Saturday, 15 December 2007

Not Quite the Madding Crowd

Just finished reading The Wisdom of Crowds by James Surowiecki. Very good reading, and I now wish I knew how to pronounce his name correctly, so I could drop it in intellectual conversations.


  1. ehh

    most of this stuff has been done before in game theory and its applications.(why is the line you stand in the longest, if you had to meet in paris where would you meet etc), from whatever little i have read describing the book from the posted link.

    also, his little example of 'who wants to be a millionaire' ignores the fact that usually the audience is asked easier questions, while the harder questions go to the individual phone call.

    the excerpt on the path of the agents is interesting, but i'd want to see how this guy computed the group strategy. It could be that the way the group strategy is set up enforces the best i.e. minimum step strategy.

    after all, when social planner optimizes he will always reach optimum when individual does not always...

  2. It's a popular-economics type book a la freakonomics/blink etc. Those sort of books are meant to make the academic stuff more accessible, not drop something earth-shatteringly new on an unsuspecting public. Besides, blurbs are written to make books seem more accessible as well as up the 'cool' factor of the book - a consequence of Freakonomics' success I'd say. Hence those examples.
    The point of the book is that given certain necessary conditions-diversity, independence and decentralization amongst the individuals that form the crowd, as well as a mechanism to aggregate the opinions of the group, the final decision/answer of the crowd on average turns out to be better than the smartest individual in the group.
    As for social planning, the point he makes is that you may not need a social planner for social optima, even without assuming perfectly rational forward-looking individuals.
    Thanks for the comment though - good to know that someone's actually reading my blog :-)

  3. wouldn't you also need a sufficiently large number of people?

    i may be stretching it but i think some kind of auctions exploit this kind of decision making.

    anyway it's interesting though i suspect some of the stuff isn't new like you mention.

  4. Actually the book also deals with smaller groups eg shop-floors (getting the workers on the shop-floor to make decisions rather than leaving it to supervisors) or smaller teams (groupthink as a negative example).
    Yes, auction markets are a great way of aggregating individual valuations. In this book he discusses betting markets, which I don't think have received much coverage from the popular economics press so far.

  5. by that last statement, I mean that the betting markets have not really been covered in Freakonomics-type books too much. Though I could be wrong. I haven't read 'em all...