Black Swan: Second Edition: The Impact of the Highly Improbable: With a new section: "On Robustness and Fragility", The
Review by Derryn:
Nassim Taleb has been praised for writing a no holds barred critique of quantitative risk, described as one of the most important books of the 21st century. Important due to his insightful description of risk models in the finance industry. TBS was published just months prior to those same models failing spectacularly during the GFC.
This was an enjoyable book. It provides a great narrative with a style uncommon to non-fiction. One reviewer described it as being ‘stuck in a cab with a highly opinionated cab driver’. That is exactly how it feels, a passionate and witty description against, well pretty much everyone. Mr Taleb jumps from topic to topic in a relatively smooth manner that keeps you interested.
The Black Swan synthesises a lot of ideas from different domains into his story and adds some of his own descriptions of concepts. His identification and naming of the ‘Ludic Fallacy’ is very insightful. This is the idea that the real world can’t be reduced to perfect mathematical formality. You are better off thinking about your risks logically and what you can do to contain them.
The book has a lot of problems. Firstly, there is no structure. Without structure he has just produced a long stream of consciousness rant against things he doesn’t like. Without structure there is no position that he can actually be criticized on. He doesn’t provide a comprehensive summary of his thesis, if there even is a thesis. This is a major problem. ‘The Black Swan’ has become an overused term with no clear definition. I’d bet that no matter how you take this book I’m certain Mr Taleb would tell you you’re wrong.
Mr Taleb also claims a lot of originality for his idea. But so far as I can tell it is just a long way to describe the concept of ‘unknown unknowns’.
The Black Swan focuses a lot on the use of the normal distribution. As such, a lot of effort is spent talking about ‘fat tails’ and how they don’t capture ‘rare events’. A branch of statistics called extreme value theory has been developed that captures most of the problems he identifies with the normal distribution. Admittedly, this theory can’t capture ‘unknown unknowns’ but he does ignore this valuable theory completely. Mr Taleb’s response to this critique was a 4 page paper that doesn’t make sense. Given he can write 300 pages about the normal distribution I’m sure he could do a better job than four pages on extreme value theory.
Mr Taleb offers some periods of humility, but by the end of the book his rant is in full swing. He is quite offensive in tone to a lot of people, economists and journalists are specially targeted by his ire. Where he isn’t offensive, he is a sycophant, e.g. with his praise of Benoit Mandelbrot.
The book does offer the reader some great lessons. I particularly liked his concept of how he tries to act under uncertainty, his ‘bar bell’ description. As a default act with extreme risk aversion, something like a Kelly Criterion, in most decisions. This makes you robust to uncertainty. But at the other end of the bar bell, when a positive black swan is likely, put all your eggs in that basket to make the most of that opportunity.
This book was very hard to review. There is so much in it that can be liked and a lot that can be ignored. I think it is worth the read nonetheless. Maybe you will find diamonds in the rough that I didn’t quite understand.