The Wisdom of Crowds idea took off in 2005 when James Surowiecki published a book with the same title. His thesis was that, for many tasks, larger groups of non-specialists can be smarter than a smaller group of specialists. Both Jens Krause and Andy King, together with colleagues, have tested this idea by asking people how many sweeties or marbles are in a jar. It seems that, although the guesses of the individuals that make up the group are wide-ranging, the mean or the median is often close to the true answer.
I am currently writing about the Wisdom of Crowds idea in the context of betting on football. Last week, I decided to do a test in Stockholm while at the zoology department there. I asked non-footballing experts if they could predict the number of corners in the FIFA Women's World Cup match between Sweden and the USA. This task must be harder than sweets, since no-one knows the answer. It is in truth impossible to predict.
But when forced to predict, the group didn't do too badly. The histogram of guesses shows that they are wide-ranging. From zero up to 34. What is interesting is the mean and the median. The average guess was 11.26, the median was 9. When we went online to Bet 365 and checked the odds, 9-10 was their spread for corners. The bookmakers estimated that scoring less than 9 corners was equally likely to scoring more than 10. The Stockholm zoologists, who had no expert knowledge, made the same guess as the gamblers and bookmakers who were determining the corner odds.
But both the bookmakers and most of the zoologists were wrong. There were 18 corners in the match, more than is usual for a football game. In the previous Sweden game against Nigeria there were 12 corners, two of which resulted in goals for Sweden. So it may be the spread moves up for the next match against Australia.
The experiment is interesting because it suggests that crowds of non-experts can come up with reasonable answers to questions. However, crowds can't truly predict future events like numbers of corners. There is always too much unpredictability involved.
A 'Wisdom of Crowds' experiment takes only a minute to perform and can make watching a football game a bit more fun. Please do your own, and tell me how it went. You can also have a go at Simon Garnier's count the dots experiment.
As a little incentive in my own experiment I offered to buy a bottle of wine for the person who got the corners right. A bottle of wine is soon on its way to Love Dalen, who guessed 17. Well done.