A fascinating article from Kevin Pullein in today's Racing Post got me thinking. Pullein used a book called "The Wisdom of Crowds" by James Surowiecki to argue that the more people to whom you ask a question, the more reliable the answer becomes.
The theory is introduced by citing the findings of one Francis Galton who visited a country fair in 1906 at which people were invited to guess the weight of an ox. Galton asked for all the entries and discovered that the average of all the entries was within 1 lb of the correct answer. Now, there were some experts (farmers) at this fair but they were vastly outnumbered by non-experts.
Using Galton's experience and other evidence, Surowiecki argues that crowds are actually very wise but only under two conditions - one, the crowd must be diverse and two opinions must be given independently.
This explains how polling should work and why opinion polling, if done properly, can be a very reliable guide.
It's also a guide of sorts to betting - how often are odds manipulated by punters NOT acting independently - everyone sees a pride shortening and jumps on the gamble - and can it really be argued that bookmakers are a diverse bunch ?