The BlackSwan: The Impact of the Highly Improbably

Black Swans are unexpected occurrences that disrupt predicted outcomes. Though seeing one-thousand white swans does not disprove the existence of black swans, a single black swan can prove their existence.

This book is filled with examples of the human tendency to think they know more than they do and fail to expect the unexpected. People tend to look at data and events that confirm their expectations rather than seeking for ones that disprove it. This is exacerbated when the losers drop out. (Thus gamblers seem to have "beginners' luck" because lucky beginners are most likely to continue.) While "rules" may provide insight, breaking the rules is required for success. In hindsight, we can often create "rules" to predict improbably events (like the rise of the internet or Microsoft.) However, rarely can we accurately predict them in the future. (And when we do, it tends to be due to dumb luck.)

The arguments are good. However, they seem to bounce around, often arguing opposite points from different angles. (I guess this is par for the course for a book on highly impactful improbabilities.) The author does tend to ramble on, making this a good target for a "condensed book." His tone also appears very condescending. While he stresses the importance of not getting too involved with minutia and our own knowledge, he comes across as perhaps too confident in his own knowledge. He tends to go too overboard in defense of randomness. While the random events may be the most influential, without some degree of predictability, there would be little chance for the serendipitous events to occur. (If we knew we would know something in the future, then we would already "know" the event happened.)

He spends some time criticizing the European/French elite intellectual establishment, then proceeds to quote various obscure intellectuals. (Perhaps he does not realize that he is more closely associated with the establishment than he thinks.)

The strongest attacks, however, are saved for economists and statisticians and their Gaussian normal distributions. While normal distributions may model many theoretical natural conditions, it has significant problems in the real world. He has much more respect for Mandelbrot's fractal analysis. Modern "portfolio theory" is heavily dependent on the Gaussian statistics and thus susceptible to uncertainties. His attacks are strong and convincing. However, he appears to go overboard. While he acknowledges that statistics can be used to managed the "closed" situations of Casinos, he doesn't acknowledge that financial markets may have some aspects of the controlled casinos. Just as a casino can hit unexpected events (such as a tiger mauling or a disgruntled former employee), the financial markets can have unexpected events occur. Does this justify throwing out the whole statistical world?

However, even this refutation comes with a caveat. Trying to assume a belief set as a package is a dangerous proposition. Why do philosophers doubt their mere existence, but trust their retirement money to portfolio theorists. Why do some doubt religion, but put faith in financial hocus pocus? In the end, he provides his own refutation. Everybody is their own black swan. They do wild unexpected things that make them who they are. Perhaps one of these crazy things is to trust economists.