Introduction to Prediction Markets

I wrote an article in the June 2016 issue of Microsoft MSDN Magazine titled “Introduction to Prediction Markets”. See

A prediction market is used to make predictions in hard-to-predict situations such as a political election. In the article, I explain what prediction markets are like this:

Suppose you want to predict the outcome of an upcoming championship football game between the Xrays and the Yanks. You find a group of 20 football experts and give each of them $500 in tokens. The experts are allowed to buy and sell shares of each of the two teams, in a way that’s somewhat similar to how the stock market works.

When an expert buys shares in one team, say the Xrays, the price of a share of that team increases and the price of a share of the other team decreases. Over time, the experts will buy and sell shares of the two teams until prices stabilize, and then you’ll be able to infer the probability of each team winning.

You halt trading the day before the championship game. After the game is played and the winner is determined, you pay experts who have shares in the winning team according to the last price of the team when trading closed. Because the experts know they’ll be paid, they have incentive to give their true opinions during trading. What I’ve just described is called a prediction market.

In my article, I explain the math equations for determining the cost of shares, and the probabilities of each outcome winning.


This entry was posted in Machine Learning. Bookmark the permalink.