Superforecasting

By Philip E. Tetlock

**Estimated Reading Time:** 45 minutes

**What You'll Learn**

Why some people can predict the future better than experts with classified information. How a specific set of thinking habits transforms ordinary people into remarkably accurate forecasters. What it takes to see the world in probabilities rather than certainties.

**Who This Book Is For**

Anyone who makes decisions under uncertainty. That includes investors, business leaders, policymakers, analysts, and anyone who has ever wondered whether their predictions about the future are any better than random guesses.

In the fall of 1984, a young psychologist named Philip Tetlock began asking experts to predict the future. He wanted to know something simple: when people who study politics, economics, and international affairs for a living say something will happen, how often are they right? The answer, when it finally arrived after two decades of data collection, was uncomfortable. Most experts performed no better than random chance. Some performed worse. A dart-throwing chimpanzee, Tetlock observed, would have beaten many of them. The experts were confident, articulate, and frequently wrong. Worse, they were rarely held accountable for their errors. The world moved on, new predictions were made, and nobody kept score. This finding, published in Tetlock's 2005 book *Expert Political Judgment*, was not exactly welcome news. It suggested that the entire enterprise of expert forecasting was built on sand. But buried in the data was something curious. A small minority of forecasters did better than chance. They did better than their peers. They did better, in some cases, than the collective wisdom of markets. Tetlock wanted to know what made them different. That question led to the Good Judgment Project, a multi-year forecasting tournament sponsored by the Intelligence Advanced Research Projects Activity, or IARPA, the research arm of the American intelligence community. The tournament was simple in design and brutal in execution. Thousands of volunteers were asked to predict real-world events: Would Greece leave the eurozone? Would North Korea test a nuclear weapon? Would a particular dictator survive the year? Every prediction was scored for accuracy. Every forecaster was ranked. The results were startling. A small group of forecasters, roughly two percent of the participants, consistently outperformed everyone else. They beat the average forecasters. They beat the experts. They beat prediction markets. Most remarkably, they beat intelligence analysts who had access to classified information. These individuals came to be called superforecasters. What made them special was not obvious. They were not geniuses. They were not clairvoyants. They were not insiders with secret knowledge. They were ordinary people: a retired computer programmer, a homemaker, a former ballroom dancer, a software engineer. What they shared was not talent but a way of thinking. They approached prediction…