Just finished Nate Silver’s book The Signal and the Noise, only a few years after it was published.
I enjoyed most of the book, especially the philosophical discussion behind it: observations of the world inherently involve uncertainty, so predictions based on those observations must incorporate those uncertainties. But as more data pour in, we should be humble enough to revise our models and predictions.
As a whole, the book was easy to follow and didn’t get too technical, giving simple, worked examples of how to apply Bayes’s Theorem, for example, to evaluate the probability that a positive result on a mammogram actually means a woman has breast cancer.
For my money, the chapters on baseball and chess were the best. I learned that in the 1997 rematch between Garry Kasparov and Deep Blue, Deep Blue pushed a pawn on the 44th move, apparently baffling Kasparov. Silver suggests Kasparov “concluded that the counterintuitive play must be a sign of superior intelligence” and lost his nerve, resigning the next game even though it could have been played to a draw. Turns out the move may have been a last-ditch, random choice by Deep Blue.
The book falls a bit short in other chapters, though. Prof. Michael Mann, famous climate scientist at Penn State, had some pretty strong criticism for the climate change chapter. I found the concluding chapter a little platitudinous, without a lot of the interesting technical discussion in previous chapters.
Still, a really engaging read and, like me, hopefully readers come away with a strengthened commitment to scientific skepticism.