Order The Black Swan: The Impact of the Highly Improbable

There are many good things about this book. Firstly, it is a very enjoyable read. The examples and stories may hurt the organization of the ideas somewhat but they make the book hard to put down. Secondly, Taleb makes some very reasonable criticisms of the use of the normal curve for variables that are clearly not normally distributed such as book sales or market prices. He exaggerates the problems with statistics quite a bit, however. I hope that non-scientific readers do not take these to heart and cease trusting anyone who uses traditional data analysis methods- in most scientific endeavors these are appropriate and superior to alternatives. Later in the book he does admit that there are legitimate uses of this approach- he gives the example of empirical psychology specifically(it was refreshing to hear his praise of this science).

His idea about using a fractal distribution for data with 'fat tails' is an interesting alternative, though I'm not sure how reliable such a technique really is. This may be part of the point- the normal curve and traditional statistics simply don't work in these cases, but we don't really know how these variables are distributed, so using a fractal tail is an alternative but imperfect tool. The major problem that I see with this tool is that you have to estimate what the power exponent of the tail will be- and making a mistake will have huge consequences for the model. Taleb himself admits that you could give ten different analysts the same data and they could all come up with a different value for the power exponent! This seems like a major Achilles Heel to using a fractal approach to these situations. I wish instead Taleb had simply emphasized that these variables don't behave in a way that lends themselves to exact modeling, especially because of the higher possibility of 'Black Swans' or what could simply be called very extreme cases.

My gripes with the book mostly come down to Taleb's arrogance and exaggerations. Taleb can be pretty full of himself, which gets old fast. As I mentioned above, I wish he hadn't exaggerated the problems with statistics- in the first part of the book he acts like nothing interesting calls for the use of statistics and calls the normal distribution an intellectual fraud. The normal distribution is not a fraud- it is an excellent description of many phenomena, despite the fact that there are situations to which it should not be applied. Taleb is also extremely critical of certain disciplines and professions, especially economics. He does make legitimate criticisms of economic ideas and models, but he does so in an insulting and tactless manner.

Despite Taleb's arrogance and exaggerations, this book has many interesting ideas and several great discussions of the philosophy of knowledge and science that make it worth reading. Just make sure to exercise a healthy dose of skepticism when you do so (as Taleb himself advocates).Get more detail about The Black Swan: The Impact of the Highly Improbable.