Does Group Wisdom Outperform Individual Wisdom? What We Can Learn from the Film Industry

Delfiner, Pierre, Total E&P, Paris La Defense Cedex, France

It is widely accepted that a group of experts is collectively better at estimating than an individual expert. Indeed there is more information and more power in several brains than in a single one, but in technical or scientific matters an individual may well be right against all others. This is especially true for disciplines such as geology which involve “flair”. It would be very interesting to find out from real oil exploration data.

Short of such data we will investigate this question using a dataset borrowed from an industry which is at least as risky as petroleum exploration, the film industry. Every Tuesday a group of about 40 media professionals log on a website and make forecasts on the atten­dance of new movies on their first day of release, i.e. the next day, in the Paris region. When results become known, scores are computed and a ranking of the players is established. As in petroleum exploration, the level of information varies among players and they analyze it differently. While each movie is unique, a key component of the forecast lies in the use of analogs.

A database of about 100 movies (prospects) evaluated by 40 players (geologists) is available to answer various questions, and in particular this one: Does Mister Average, our new player, consistently beat the other players? How do players compare among them­selves? If the estimates were used as bids in an auction, would the winner necessarily over­pay, and by how much? We may learn something from movies.