November 20, 2015

We’ve tried it all: tea leaves, tarot cards, palm reading. We’re obsessed with figuring out what’s going to happen before it happens. But can we?

Since 2011, the US intelligence community has sponsored a “forecasting” tournament called the Good Judgment Project, in which ordinary people take a crack at predicting global political events. Researchers hope that the best forecasters can teach us how to better anticipate the future.

Philip Tetlock, co-author of the book “Superforecasting: The Art and Science of Prediction,” runs the tournament. He began organizing similar tournaments in the 1980s.

Along the way, he discovered something curious. There are people who are exceedingly good at predicting what's around the corner.

Tetlock says the key to being a "superforecaster” isn’t necessarily about knowing something with 90 percent certainty. It’s being able to create a hierarchy of probabilities, much like an expert player would in poker.

“In poker, you’re sampling from a well-defined universe,” Tetlock says. “The basic laws of probability you learned in school obviously apply. It’s not so obvious the laws of probability apply when you’re estimating the likelihood of ‘unique’ historical events. Superforecasters are very good at seeing through the uniqueness label and seeing events falling into larger comparison classes and basing their probability estimates on that.”

And you can get better at making those estimates through practice, Tetlock says.

“It’s not that superforecasters... have higher IQs. I don’t think the answer is that they know more about politics… I think in this tournament what really mattered was a belief among the best forecasters that it was possible to get better at this.”

So put down the crystal ball and start forecasting.

Philip Tetlock, Sci and Tech, the future, Forecasting

Previous Post

Farming in Your Apartment

Next Post

The People Powering AI Decisions

comments powered by Disqus