Nate Silver

2012

---

An essential book for anyone concerned with "big data" or any aspect of science of forecasting. Silver casts an experienced (and somewhat jaundiced) eye over a range of commonly-encountered forecasts, including politics (his own main area), poker, finance, and climate change. In each area he manages both to convince that forecasts can be made to good effect -- and to demolish many of the current practices one finds in these areas. On the way he discusses Bayesian statistics, the psychology of a good forecaster (be a "fox," not a "hedgehog"), how to spot bias, and gives some critical advice that would be of useful to anyone looking to apply such techniques. Should be required reading for all science PhD students.

5/5. Finished 02 December 2013.

(Originally published on Goodreads.)