Tucson Citizen

Prediction markets being used by companies

Published: 07.11.2007
Want to know whether Hillary Clinton or Barack Obama will win the Democratic nomination in the 2008 presidential election? How much money Transformers will rake in at the box office? Or whether the New York Mets will win the next World Series?
Thanks to a growing body of academic research and media attention over the past decade, interactive prediction markets have become a popular way to try to answer such questions.
The basic concept is like Fantasy Football. Instead of trying to predict player performance, participants buy "shares" of certain outcomes - the topic can really be anything - and the person whose guess most closely matches reality wins. The payoff for the people creating the markets comes when they sell these predictions as market research.
A Scottsdale entrepreneur has recently gotten in on the trend with his own video-game prediction market.
The simExchange tries to answer questions such as how many Xbox 360 units will Microsoft Corp. sell in a particular month. Or will sales of the video game Final Fantasy XII surpass those of God of War II?
The idea for the simExchange came to founder Brian Shiau at Princeton University, where the Arizona native earned a bachelor's degree in economics and a certificate in finance in 2005. Shiau launched a beta version of the simExchange in November as a way to marry his passions for video games and the stock market.
"I figured that there was a way to use the (prediction market) concept in video games," the 23-year-old says.
It also didn't hurt that he could provide an outlet for hard-core gamers to bet on how successful a new video game or console system would be while providing meaningful information to the game developers and console manufacturers.
In many respects, the simExchange works like other popular prediction markets. One of the best-known prediction markets is the Hollywood Stock Exchange.
Started in 1996, the Hollywood Stock Exchange was formed to predict how new movies will do during their first four weekends at the box office. Users buy and sell fake shares of movies based on news related to the progression of films that are in production.
Just like the actual stock market, certain events might trigger the value of a movie's shares.
For example, news of a particular actor signing on to a movie could cause the film's share price to rise or fall, depending on the actor's popularity.
"The original intent of the exchange was to basically facilitate this interaction . . . between moviegoers and the studios," says Alex Costakis, managing director of the Hollywood Stock Exchange LLC in Los Angeles.
Today, the site sells research reports and demographic data to movie studios in the early stages of developing films.
Like most prediction markets, the simExchange does not use real money. Instead, users sign up to buy, sell and trade shares of video games and game consoles using a fake currency called Dragon Kill Points, or DKP.
The video-game industry is particularly fitting for the prediction-market model, Shiau says, given the ferocity with which fans argue over which new gaming system will come out on top after their releases. There are even references to DKP in other video games.
User Robert Witko says he joined the simExchange because it provides "more intelligent debate on the console wars."
"At least with the simExchange there's research going on behind it," says Witko, 19, a software engineering major at the Rochester Institute of Technology in Rochester, N.Y.
While others may question the significance of a market in which users trade imaginary shares of video games, "That's actually the point of it to me," he says.
"It's a game, and as gamers, that's what we do. We play games. It's just a way of tying our hobbies into the gaming industry."
Witko's dedication - he logs on to the simExchange for at least an hour every day to check the performance of his stocks - illustrates one of the key ingredients needed for a prediction market to be successful.
Market creators can't collect data to sell to companies unless their market provides some entertainment value to entice users to participate, says Robin Hanson, an associate professor of economics at George Mason University in Fairfax, Va.
"The main issue is what do traders get out of participating in this market?" asks Hanson, who has been studying prediction markets since the late 1980s.
He was the chief architect of a prediction market developed for the Pentagon around 2003 to determine the likelihood of a given set of geopolitical events. However, the project was scrapped after critics - who dubbed it the "terrorism futures" market - complained that it may lead terrorists to think they could make money from the predictions, thus encouraging attacks.
A successful prediction market also needs a diverse group of participants, says Richard Borghesi, an assistant professor of finance at Texas State University in San Marcos, Texas.
As of May, Shiau says the simExchange had about 3,000 registered users.
Shiau says his goal is to eventually make money from the information gleaned on the exchange, but whether the video-game industry will bite remains to be seen.
"I really don't know if that's something that would become a very valuable resource to the industry itself, at least not right now," says James Brightman, lead business editor of GameDaily BIZ. The online publication tracks video-game industry news.
A major obstacle for the simExchange is obtaining sales figures, which can be hard to come by in the video-game industry, Brightman says.
Part of the problem is that NDP Group, the primary company that tracks retail sales for the video-game industry, only releases monthly sales figures.
Shiau acknowledges the challenges involved in obtaining current sales figures, but notes that the simExchange structures its contracts differently than prediction markets such as the Hollywood Stock Exchange or HedgeStreet, a prediction market for hedge funds.