Superforecasting The Art and Science of Prediction by Philip E. Tetlock and Dan Gardner - Paperback

I’m not being funny or anything but I bought this book because I thought it would help me sort my fantasy league team out. I’m in second to last place, 3 points ahead of Jimmy the Turk and I’ll level with you I’m worried. Anyway Dom Cummings suggested I buy this book and since he was spot on about the 300m per week from the NHS I though I would go for it. Alas, the last two weeks before lockdown I lost both of my fantasy fixtures despite super forecasting hard all night. I even emailed Dom Cummings to ask for my money back but I superforecast that is unlikely to happen.

This book is a great book for understanding forecasting. It explains the methods and personalities of 'Superforecasters', people who Tetlock have studied who consistently outperform experts and non-experts alike in forecasting future events. Not perfect predictions, mind you, but consistently better statistically. What do they do differently than ordinary people to perform so well? Tetlock gives his best explanation in this book.One criticism I have is that I would've liked it to better slightly less "popular" science; include a bit more hard data, remove a little of the padding. However even with this criticism, there was much for me to learn. And it did include substantial references to evidence.Prediction is an extremely important component to testing whether your hypotheses are correct. Therefore, knowing about prediction is a key issue in science. Anyone who cares a lot about science should read a book like this or something similar. For any such person, I would gladly recommend this book.

The book clearly presents the results of a year long and very thorough work of the authors, dealing with how to improve our forecasting skills. It has given me a different perspective on the validity of many forecasts we see in the media and made me reconsider, how I am approaching forecasting in life at at work. However, after reading 50% of it, I felt like the concepts become repeatable, and the author is just trying to prolong the story by dwelling more on the same several key ideas. In any case, I would read it again, as it gives a different perspective on an interesting and important topic.

Super-forecasting is a book about forecasting, of course.If you didn’t know already, superforecasting has become a way for already people to make predictions about the world. The website sets challenges for you to forecast and then the tribe of forecasters set about researching and maybe even guessing, but they come up with a suggested outcome and forecast.In the book, Tetlock and Gardner, take you through the process that you too can start forecasting, actually with little or no data. For example, how many piano tuners are there in Chicago? OK, a simple question but a starter before working on a question such as “What will be the price of oil, when President Putin leaves office?”I was recommend this book and was also recommend the book “thinking in bets” by Annie Duke.

"Experts are about as accurate as chimps when predicting the future". This tidbit, so often mentioned when discussing (or dismissing) expert opinion or predictions, originates from the research of Mr. Tetlock on Expert Political Judgement. A natural next step was figuring out if anyone could reliably answer questions about the not so distant future and the result was the Good Judgment Project. The main results are detailed in this book, there indeed exists a group of super-forecasters who manage to constantly out-predict the chimps and experts in the intelligence community. The book describes some of the characteristics of a super-forecaster. Not surprisingly they are, in general, good with numbers and ingest a lot of information. They tend to be slow thinkers, in the sense of Kahneman, and at least in some cases not as much affected by cognitive biases. Super-forecasters, however, are not super-human. Forecasting is a skill that can be learnt or improved.