The Wisdom of Smart Crowds
The wisdom of crowds is an interesting topic that is very open to debate. Does it actually work? Can groups of people, experts and non-experts, really work together to create a good prediction market? James Surowiecki put together a great explanation, and a really good read, in his aptly named book, The Wisdom of Crowds.
So, it seems that there are many working examples of the wisdom of crowds online, but one just came across my feed reader that seemed to interest me more than others.
The site is called PicksPal. It's an online sports betting market with no money, only bragging rights. Users sign up place bets on almost all North American sporting events in a wide variety of categories and ways. It is rather cool and if you're interested in sports at all, it's probably worth checking out.
The interesting part to me, is not the betting, rather, that the PicksPal team noticed that some users, when grouped together, created very high quality predictions.
"... the PicksPal team noticed that a very small percentage of users tend to be correct in their picks significantly more often that they should be statistically. When they grouped these special users they found them to be a powerful predictive force." from TechCrunch
While this may seem like the wisdom of crowds at work, it's not exactly. The wisdom of crowds expects both experts and non-experts to be a part of the crowd. PicksPal is simply taking the experts and using their picks to create aggregate picks. "The Wisdom of Smart Crowds" is a more appropriate title for this site.
So, PicksPal is actually creating a paid service that allows you to get the data from this group of geniuses. I hope that they release the data after the sporting events are finished, so that we can see how the geniuse markets do over time.
While this is an extremely interesting usage of crowds to create expert data, to prove that the wisdom of crowds is true, wouldn't the aggregate of the entire user base have to be the same or better than a group of the top performing users?
Greg, a really interesting idea. Totally think you are right on about what is the right amount of people to show the wisdom from. Intriguing that they are taking the top subset. Wondering too if the mass number of all members might not be more accurate, at least according to the book. On the other hand would not that always just be the 50% on each side of the spread, seeing as how that is usually balanced??? Not sure.
Yeah, that’s an interesting question in itself. How balanced are these types of markets? Do they resolve in a 50 / 50 split or close to?
My take from the book was that the mass number of members should be even better than the group of geniuses. But, there are a lot of factors at play that may make that disputable and it is definitely not a fact!