The Black Swan by N. Taleb/Book Review: J. Colannino

A few weeks ago I heard Vinod Kholsa speak at the ARPA-E conference in Washington D.C. He recommended this book, written by Nassim Nicholas Taleb (NNT), so I read it. NNT is in some ways what I aspire to be -- a renaissance man. Three years my junior and born in Amioun, Lebanon he speaks fluent French, Arabic, and English. He has a reading knowledge of classical Greek, Latin, Aramaic, and Hebrew. NNT received his MBA from the Wharton School of Business at the U of Pennsylvania and a Ph.D. from the U of Paris, Dauphine, in Management Science. He reportedly made a personal fortune predicting the collapse of the financial markets and Fannie Mae. In fact, that prediction is in this book published in 2007, a year before the crisis. The Black Swan is Taleb's magnum opus -- at least so far.

Swans are white, by definition, until the first sighting of a black swan in Australia in 1697. The view was game changing and the definition of what a swan was had to be revised. By analogy, NNT defines black swans as rare, unanticipated, and game changing events; things like the laser, the Internet, and the personal computer.

Clearly an iconoclast, and I dare say angry man, NNT rails at the current state of academia and business, especially the statistical sciences as applied to economics and finance. He is angry because mistakes in finance and war destroy lives. One thing that NNT cannot brook is "epistemic arrogance" -- gasbags that bloviate about things that matter but that they only pretend or delude themselves into thinking that they understand. According to NNT, people become arrogant because arrogance grows faster than knowledge. Experts are the worst (most arrogant) and thus make suckers out of their clients/victims. If this language sounds extreme, you should read the book. I am toning it down. Taleb has to self-censor his own expletives with asterisks.

The Black Swan compares two hypothetical countries -- Mediocristan and Extremistan. In Mediocristan, all of the usual rules apply, including the normal distribution and derived statistics. In Extremistan, the bell curve is the "great intellectual fraud (GIF)." The problem, according to Taleb, is that we live in Extremistan, not Mediocristan. For example, market fluctuations of 20 standard deviations have arisen historically -- this is simply impossible in a world governed by the normal distribution. As a remedy for statistics based on the normal distribution, Taleb recommends Mandelbrot's fractal geometry; this does not give the precision of classical statistics, but it does give the best one can hope for in Extremistan -- a gray swan. A gray swan is a partially anticipated black swan. Black swans are hatched by nature, not created by man. The best one can do is maximize opportunities and conditions for hatching black swans.

Black Swans (game changing innovations) are accidents. They only seem predictable because of a post-hoc story that is irresistibly constructed to make sense out of the surprising. For example, there is no end of books about World War I and its cause. But consider this: if anyone really knew what caused World War I, it would have been prevented. One of the many things that really irritates NNT is that the experts who destructively lull the general public into believing financial markets are stable are never held accountable for the harm they do. That is, being perennially wrong has no consequence to their career.

I can draw many parallels in my own R&D career. Almost anything beyond incremental progress is serendipity. I agree with Taleb that the best one can do is to maximize opportunities for favorable outcomes. In my own experience this happened by moving incremental R&D to other groups, and maximizing the amount of breakthrough R&D we performed internally. In other words, I deliberately set about to maximize the opportunities for R&D by having more bites at the apple. The result was breakthrough research.

Part (or all) of NNT's pessimistic outlook seems to me to be conditioned by his worldview: that we are a chance result of grander forces. Hence, matters that count are random, unpredictable, and non-sensible; we merely invent post-hoc narratives as comfort food. Taleb is a pragmatist, and whether this is absolutely true or apparently true makes no difference to him. In everyday living, the end result and the proper actions to take are the same in both cases. Ironically, he seems to base his philosophy on this evolutionary paradigm, itself the ultimate example of the post-hoc narrative fallacy that he decries. While I don't share his worldview, I do think that he rightly shows that extreme events do not operate in a business-as-usual environment and that there is much money to be made by properly recognizing this difference.