The Signal and the Noise: Why Most Predictions Fail but Some Don't, by Nate Silver
/
I like this book, not because it provides much new insight, but because the author seems genuine. In this day and age someone who cares about their issue, more than selling books, is refreshing.
Two passionate, and well argued, points are provided in the book.
- There is data that clearly demonstrate that mainstream media is a destructive force when it comes to understanding the future as experts that are visible in media are those that with a bad track record of predicting the future. The reason for this is that the logic of media (sensationalism and simplicity) attract "hedgehogs" i.e. people with strong options and simplistic models for understanding the world.In a complex world we need to move beyond bumper stick logic this is a problem. What disturbs Nate, and should bother anyone who cares about a society built on knowledge, is that these sound bite providers manage to be constantly wrong, without ever being challenged. The kind of blog that Nate created for elections have a better track record, but it does not follow the media logic (e.g. it ignores the small things everyday and don’t over emphasize when new information becomes available that challenges established knowledge, instead it is weighted and incorporated).
- That society must become much better in dealing with probabilities, especially low-probability high-impact situations. The problem with the book as that it does not address totally new challenges where there is no earlier data set to base any estimations on. Still understanding that most of the big problems are “outside” what exports traditionally include in their planning.
I also want to give a big plus for proper footnotes. I'm so happy to see an author that actually provide footnotes, but it also makes me frustrated as you can't click on the links (or even copy the links) on Kindle... A simple software update from Jeff Bezos would be more than welcome (and a contribution to global knowledge).
There are also a number of "not so good things" in the book. The fact that Nate is enthusiastic result in a lot of "noise" in the book. Maybe Nate had to compromise with the publisher to get the book out? There is a lot of small stories and issues that are not related to the book, but maybe Nate thought that his book in itself could help train the reader to search for the signal among a lot of noise. Large parts reads like a (reasonable) entertaining, but rambling, dinner conversation.
It would have been interesting to read about Nate’s own reflections regarding his own role and contribution to noise/signal. After his “success” in the election he now seems to be invited to talk about a lot of things that he has no/little clue about. How can he ensure that he will not turn into a kind of “hedgehog” that applies his thinking to subjects where other approaches would be better? He seems like a cleaver guy who reflects so I’m sure he’s been thinking about it.
For those not interested in sport you can skip 20-40% of the book that does not add very much to the story (but I guess it is through his interest in sports that Nate wants to explain how he became interested in numbers).
As the general message about being careful and think about the possibility of being wrong is so well written I think this could be recommended as a book about scientific methodology at universities. Not the least in the US as the understanding of science seems to be moving in the wrong directions. But the teachers need to explain that Nate is a bit confused when it comes to the difference between policy making and science. All important issues cover both fields and there is no clear distinction (the difference between “ought” and “it” is more a theoretical one than a real).