Skip to content

January 27, 2013

Review: The Signal and the Noise: Why So Many Predictions Fail – But Some Don’t

by brainoids

The Signal and the Noise: Why So Many Predictions Fail - But Some Don't
The Signal and the Noise: Why So Many Predictions Fail – But Some Don’t by Nate Silver
My rating: 4 of 5 stars

As with some others on Goodreads, I found this book a little hard to rate, thinking it a “3.5” and opting for a 4 star rating from an “E for Effort” standpoint. Part of this is high expectations on my part based on affinity for Silver’s FiveThirtyEight election prediction work.

The book is well researched and covers a nicely diverse array of example topics, including but not limited to economics, betting, sports, weather, climate, earthquakes and terrorism. The diversity keeps the interest going. A challenge here is that few of the examples were unfamiliar to me; ironically as the book is ultimately about Bayesian inference, there may be a little bit of a Bayesian thing going on relative to those most likely to buy/read and those most likely to have prior exposure and be left wanting more. The same thinking might suggest that the book is targeted more towards readers attracted by Silver’s political forecasts than those with a wonkish or professional interest in prediction itself.

For the latter, Silver redeems by offering something hard to find in similar popular literature, a high level synthesis across both realms and disciplines in prediction. A contrast with Kahneman’s Thinking, Fast and Slow and Surowiecki’s The Wisdom of Crowds (both of which Silver draws from) helps illustrate: While these two books are by no means peers (Kahneman’s represents a lifetime of scholarship, Surowiecki’s is more management faddish), as books, both suffer a bit from “the curse of knowledge” – the authors’ over familiarity with the often contradictory details leaves the reader rudderless on how to apply the findings in practice.

Silver, instead, takes a first step towards synthesis. This is welcome, although occasionally questions do arise about the formal correctness of mixing and matching themes and findings from very the different predictive methods (regression, classification, physics based modeling, simulation, etc) covered in the book. Absent a unifying framework to relate these methods (Silver is clearly an applied forecaster rather than a theoretician) the reader must rely on his claims to authority by experience (as well as the depth of research indicated by heavy citation) in trusting the synthesis and recommendations.

Overall, Silver ends up on the positive side of the trust ledger sheet, and even for readers already familiar with the topical examples, he provides enough additional color, as well as thought provoking commentary, to make it all worthwhile.

View all my reviews

Advertisements
Read more from Uncategorized

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Note: HTML is allowed. Your email address will never be published.

Subscribe to comments

%d bloggers like this: