The Basic Principle of Prediction Markets
Prediction markets are a method to aggregate beliefs about the outcome of future events. They provide financial incentives, to tell the truth.
Let us assume you want to forecast the outcome of a presidential election with two candidates A and B. You could ask people in the street for who they will vote. But your interviewees have no reason to tell the truth. This is where prediction markets enter the stage.
The market operator creates a system where you can buy a complete set of shares. The complete set consists of a share A and a share B. Share A represents the outcome that candidate A wins the election. And share B represents the victory of candidate B.
The market operator sells this complete set for 1 ETH. He promises you to buy back the winning share for 1 ETH after the event takes place. The losing share is worth noting.
So, if you buy a complete set, you would not win or lose anything. But if you believe that candidate A has a probability of winning of 80%, you could calculate the expected value of the share A and B, respectively.
Share A would be worth 1 ETH * 0.8 = 0.8 ETH and share B would be worth in your opinion 1 ETH * 0.2 = 0.2 ETH. The sum of all expected values must equal 1.
If another participant thinks A’s probability is 95%, he would value the A share at 0.95 ETH. This creates room for trading the shares. You would be willing to sell your share A for at least 0.8 ETH, and he would be willing to buy it for a max 0.95 ETH.
The resulting price is the overall probability of candidate´s A victory. The aggregation works best if there are many traders trading every time they spot a share that is cheaper than their expected value.
The underlying theoretical model is the market efficiency hypothesis that states that asset prices reflect all available information. However, this has gained some criticism. Critics say that people have bounded rationality and are not able to process all information. Besides that, people tend to herd behavior and follow the crowd.
