Very U.S. centric, but some interesting chapters nonetheless,
This review is from: The Signal and the Noise: The Art and Science of Prediction (Paperback)
Customer review from the Amazon Vine Programme (What's this?)
Nate Silver became a household name when his fivethirtyeight website correctly predicted all 50 states's results in the US Presidential election. I started this book with high hopes of a clear treatise on prediction. I very quickly got bogged down however, in the highly US-centric nature of this book. Starting with lots of detail about US political pollsters and pundits, we quickly move to a highly detailed chapter about baseball predictions. Although this is ok to understand for a non-baseball afficionado, the endless baseball terminology quickly reaches critical mass. The theory is ok - baseball is easy to make statistical models for because the batters just bat and the pitchers just picth - only the fielding is the team part. So it's not like football (or soccer as Nate would no doubt call it), because it's fairly pure. In fact it maps fairly well to cricket, but there is no mention of cricket in this book. And in fact there is barely any mention of anything outside the US in the whole book.
Having said that, this is a fairly readable book with some interesting chapters. It's not as clearly written as I hoped and the sections on weather forecasting (again mostly concerned with US Hurricanes) and poker (international I know and something Nate himself played professionally) are ok, but a bit woolly in their exposition.
Later chapters are much better - the one on chess playing computer programmes (and the effect of bugs therein) and another on climate change are well written, although the latter - like the one on stock market predictions - is steady rather than having astounding revelations.
As another review has pointed out the book is light on the maths itself, but we are constantly reminded of Bayes theorem and how we can use it for real world, complex problems to check whether we are on the right lines when we predict things.
I wasn't convinced by everything in this book, but it improved as it went on and if one can ignore the parochial nature of its US author it is certainly worth a read.