This book is about prediction and how various people do it, and how people ought to be doing it. It is well set out, with sections on all the stuff we especially like to predict; sports, politics, stock markets and the weather.
I have recently read The Drunkard’s Walk and The Black Swan and various popular economics and psychology books. I think I may have overdosed on this sort of thing, because this just felt like it was nothing new. Well, it’s either that or my previous ten year career in commodity trading analysis and electricity demand forecasting. I like forecasting, I think it’s interesting, but I judge a forecast intuitively now, I cannot help but consider the data available when someone makes a predictive statement, it has been three years after leaving that previous career and I still do it, so maybe I always will.
This is a nice introduction to that way of thinking, though, and I think all “normal” non-analytic folk should read this or something like it. All too often people accept a prediction without question (eg, the housing market has been recovering, so that will continue) and that can so often lead to disaster.
There is a particularly interesting bit on the behaviour of people who live in earthquake zones. Just as my experience of living in Las Vegas leads me to a general assumption that rain will fall over my house 15% of the time it is predicted people in earthquake zones rarely leave town when one has been predicted. Silver explains why.
Silver outlines the skills of a good forecaster. Get lots of data, then some more, and then some more, if you can. Be aware of your prejudices and question them. Re-forecast frequently. Do not force a pattern to appear if there isn’t one. Calculate and consider probabilities. Hedge your bets. Use a human brain to sense-check your models’ outputs. Do not trust the weather forecasters. Accept that the world is intrinsically uncertain.