Nate Silver is the statistician who received a lot of attention for correctly predicting the winner of the 2012 presidential election in all fifty states. His blog FiveThirtyEight (named after the number of electoral votes) is a treasure trove of analysis and predictions for politics, sports, economics, and culture.
I expected The Signal and the Noise to describe, for a general non-mathematically inclined reader, the mechanics of statistical prediction and how to interpret statistical forecasts. The book does include some of that, but most of its chapters explore areas where we try to predict the future -- the stock market, elections, sports, the weather and climate, earthquakes, national security -- and why we mostly fail to make good predictions. We fail to distinguish the signal from the noise due to bias, overfitting small samples, or pure complexity.
The through-line of Silver's argument isn't always clear, but ultimately his more tangible approach is rewarding. Wouldn't I rather think about real-world applications than abstract theory? Well, I'm not sure actually, but the book always held my interest.
I expected The Signal and the Noise to describe, for a general non-mathematically inclined reader, the mechanics of statistical prediction and how to interpret statistical forecasts. The book does include some of that, but most of its chapters explore areas where we try to predict the future -- the stock market, elections, sports, the weather and climate, earthquakes, national security -- and why we mostly fail to make good predictions. We fail to distinguish the signal from the noise due to bias, overfitting small samples, or pure complexity.
The through-line of Silver's argument isn't always clear, but ultimately his more tangible approach is rewarding. Wouldn't I rather think about real-world applications than abstract theory? Well, I'm not sure actually, but the book always held my interest.
No comments:
Post a Comment