Book Review - The Signal and The Noise by Nate Silver
This is basically a review in two parts. One of them covers the rest of the book and the other covers the last chapter. You'll see why I had to do this later.
I've always liked stats (and it rescued me from absolute failure at A-Level maths), but that also means I'm well aware that it can be used to confuse rather than illuminate. The book appealed to me as Nate Silver seems to be the stats wonk du jour and it always nice to see stats at work.
Reading it also caused me great amusement when the Guardian started up their series of pieces on football statistical journalism. Because it wasn't quite lifted directly from but both the order and the shape of the ideas were blatantly straight lifts from The Signal and the Noise.
The first chapter is a little uneven, but it picks up after that. I particularly liked chapter 3 (which is basically about baseball, moneyball and how statistics will only take you so far), chapter 4 (which is about what it means to be accurate and what unexpected biases can creep in and communicating your results), chapter 6 (which is about what happens when that communication goes wrong), chapter 7 (about models and modelling, which obviously gets mentioned at different points but this one goes into depth about it), chapter 8 (about Bayesian statistics) and chapter 11 (which is about how forecasting can go right and wrong when you group forecasts).
Silver seems to write more clearly when the chapter is about sports (where he knows a lot) or things where he's not necessarily an expert in the field but understands how the number are produced. When he moves into politics, he seems a lot more smug and didactic.
The book is exceedingly well footnoted (I suspect there has been a thesis at some point in Silver's life, I've never seen this level of footnoting and citing outside of people who have been forced into writing one). There was at least one occasion where I could have done with an extra footnote - in chapter 5 which is about earthquake prediction and the danger of over-fitting data), where it is mentioned that the "Richter" scale we use now is actually a modification of the old one but at no point is the new one, or rather the equation used, given.
At the start of the book, the figures are clearly referenced in text but this falls by the wayside later on, which is a little annoying because visualisation helps in the understanding of statistics. I'm also not fond of figure 1.1 because it uses area to describe something and it's really hard to compare the two things if your brain is not wired for spacial awareness.
Overall, if the book had stopped at the penultimate chapter, I would have been happy to give it a 4/5.
Unfortunately, it doesn't.
The final chapter is on predicting terrorist attacks, and how the expectation of an event of size x can be quantified in a very similar way to earthquakes. So far, fine.
And, in a short description of Donald Rumsfeld, it explains Bush 2 era foreign policy better than 1000 words could have done ("He [Donald Rumsfeld] was eight years old on December 7, 1941, and was listening to his beloved Chicago Bears on the radio when the broadcast was interupted by a bulletin with news of the surprise attack", page 415).
The book also explains why Rumsfeld's 'Known Unknowns' soundbite was, in fact, perfectly reasonable rather than the nonsense it was accused of being. Again, perfectly good, because, the known unknowns bit is one of the best descriptions of why plans and predictions go wrong that you're going to hear.
It's after that it goes wrong. Part of the problem is that it focuses on the American experience of terrorism (which is fine, it's an American book), but uses NATO data as its data. And the experience of terrorism in other NATO countries is not the same as that of the US.
So when the book asks "Why don't terrorists blow up shopping malls?" (pg 440), one is tempted to ask, yes, why don't they? (The two examples I have used were within 30 minutes drive of my home town, I suspect Northern Irish people will have many more, and I suspect ETA also did similar things.)
Eyeroll-inducing parochialism I can live with.
But when, on the topic of the Israeli response to terrorism Silver comes out with "For instance, police typically try to clear the scene of an attack within four hours of a bomb going off, letting everyone get back to work, errands or even leisure,"(pg 441) one is tempted to ask whether that everyone includes Palestinians or is the book just ignoring them?
And then my response is to go 'Nate, you are too clever to be this stupid'. Which is a bad response to a book by someone I'm sure is leagues cleverer than I am.
The last chapter is so bad it drags my overall rating down to 3.
The Theory that Would Not Die: How Bayes' Rule Cracked the Enigma Code, Hunted Down Russian Submarines, and Emerged Triumphant from Two Centuries of Controversy by Sharon Bertsch McGrayne
Antifragile: Things That Gain from Disorder by Nassim Nicholas Taleb
Thinking, Fast and Slow by Daniel Kahneman
The Black Swan: The Impact of the Highly Improbable by Nassim Nicholas Taleb
Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die by Eric Siegel
The Information: A History, a Theory, a Flood by James Gleick
Introduction to Bayesian Statistics by William M. Bolstad
The Big Short: Inside the Doomsday Machine by Michael Lewis
Big Data: A Revolution That Will Transform How We Live, Work, and Think by Viktor Mayer-Schonberger
Introduction to Probability and Statistics from a Bayesian Viewpoint, Part 1: Probability (Pt. 1) by D. V. Lindley
Black Swan is on my to be read eventually radar.
I've always liked stats (and it rescued me from absolute failure at A-Level maths), but that also means I'm well aware that it can be used to confuse rather than illuminate. The book appealed to me as Nate Silver seems to be the stats wonk du jour and it always nice to see stats at work.
Reading it also caused me great amusement when the Guardian started up their series of pieces on football statistical journalism. Because it wasn't quite lifted directly from but both the order and the shape of the ideas were blatantly straight lifts from The Signal and the Noise.
The first chapter is a little uneven, but it picks up after that. I particularly liked chapter 3 (which is basically about baseball, moneyball and how statistics will only take you so far), chapter 4 (which is about what it means to be accurate and what unexpected biases can creep in and communicating your results), chapter 6 (which is about what happens when that communication goes wrong), chapter 7 (about models and modelling, which obviously gets mentioned at different points but this one goes into depth about it), chapter 8 (about Bayesian statistics) and chapter 11 (which is about how forecasting can go right and wrong when you group forecasts).
Silver seems to write more clearly when the chapter is about sports (where he knows a lot) or things where he's not necessarily an expert in the field but understands how the number are produced. When he moves into politics, he seems a lot more smug and didactic.
The book is exceedingly well footnoted (I suspect there has been a thesis at some point in Silver's life, I've never seen this level of footnoting and citing outside of people who have been forced into writing one). There was at least one occasion where I could have done with an extra footnote - in chapter 5 which is about earthquake prediction and the danger of over-fitting data), where it is mentioned that the "Richter" scale we use now is actually a modification of the old one but at no point is the new one, or rather the equation used, given.
At the start of the book, the figures are clearly referenced in text but this falls by the wayside later on, which is a little annoying because visualisation helps in the understanding of statistics. I'm also not fond of figure 1.1 because it uses area to describe something and it's really hard to compare the two things if your brain is not wired for spacial awareness.
Overall, if the book had stopped at the penultimate chapter, I would have been happy to give it a 4/5.
Unfortunately, it doesn't.
The final chapter is on predicting terrorist attacks, and how the expectation of an event of size x can be quantified in a very similar way to earthquakes. So far, fine.
And, in a short description of Donald Rumsfeld, it explains Bush 2 era foreign policy better than 1000 words could have done ("He [Donald Rumsfeld] was eight years old on December 7, 1941, and was listening to his beloved Chicago Bears on the radio when the broadcast was interupted by a bulletin with news of the surprise attack", page 415).
The book also explains why Rumsfeld's 'Known Unknowns' soundbite was, in fact, perfectly reasonable rather than the nonsense it was accused of being. Again, perfectly good, because, the known unknowns bit is one of the best descriptions of why plans and predictions go wrong that you're going to hear.
It's after that it goes wrong. Part of the problem is that it focuses on the American experience of terrorism (which is fine, it's an American book), but uses NATO data as its data. And the experience of terrorism in other NATO countries is not the same as that of the US.
So when the book asks "Why don't terrorists blow up shopping malls?" (pg 440), one is tempted to ask, yes, why don't they? (The two examples I have used were within 30 minutes drive of my home town, I suspect Northern Irish people will have many more, and I suspect ETA also did similar things.)
Eyeroll-inducing parochialism I can live with.
But when, on the topic of the Israeli response to terrorism Silver comes out with "For instance, police typically try to clear the scene of an attack within four hours of a bomb going off, letting everyone get back to work, errands or even leisure,"(pg 441) one is tempted to ask whether that everyone includes Palestinians or is the book just ignoring them?
And then my response is to go 'Nate, you are too clever to be this stupid'. Which is a bad response to a book by someone I'm sure is leagues cleverer than I am.
The last chapter is so bad it drags my overall rating down to 3.
The Theory that Would Not Die: How Bayes' Rule Cracked the Enigma Code, Hunted Down Russian Submarines, and Emerged Triumphant from Two Centuries of Controversy by Sharon Bertsch McGrayne
Antifragile: Things That Gain from Disorder by Nassim Nicholas Taleb
Thinking, Fast and Slow by Daniel Kahneman
The Black Swan: The Impact of the Highly Improbable by Nassim Nicholas Taleb
Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die by Eric Siegel
The Information: A History, a Theory, a Flood by James Gleick
Introduction to Bayesian Statistics by William M. Bolstad
The Big Short: Inside the Doomsday Machine by Michael Lewis
Big Data: A Revolution That Will Transform How We Live, Work, and Think by Viktor Mayer-Schonberger
Introduction to Probability and Statistics from a Bayesian Viewpoint, Part 1: Probability (Pt. 1) by D. V. Lindley
Black Swan is on my to be read eventually radar.