The Signal and the Noise

"The Signal and the Noise: Why Most Predictions Fail, but Some Don't", by Nate Silver, is a study on how people make predictions in a wide variety of fields, and why they are wrong most of the time. Silver is the author of the 538 Blog, which gained a bunch of Fame over the last two election cycles for very accurately predicting congressional and state presidential elections. In this book, he explains how people in everything from baseball to earthquakes predict things in their fields and why they seem to do so poorly most of the time (it is normally for one of a few reasons - their models are wrong, they have an agenda besides correctness, or they don't know how to account for rare events). He does love weathermen though - on average, their prediction accuracy had increased significantly even over the last 20 years. Their predictions are good because they make a ton of them, it is easy to see if their predictions are right, and they improve their models when they fail.

Overall, I liked this book, but it felt like a collection of essays rather than a coherent whole. It took me a looooong time to get through it, but that is probably because it is so densely packed with interesting information. An interesting read.