The Signal and the Noise, by Nate Silver
Feb. 28th, 2014 04:00 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
essentialsalteswhy so many predictions fail -- but some don't.
One of my favorite things is the cover, with all the additional 'noise'.
Each chapter takes on a specific realm of prediction, or a particular aspect. Weather, baseball, poker, the economy, Bayes' Theorem, etc. Probably the best part of the book is Silver's interviews with various experts who explain a lot about how they make predictions, and in particular the focus on the limitations of these predictions. As he tells it, the book is divided in two halves: "The first seven chapters diagnose the prediction problem [this is the good stuff] while the final six explore and apply Bayes's solution."
Silver's own voice is not as compelling as that of some of the experts, and though he talks about his work in baseball stats and his brief time as a professional poker player, I was surprised that he doesn't talk all that much about his political predictions and poll-mashing. I also have some quibbles with some of his examples and analogies. But still, especially in the first half of the book, there's a lot of good material about modeling, prediction, and probability, and how to identify and avoid the traps we inevitably fall into (even when we know how to identify and avoid the traps).
One of my favorite things is the cover, with all the additional 'noise'.
Each chapter takes on a specific realm of prediction, or a particular aspect. Weather, baseball, poker, the economy, Bayes' Theorem, etc. Probably the best part of the book is Silver's interviews with various experts who explain a lot about how they make predictions, and in particular the focus on the limitations of these predictions. As he tells it, the book is divided in two halves: "The first seven chapters diagnose the prediction problem [this is the good stuff] while the final six explore and apply Bayes's solution."
Silver's own voice is not as compelling as that of some of the experts, and though he talks about his work in baseball stats and his brief time as a professional poker player, I was surprised that he doesn't talk all that much about his political predictions and poll-mashing. I also have some quibbles with some of his examples and analogies. But still, especially in the first half of the book, there's a lot of good material about modeling, prediction, and probability, and how to identify and avoid the traps we inevitably fall into (even when we know how to identify and avoid the traps).
