The SimExchange is a website which uses play money, and markets like those we have discussed in class to forecast video game sales. Predictive markets are a powerful use of concepts from class which one could imagine applying to a huge range of topics, including ones far more important than video games. In fact, there are predictive markets about other topics such as the notable Iowa Electronic Markets which uses real money markets to predict events such as political outcomes. See this article for a nice overview of the SimExchange’s concept, and history.

Users buy and sell stock based on their attitude toward a gaming system or piece of software. This allows gamers to collectively give their opinions of games, and it is reasonable to think they often know things the experts do not. The mechanics of this are very similar to the those of a real stock market. You can look at the bid and ask prices, and submit your own bids. Users also are rewarded for producing relevant news articles (such as work on a new game being announced) and submitting other content.

So, how effective are predictive markets? According to the site’s main page:

The simExchange applies prediction market technology to the video game industry to generate better quantitative and qualitative indicators to what’s coming in the video game world.

What is prediction market technology?

Prediction markets are basically stock markets that trade for the purpose of forecasting and are one of the most effective ways to forecast the future.

Where’s the proof?

The Iowa Electronic Market applies prediction market technology to predicting the US Presidency. Compared with 596 polls, the prediction market was more accurate 75% of the time.

HP applied prediction market technology to predict its printer sales. Over 3 years, the prediction market outperformed the company’s own expert analysts 75% of the time.

Many Fortune 500 companies operate prediction markets to forecast product potential and make business decisions.

The site produces a lot of interesting data, presented in their research reports.

For example the December 2007 Review Report claims:

Software sales came in at $2.37 billion, a 36% increase year-over-year and in line with the prediction market’s expectation for $2.38 billion in December software sales, a mere -0.28% difference.

This seems like an extremely impressive prediction, displaying the power of these markets. The same report also shows that the predictions are not perfect, Call of Duty 4 sold 58% more units than expected.