– Crowdsourcing The Crystal Ball – by Forbes’-s James Surowiecki – 2007-10-15
[…] So what’-s the catch? Only this: We’-re still not sure how far into the future prediction markets can really look, or whether they’-re going to be able to foresee the kind of world- or business-altering events that Tetlock, for instance, asked his experts about.
So far, prediction markets’- track record has been built on predicting events that will occur in the near future, and where the range of variables that might determine that future is reasonably small and well defined. (Elections, sales forecasts and product launch dates all fall into this category.) But prediction markets haven’-t, for the most part, been used to try to predict things like the fall of the Soviet Union, and so it’-s not clear whether a market would really be able to foresee events that represent a dramatic break with the past, rather than an evolution from it.
This hardly means that prediction markets are of little use: The kind of forecasting problems that these markets are good at are fundamental to any business. Using prediction markets internally should have a beneficial effect on a company’-s bottom line.
But it is fair to say that we don’-t know enough yet to say that prediction markets really solve the crystal-ball problem when it comes to the long-term future. What we need now is to start using prediction markets to ask bigger questions, which will eventually help us understand what the problem with forecasting really is: Is it how we’-re trying to predict the future? Or is it that we’-re trying to predict the future at all?
What do you guys/gals think? I’-d go with the idea that it’-s quite impossible to use prediction markets to forecast the long-term future.
It really depends upon what one means by the word “dramatic.”