# The Signal and the Noise

## Why Most Predictions Fail-- but Some Don't

Book - 2012
Silver built an innovative system for predicting baseball performance, predicted the 2008 election within a hair's breadth, and became a national sensation as a blogger. Drawing on his own groundbreaking work, Silver examines the world of prediction.

Publisher:
New York, N.Y. : Penguin Press, c2012.

ISBN:
9781594204111

Characteristics:
534 pages :,illustrations ;,25 cm.

Alternative Title:
Signal & the noise

## Comment

Add a CommentThe diversity of topics treated by applied statistics by the author in this book is quite broad: earthquakes, baseball, congressional and presidential politics, stock markets, and even terrorism. Its not overly complex, rather informally written, but clear. The insights gained are explained, as are the limitations of statistical decision or prediction making. Also well covered is the problems of models (overfitting of the data), and difficulties in getting enough data, and good data. It even earned a special place in my heart for its commitment to Bayesianism. (I should add that Richard Jeffrey's little book on Bayesianism "Subjective Probability: the real thing" is available as a free pdf e-book at fitelson dot org.)

A very good discussion of statistics and probability. This is a fairly technical book, which is a bit dense in part even though it is written for a general audience. Silver talks about the math behind predictions from politics to finance to weather to natural disasters, and explains why some things are really hard to predict accurately.

An interesting read if you're looking for some basics about statistics and probability.

Silver does a good job of maintaining a narrative throughout the book. He ties it together nicely with a sort of call to action to think more critically about predictions as you move forward.

Interesting book but I tired of it quickly. I could skim chapters once I understood the perspective of each subject covered in the chapter.

The Signal and the Noise is the most practically applicable statistics book I am aware of that also manages to be accessible.

Throughout the book, Silver explores various uses and misuses of statistical analysis, illustrating what makes the difference between a good prediction model and a bad one, by looking at some of the most successful statistical analysts out there, from his own experience with baseball statistics and elections analysis, to climate change science and weather forecasting, beating the stock market (or not), and poker strategies.

If nothing else, the book will convince you that these things have little-to-nothing to do with luck.

I really liked this book, but I can't give it a five stars. Here's what I feel:

- What I liked:

The topic of predictions is right up my alley professionally and personally. A lot of the topics he covers about limitations of predictions as well as how to improve them are something I live day to day.

Nate Silver's interests seem to closely reflect mine personally. I'm not a big baseball fan, but I'm a soccer fan and I can totally abstract a lot of his baseball commentary to soccer issues.

I love the connections and references he makes. He refers to other books that I've read (and recently), and he also goes back in time to Kasparov vs. Deep Thought (and Deep Blue) which I followed very closely when those were current affairs.

There are a lot of topics that he explains in reasonably simplistic terms which improved my understanding of the topics.

This is the best explanation of the poker bubble that I've ever read or heard.

-What I didn't like

The book is too long - at least 20-25% too long in my opinion. There are some points he just goes on repeating and belaboring.

The books is mostly organized as a series of "vignettes on prediction" the way I see it. The overall organization was a little confusing to me, not sure if I was missing the takeaways from sections or if I was supposed to draw some own conclusions from the different examples. The attempts to cross-reference prior examples only confused me more. I understand this is not intended to be a text book on predictions, but I had trouble crafting an overarching storyline.

The characters in the different anecdotes mostly flit in and out and in a lot of cases we don't hear enough to care about them. Even in non-fiction, I like it more if the characters in the anecdotes are more developed that we see things from their point of view (whether we agree or disagree)

Outstanding look at how we make predictions and how we can improve our ability to see into the future using big data

It's election time! Win your office election pool (that's a thing, right?) with some forecasting help from Nate Silver.

A look at statistical analysis and randomness with our inability to make predictions due to the complexity of dynamic systems. The narrative and layout is a bit disjointed but is otherwise an insightful read. Easily confirms my personal bias that most pundits have little idea of what they are talking about. Worth reading if you are interested.

This is a worthwhile book for those who would like a basis for skepticism about the information we get from news programs, although it could have been more concise. The most valuable chapter deals with the Bayes approach to making and updating predictions. If you can multiply, divide, add and subtract, you can use this formula as he directs. There is an error in the graphs on page 357 in that the grey areas represent "individual investors".