Friday, January 29, 2010
The Black Swan - Nassim Nicholas Taleb
I've been digesting this book for a day or two, and I still don't really know what to say about it. Taleb is very intelligent and can write about difficult philosophical and epistemological subjects with ease, and he usually has the courtesy to bring you along with him. You know he understands what he's trying to say, but sometimes it seems like he's trying to say contradictory things. The book is a discussion of how people treat probability, risk, statistics, and prediction incorrectly. He marshals his experiential anecdotes, fictional stories, theories, philosophical ideas, and statistical breakdowns in the main argument that we need to pay more attention to Black Swans. These are events or quirks in existence that are nearly impossible to predict, very important and impactful, and rare. His argument is that usual statistical models (the bell curve, Gaussian statistics) aren't useful because they don't take into account these rare and important events. This causes people to build their lives around safe assumptions of what will happen, ensuring they remain unprepared for Black Swans, which can be positive or negative.
I think it's an interesting argument, and he managed in 2007 to talk about the dangers of Wall Street ignoring the warning signs of a large collapse in the housing market and hedge fund/investment bank over-leveraging. That's an important insight, and I'd like people like him weighing in on how our financial sector should be organized.
My troubles with the book mainly reside in the organization - he meanders a great deal. At times he seems like he wrote it to settle scores and write wrongs by explaining himself to people he feels have slighted him (similar to Richard Dawkins' The God Delusion). I found myself noticing increasing instances of not-so-subtle digs at the French, Nobel Prize winners, most philosophers, economists, Wall Street traders, people who think they know anything, and anyone who's looked askance at Taleb's theories. Additionally, he veers between "we don't and can't know anything" and "I'm not saying we don't and can't know anything" as we speed through the book.
I got the point that narrow uses of statistical models are dangerous and panning out in the picture of reality is important. I got that a lot of "experts" are too focused and blinded by their dogma to be safely useful for dictating policy and research. I got that good ideas are often ignored, important events aren't predicted well, and history is often unexplainable. Skepticism is underutilized. But other than vowing to think critically-er, I'm not sure exactly what else I'm supposed to take from the book.
I'll still be ruminating about what I got from this book for some time, and I'm glad I read it. But I think it could have been a better book.