While it's hard to predict the future, some people are exceptionally good at it. The so-called "super-forecasters" can predict the likelihood of future events with astonishing accuracy, often with no particular prior expertise. Historically, evidence suggests that they are mostly generalists who dabble in all sorts of fields. They're simply less beholden to their own biases. The same applies when reversed, people who have built up an impressive but narrowly-focused expertise are usually less-accurate with their future predictions because they're limited by their own worldviews more.
Surprisingly, this is also true with people who have dedicated their lives to one field of study when they're trying to guess where that field is going. And we have data on that, emerging after a 20-year experiment that began in 1984 at a meeting of a National Research Council committee on American-Soviet relations. At the time, the psychologist and political scientist Philip E. Tetlock was 30. After listening to other members discuss Soviet intentions and American policies, he was interested in the authoritative predictions delivered by renowned experts because many of them contradicted one another.
Soon after, he decided to put expert political and economic predictions to the test. Tetlock collected forecasts from 284 well-educated experts who averaged more than 12 years of experience in their fields. Ensuring that the predictions were concrete, experts had to give specific probabilities of future events. Tetlock wanted to collect enough predictions that he could separate lucky and unlucky streaks from true skill. The project lasted 20 years and comprised of 82,361 probability estimates about the future.
The result: the experts were, by and large, terrible at predicting the future. Surprisingly, their areas of specialty, years of experience, and (for some) access to classified information made no difference either. They were bad at both short-term forecasting and at long-term forecasting. They were simply bad at this.
#9
#11
#15
#16
#17
#19
#20
