Improving the wisdom of crowds


A new approach to predicting the outcomes of major events could give people an incentive to seek out more diverse sources of information, according to a new study.

Market economies and democracies rely on what is referred to as crowd intelligence, or the wisdom of crowds. This is the idea that large groups know more about what is best than a single individual. This knowledge is the basis for stock markets, betting exchanges and special investment vehicles called prediction markets.

However, a new study led by Dr Richard Mann from the School of Mathematics at Leeds highlights the recent failing of prediction markets in a number of high profile events, including last year's Brexit referendum and US presidential election.

Consistent inaccuracies about such events points to possible flaws in the assumptions made about crowd intelligence. The study suggests the lack of diverse information sources among decision-making individuals has contributed to this failure in prediction markets, and may also undermine other collective endeavours such as academia and democracies

Dr Mann and Professor Dirk Helbing at ETH Zurich have developed a new theoretical model that overcomes this problem by giving people an incentive to seek out new sources of information, and an extra reason to share it.

In this new prediction market system, people would only be rewarded if they expressed accurate views but were also in the minority.

This "minority rewards" system encourages groups to bring together wider sets of information, leading to more accurate collective decisions.

In an article written for The Conversation, Dr Mann describes his "minority rewards" model and gives insight into why current forms of collective decision-making may be prone to failure. 

Further information:

Image credit: David Ragusa

Dr Richard Mann’s article A simple reward system could make crowds a whole lot wiser was published on the Conversation UK 1 May 2017

The research paper ‘Optimal incentives for collective intelligence’ is published in Proceedings of the National Academy of Sciences.

For interviews and additional information please contact University of Leeds Media Relations Officer Anna Martinez on or +44 (0)113 343 4196.