Superforecasting: The Art and Science of Prediction

5* and must read for people interested in forecasting and geopolitics, super useful for business leaders too. 4* for more casual readers, still interesting though with fewer day to day takeaways. (The best being that commentators on TV/newspapers generally have little actual insight, should be obvious but is not!)

This is the ongoing story of a research project that bore serious fruit. Showing that it is possible to both make useful, and also systematically improve, forecasts about the future - albeit with some limitations.

I knew the jist of the book from articles in the press, this is deeper on both the initial findings, the improvements during the tournament, and the future. Worth the extra effort.

Well written and nuanced, takes the lay reader from basic principles to expert insight without overuse of jargon and with good examples. Excellent name/story dropping with Tversky & Kahneman, and Taleb, well done sections.

Superforecasting skills are usually held by pragmatic people who work hard on themselves - especially in the acknowledgement and ability to put aside their own biases, who stay on top of news, and who have strong maths skills (though they don't need to use it much - more about the ability to think probabilistically).

It becomes clear that it's more how you think as opposed to what you know that helps. There is a system, to be more of a fox as opposed to a hedgehog, to be able to change your mind with new information, to actively seek out that new and opposing information. Working in the right type of team helps too.

The book is a little US based and they are openly working with the US intelligence services, that may not be a good thing depending on your POV - there are deep discussions on the Iraq war intelligence failure. Would have better assessment avoided war? We don't know. I can't see better predictions making it any worse however.