The Signal and the Noise: Why So Many Predictions Fail - but Some Don't

I love statistics and data analysis and Nate Silver is very smart and knows a lot about data analysis, yet Nate Silver seems to have an “unhealthy obsession” with data analysis. There are quite a number of interesting bits in The Signal and the Noise but it starts with a long and in depth discussion of baseball statistics, which (if you are apathetic to baseball) is only mildly interesting at best. The overall theme of this book was to show the value of Bayesian statistics. I am a big fan of Bayesian stats, but the descriptions here seem weak and the author sometimes seems to delve into data analysis just for sake of data analysis. It felt a bit like being in the room with an obsessive compulsive geek with a calculator. You may learn a lot, but it might not be a lot of fun.

Nate Silver is hot right now. As I write this, it is three days before the presidential election and he is predicting an Obama win (82% chance of winning). His insights about stats, opinions, signal and noise are spot on. Although I am still not 100% sure what Bayesian logic is. Overall a great listen full of insight. A note on the narrator. I take back every negative comment I've ever made in my reviews of his performances. He was excellent in this context.

Sensible, thought-provoking, data-based

The examples were very relatable and understandable.

The baseball example was the most illustrative.

Seek the truth...

This book's release coincided roughly with the author's correctly calling the Obama reelection, and I know there was some promotional buzz. I enjoyed the book, but like so many others, it is laden with all sorts of random anecdotes and stories that don't really cohere well, so it serves as a very light repast of actionable, penetrating statistical methods, slathered over with plenty of entertainment and filler. So, it probably nicely fulfills the contemporary book industry formula for popular writing and its marketing.

It was clear and generally well presented (accessible to a wide audience). I was well aware that people generally are not good at making predictions, relying too heavily on their heuristics and biases (e.g. failure to use prior probabilities as called for by Bayesian statistics); and the importance of putting "band widths" around probability estimates. I also thought that much too much text was devoted to each of the major topics covered (e.g. weather forecasting, political forecasts, economic forecasts, picking stocks, gambling/poker strategies, etc.). These sections could have been considerably shorter.

"Is that all there is?"

N/A

No

Generally disappointing, in that it did not expand my knowledge -- though I don't fault the book, as I've read widely in this area.

I would like to listen to parts of it again. There are some sections which had far too much detail for me, but at the same time, that is likely a plus for others.

I enjoy the parts about how he comes up with his predictions - which I thought was a bit light in the book. I would have liked more insight on the political polling - but maybe that will be the content for his next book - which I will also buy.

It is a tough read and a long book - but it is worth it. If you like predictive math and logic - and want some insight into the future of market research - it is quite valuable.

Reviewing the mortgage crisis and the responsible participants was very interesting, and the fact that a hybrid approach to probabilities (number analysis and expert opinion combined) was superior to the pure Moneyball approach was certainly revealing and practical.I also now appreciate much more the work and efforts of the national weather service, what they're able to accomplish, and why their mission should be important to all of us.

Heck, I thought it was Nate Silver the whole time. I'm sorry, Mike! I guess that's how authentic it seemed.

I really enjoyed the content and insight it provided, so I burned through it at 3x speed.

Really enjoyed it!

If your looking for how Nate did some great research and did some very good successes I recommend the book. If you like a story with drama, mystery and suspense it is a little dry and not for you.

Nate Silver is really knowledgeable, and he covered an array of topics that he knows a lot about.

Silver gives an interesting perspective on many topics on which he is knowledgeable. But he does not do nearly so good a job of explaining "why so many predictions fail but some don't." My summary of his explanation of the subtitle is "There is not enough information (or signal) to make good forecasts in the areas where forecasting is bad." This provides little insight if one is considering forecasts on a topic not covered in his book. I was disappointed because I was interested in learning more about forecasting in general.

Silver loves focusing on the rich details that relate to forecasting in politics, sports, economics, and more. I believe that most of the readers who follow Nate Silver will enjoy this book.

One of the best audiobooks I've listened to so far

Nate Silver gives an advanced, yet comprehensible lesson in statistics using exciting real world examples of how statistics were used correctly or incorrectly in each case. Topics range from earth quakes to political elections, which he is most recenty famous for.