- Author: Nassim Nicholas Taleb
- Amazon Link
I normally start a review with the reason for selecting this book. In this case it was rather simple: I had lent the book "Born to Run" to one of the engineers in my group. Seeing that I liked to read books, he gave me this book to read as he had greatly enjoyed it himself.
At the outset, I was prepared for the book to be a not-so-easy-read and hence had saved it for a long airline flight, where it would get my undivided attention. The opportunity presented itself fairly soon and in August I made a trip to India and Israel and this was the perfect book to take along for the ride.
The premise of the book is about the high impact of improbable events. There are numerous examples cited in the book, ranging from the market crash of '89 to 9/11 as well as some added material on the financial meltdown in 2008. Taleb is very critical of the "quants" in the financial sector and the models that they base their analysis on. He draws similarity with the turkey who is living peacefully for 1000 days with its confidence mounting with every passing day, until day 1001 when it gets slaughtered. His point here is that past experience can be a poor basis to build a model for predicting the future. Improbable events can and do occur and most importantly have devastating consequences.
The above observation may be significant, but certainly not the stuff you can fill 400 pages with. Taleb rambles on with what seems like no main purpose and the book comes across as the pseudo-intellectual musings of an idle mind. He spends a good deal of time and effort in bashing Gaussian distribution and the bell-curve and expresses his fascination with Mandelbrot and fractals. I don't know if he is aware of the solid mathematical foundation on measure theory that underpins the probablity and statistics that we use today. Clearly assuming that random phenomena are occuring according to Gaussian (or any other for that matter) distribution is prone to disaster. However, suggesting that power-laws apply is also a best-fit extrapolation that is unlikely to apply when you are looking to predict improbably events.
All in all, I would not recommend this book unless you want to appear "intellectual" at your next social event.
