
The markets are accurate AND influenced by people who can move the thing being measured.
Most people land on one side. Idealists see the influence and conclude the markets are rigged. Finance types see the accuracy and conclude it’s pure signal.
Both are half right. The accuracy is real. The influence is also real. They coexist.
The part people miss is the “also.”
I covered prediction markets for Trends.vc and made a documentary about them going mainstream. The more I studied the space, the more I saw this pattern. The signal gets treated as objective because it’s crowd-sourced. The crowd is real. The influence that some of those participants have over the outcome is also real.
When you read a prediction market price, the question is: how many of the people betting can also move the result?
💰
Where are you reading odds set by people who can change the outcome?
