Can You Predict the Future?

A new study aims to find out. You can participate.

Posted Aug 10, 2011

My dog.

My dog. He's not good at predicting the future.

I've often wondered since then how accurate forecasting markets are at predicting different kinds of world events. The theory is simple - the accumulated information of a bunch of people is often much more accurate than the opinion of any one person. And when we make people pony up their own money, they have an incentive to wager along with their confidence, and accuracy improves quite a bit. 

But the system has certain flaws. For example, psychologists have long known that individuals are most inaccurate at assessing risk at the extremes. According to Nobel winning economist Daniel Kahneman and Amos Tversky, we particularly prone to underestimate the likelihood of extremely probable events. I suspect that these biases are so strong that even with market incentives, the vox populi wasn't willing to commit to Cheney 100% (or, as I thought they should, 99.5%). So the wisdom of the crowds was approximately right, but wanted to hedge.

(For another example, see my previous column about market bubbles, linked above).

The ability to predict the future is appealing though. For a while the federal government considered using Intrade to predict terrorist attacks. The logic was that anyone who knew of the attack might try to make money on the site, and the government could monitor the action and predict the next 9/11. The public thought the plan was tasteless, and the plan was nixed.

The government is still interested in these questions though.

Uncle Sam is funding an ambitious new study you might want to participate in. It will be run by scientists from (my alma mater) UC Berkeley and University of Pennsylvania, but it takes place on the internet. The first round will gauge your predictions about world political events and, if you complete the study, will pay each participant $150. I am participating, and you, my loyal readers, are invited to join (as long as you have an undergraduate degree of higher from an accredited college or university). I am not involved in running the study, just an enthusiastic participant, but it fits in perfectly with the themes of the Decision Tree. So if you're intrigued, click here, or go to http://www.goodjudgment.info.

P.S. As I write this, Intrade has Obama at 50.6% for re-election, and Perry at 30.4% and Romney at 31.7% for the GOP nomination. If you think these are wrong, just join Intrade and play along with the markets. I'm going to sit this one out.