Superforecasting: The Art and Science of Prediction

One assumption about superforecasters is that they have above-average intelligence, potentially even falling into the “genius” range. Tetlock and Gardner admit that the 2,800 people who volunteered for the GJP were not exactly a representative sample; these people largely came from academic settings, wanted to devote considerable time to puzzling out questions about Greek bonds and Nigerian politics, and were willing to endure the study requirements (including psychometric testing) for the modest reward of $250. The GJP found that forecasters in general outperformed 70% of the general public in intelligence and knowledge tests, while superforecasters did better than 80%. The authors draw attention to the fact that “the big jumps in intelligence and knowledge are from the public to the forecasters, not from forecasters to superforecasters” (109). While superforecasters have above-average intelligence, the vast majority fall short of being “geniuses.” The conclusion is that while intelligence is a boon, beyond a certain point, intelligence alone does not enhance forecasting ability.

Having addressed the question of intelligence, the authors next explore behaviors. Unlike the general population, superforecasters apply strategic methods to answering questions. For example, the Italian American physicist Enrico Fermi asked his students to estimate the number of piano tuners in Chicago; many people essentially give up when faced with this question, offering a speculative guess, whereas superforecasters immediately ask themselves what other information they need to answer the question.