Deliberation in a Prediction Market

Lecture by Eric Pacuit (University of Maryland), organized by the Department of Theoretcial Philosophy

As epistemological and decision-making agents, groups suffer from a bipolar disorder: they are capable of displaying levels of knowledge and rationality that surpasses that of their individuals, and yet they also are capable of being more ignorant and irrational than the worst of their individuals.

Whether a given group functions as a Dr. Jekyll or a Mr. Hyde depends on many factors, such as the task being performed, the nature of the individuals that compose the group, how the individual contributions to the task at hand are combined, and so on. Much of the research into the so-called *wisdom of crowds* focuses on developing methods for getting the best out of groups, while avoiding the worst. There are now many known methods for improving group judgements and decisions, and these methods range from recruiting smarter individuals, to maximizing diversity, to improving elicitation protocols, to improving aggregation methods. In this talk, I focus on two particular methods for harnessing collective wisdom: group deliberations and prediction markets. 

Our method of investigation will be by a series of computer simulations. In this talk, I will develop a model of deliberation groups and prediction markets, and show how the two can be combined.  Using this computer simulation, I will explain the pros and cons of deliberation groups and prediction markets, and argue that the two methods can be united to form an amalgamation that is superior to the two individual methods. This is joint work with Aidan Lyon.

When & where?

Friday, June 3, 1.15 pm
Faculty of Philosophy, board room