Psychology Business Big Ideas

Superforecasting: The Art and Science of Prediction (Summary)

by Philip E. Tetlock & Dan Gardner

For decades, the CIA believed political forecasting was an art reserved for elite experts with access to classified data. Then, a government-sponsored tournament pitted these experts against thousands of ordinary volunteers. The shocking result? A small group of these amateurs—a pharmacist, a filmmaker, a retired computer programmer—didn't just compete; they crushed the experts, sometimes forecasting with 30% greater accuracy. Their secret wasn't genius or secret data; it was a specific way of thinking.

Be a Fox, Not a Hedgehog

Expert forecasters often fall into two camps: 'Hedgehogs,' who view the world through a single, grand theory, and 'Foxes,' who draw from many different ideas and are comfortable with complexity and nuance. The Foxes consistently make better predictions.

During the Cold War, many 'hedgehog' experts were certain the Soviet Union would endure because their entire worldview (e.g., Marxism or anti-communism) depended on it. 'Foxes,' in contrast, looked at multiple, smaller data points—like declining life expectancy and rising black market activity—and were far better at predicting its eventual collapse.

Precision Beats Vague Assertions

Vague predictions like 'it's possible' or 'a serious risk' are useless because they can't be proven wrong. Superforecasters make specific, probabilistic forecasts (e.g., 'a 70% chance') that can be tracked, scored, and improved over time.

When President Kennedy's advisors debated the Bay of Pigs invasion, they offered confident but vague assurances of success. A superforecaster would have asked, 'What is the precise probability of success? 60%? 85%?' This forces a clearer assessment of risk and exposes hidden assumptions, which could have prevented the disaster.

Start with the Outside View

When faced with a new problem, our instinct is to focus on its unique details (the 'inside view'). Superforecasters do the opposite: they start with the 'outside view' by asking, 'How often do things like this happen in situations like this?' This provides a statistical baseline that anchors their forecast in reality.

If asked whether a specific restaurant will succeed, don't just analyze its unique menu and great location (inside view). First, look up the base rate: what percentage of new restaurants in this city fail within three years? Start your forecast there, and only then adjust based on the specific details.

Forecasts Are Hypotheses, Not Commitments

Poor forecasters make a prediction and defend it to protect their ego. Superforecasters treat their forecasts as temporary, constantly seeking new information—even if it's contradictory—and making small, frequent updates to their probabilities.

A superforecaster might initially assign a 60% chance of a trade deal passing. After a key politician makes a negative comment, they might immediately revise it down to 55%. After a positive economic report, they might nudge it up to 58%. These small, frequent adjustments, like a ship making tiny course corrections, lead to far greater accuracy over time.

Go deeper into these insights in the full book:
Buy on Amazon
Listen to the full audio book with an Audible Free Trial.
As an Amazon Associate, qualifying purchases help support this site.