Scott Alexander shares and comments on highlights from Philip Tetlock's 'Superforecasting', discussing forecasting, cognitive biases, and organizational effectiveness. Longer summary

This post is a collection of highlights and commentary on Philip Tetlock's book 'Superforecasting'. Scott Alexander shares quotes from the book and provides his own analysis on topics such as evidence-based medicine, cognitive biases in forecasting, the importance of probabilistic thinking, and organizational effectiveness. He also reflects on the implications of these ideas for fields like intelligence analysis, politics, and rationality. Shorter summary

Scott Alexander reviews 'Superforecasting' by Philip Tetlock, discussing the traits of highly accurate predictors and the book's validation of rationalist techniques. Longer summary

This post reviews Philip Tetlock's book 'Superforecasting', which explores the qualities of highly accurate predictors. Tetlock's Good Judgment Project identified a group of 'superforecasters' who consistently outperformed others, including CIA analysts. The review discusses the characteristics of these superforecasters, emphasizing their understanding of logic and probability, ability to break down problems, and resistance to cognitive biases. Scott Alexander notes the similarities between superforecasters' methods and rationalist techniques, suggesting the book's value lies in providing high-status validation for these approaches rather than presenting new information to those already familiar with rationality concepts. Shorter summary

Scott Alexander explores the possibility of a 'General Factor of Correctness' and its implications for rationality and decision-making across various fields. Longer summary

Scott Alexander discusses the concept of a 'General Factor of Correctness', inspired by Eliezer Yudkowsky's essay on the 'Correct Contrarian Cluster'. He explores whether people who are correct about one controversial topic are more likely to be correct about others, beyond what we'd expect from chance. The post delves into the challenges of identifying such a factor, including separating it from expert consensus agreement, IQ, or education level. Scott examines studies on calibration and prediction accuracy, noting intriguing correlations between calibration skills and certain beliefs. He concludes by emphasizing the importance of this concept to the rationalist project, suggesting that if such a 'correctness skill' exists, cultivating it could be valuable for improving decision-making across various domains. Shorter summary