headerphoto

Archive for January, 2008

Preface

“Wanna bet on that?” This playground challenge attests to the sincerity of the questioner. Acceptance of the challenge attests to the sincerity of the target as well. A refusal to bet, meanwhile, suggests insincerity. In the absence of any information about the merits of the dispute, at least, such a refusal gives onlookers some reason […]

 

The Media

If democratic institutions are tools for translating public opinion into public policy, then the most important institution might be the one that contributes most to forming public opinion in the first place: the media.1 Constitutional protections for free speech and a free press such as the First Amendment to the U.S. Constitution seem to reflect […]

 

The Probability Estimate Prediction Market

Prediction markets may take many forms, and some of the designs for prediction markets that this book considers might not merit being described as markets at all. The term prediction market, however, is descriptively accurate at least as applied to the design most prevalent today, which I call the “probability estimate prediction market.” As the […]

 

The Theoretical Case for Prediction Markets

Risk and transactions costs might affect not only the amount that individuals will be willing to pay at auction but also their bid and ask prices in the subsequent market. But there is a strong theoretical reason to believe that prices derived from market activity will provide more reliable probability estimates than prices derived from […]

 

Iowa Electronic Markets and TradeSports

We can obtain an approximate sense of the accuracy of probability estimate prediction markets by eyeballing them. Let us consider the most venerable collection of prediction markets, the Iowa Electronic Markets, created by professors at the business school at the University of Iowa. Although the IEM features a range of prediction markets, the most famous […]

 

Pari-Mutuel Wagering

The probability estimate prediction market represents an improvement over what might have been the best earlier approach for aggregating bets on a particular outcome: pari-mutuel wagering, commonly used in horse racing. In pari-mutuel wagering, a bettor can place money on any possible outcome. For example, in a horse race, a bettor could place $1 on […]

 

Reduction of Heuristics and Biases

Consistent reporting of prediction market predictions and trends could demystify the sports pages, reducing events that require much attention today to mere blips on a market chart. Some prediction markets on TradeSports, for example, predict not the result of individual games but the outcome of entire seasons. For example, before the beginning of the Major […]

 

The Numeric Estimate Prediction Market

The prediction markets illustrated so far are designed to predict probabilities, but sometimes it is useful to aggregate individual predictions of numbers that are not probabilities, for example, how many games the New York Mets will win this year or how many people in Love Canal will die of cancer over the next ten […]

 

The Accuracy of Prediction Markets

The experience with the Iowa Electronic Markets for vote share suggests that such markets have relatively strong accuracy. In a paper considering fifteen elections, Joyce Berg and coauthors show that the markets on election eve were more accurate on average than were final preelection polls.40 The average absolute error of a vote share forecast derived […]

 

Manipulation

Though the creator of a prediction market cannot, absent fraud, guarantee a preferred result, there is a danger that individual traders in a prediction market might be able to manipulate forecasts. One way to do so would be to disseminate false information. To the extent that this approach is successful, it works because it changes […]