Monday, September 13, 2004

The Wisdom of Crowds

On a slightly less eerie note, I have to mention that I've come across an excellent book, The Wisdom of Crowds (James Surowiecki, Little Brown, 2004)

One sentence summary: despite the common belief that crowds operate at the lowest common denominator of intelligence, if handled right, a group can make decisions that are better than even the most intelligent person within the group.

Think about that. Most of us (me included) have a healthy skepticism about the ability of a large group of people to make a rational decision, let alone the best decision. We think about mobs, we remember our experiences sitting in various committee meetings where no matter how intelligent the individuals in the committee were, the group level of intelligence (as expressed in decisions made) seemed to gravitate to the lowest common denominator. We have our own (often elitist) distrust of "the masses", and the opinions expressed by the masses. I see this all the time in my work - in fact, I've expressed this opinion several times. We tend to trust the single wise leader, or an expert, over the (we assume) vast, ill-informed, uneducated masses.

But Surowiecki argues that, co-ordinated in the right way, the collective decision generated by a group is better than any individual's, even (and especially) the experts in a field. He suggests that the correct means of co-ordinating a group's decision is not via consensus, but via aggregation. If you've sat in a committee, you're familiar with consensus - discussion, debate, and finally a resolution of differing and diverse opinions into one "best" answer (with varying degrees of compromise or caving-in, depending on how tyrannical your chair is). Surowiecki points out the many pitfalls of this method - strong voices tend to dominate (strong either through persuasiveness, or by virtue of position or expertise) which herd the diverse opinions into one corral; once a majority of people have declared support for one idea, others tend to suppress their dissent; and discussion tends to reinforce established ideas, rather than encourage new ones, or recognise that the old ideas are no longer relevant (what he calls "groupthink".

Aggregation, on the other hand, is as simple as voting. It is the averaging out of all individual opinions, and Surowiecki offers some very convincing case studies of how this actually has worked in real-life - here's a link to an extract about this.

I like this idea. I like it because it appeals to my own fundamentally democratic beliefs. We disparage and underestimate The People too much - the very term "The Masses" has connotations of mob rule ("the blind masses") and irrational behaviour, but Surowiecki shows that properly handled, the whole is more than the sum of its parts, and we can trust the wisdom of the crowd.

Aside: one of the things mentioned in the book is how trading markets can be used to value almost anything, and how they predict future events more accurately than experts. For example, he mentions the Hollywood Stock Exchange, where you can buy shares in actors and movies, and where the aggregated opinion of the masses (expressed in the value of a movie or actor's stock) has turned out to be a reliable predictor of how well these movies and actors do.

BlogShares is another example of this - an online "stock market" for buying fictional shares in blogs, where the share price of a blog should (and probably does) indicate a blog's potential for growth quite accurately (after reading the book, I wonder if this whole blogshares thing isn't some corporation's sneaky means of identifying potentially profitable blogs so they can invest in them. hmmm.) Unlike HSX, Blogshares uses "play" money, and so players only get kudos and not real profits, but it's still interesting. Here's how this blog is doing (I wasn't even aware that this existed until T pointed it out to me).

No comments: