Prediction Markets: What are they and why do they Work?

Rate Blog: 
Average: 5 (1 vote)

Conventional forecasting methods to predict events analyze historical data using statistical tools (like time series analysis etc.) in conjunction with sampling of opinions through polls and surveys. Research has indicated that Prediction markets produce superior forecasts (relatively lower prediction errors) than conventional forecasting methods. Prediction markets (also called information markets) are forums (we call ours an ‘opinion exchange‘) for trading (we have limited it to buying for now) contracts based on the outcome of uncertain events (questions in our context). The predicted outcome of each question is compared with the actual outcome and individuals are incentivized for calling it right (could be real, play money or other non cash incentives) and penalized for getting it wrong (you loose the opportunity to win !!).

James Surowiecki in the book The Wisdom of Crowds talks about how groups together can make more accurate predictions than any of its individuals.  An example provided in the book talks about a group, to a livestock show who could predict within a pound (average weight predicted was 1197 to the actual of 1198). Other researchers have suggested that the power of prediction markets can be explained because of the incentives for truthful revelation, aggregation of multiple sources of information, freedom of participation and market based incentives and algorithms at play for more truthful information discovery. Some of the issues with convention methods like biases, convenience sampling, ease of manipulation etc. are also overcome. There are a set of detractors for predictor markets too, however there is a healthy debate and research which is ongoing which is improving the quality of predictions.

Given the proven success of prediction markets in determining the outcome of events (e.g. the record in predicting the Presidential elections in USA on the Iowa Electronic Markets) there is increased adoption by modern day corporate as well (e.g. Google, Yahoo, HP etc.). Application of Prediction Markets have been several in diversified public fields like politics, economics, sports, product sales, weather forecasting, personal decisioning as well as private corporate applications from determining likely dates of product releases , deciding on new product launches, selecting advertising campaigns and to make investment decisions.

Typically the prediction market events are short duration and the final contract value determines the probability or a likelihood of an event taking place. The accuracy of predicting events is typically found highest in binary events (this will happen or not?), however combinational applications which are combination of events and other combinations can also be predicted based on various algorithms. IndiaPredicts has implemented a winner takes all algorithm, however a stock exchange like futures, auction and MSR are other ones.

Future potential applications of Prediction Markets are several (including social ventures or determining government priorities) and what is needed are efficient, robust and impartial mechanisms to collect information from different participants. Something holding back the growth of Prediction Markets, is that in several countries it is considered illegal (including USA) to trade with real money on prediction markets (as it is associated with internet gambling) however as online communities and social media platforms grow there is bound to be a relook from regulatory and government authorities. Research has indicated that Prediction Markets with non financial benefits (like played with Play Money) can be equally effective in predicting future events. is India’s first formal Prediction Market and without doubt we will work completely within the constraints and regulatory frameworks laid down by the Indian government!!


  • Surowiecki, James (2005). The Wisdom of Crowds. Anchor Books.
  • Arrow, Kenneth J., et al. "The promise of prediction markets."