What are the chances of that?

All swans are white.” This scientific generalisation, well supported by wide observation, was unexpectedly overturned when the first travellers to Australia brought back reports of black swans.
In this book, a “black swan” is an event with similar attributes. “First, it is an outlier, because nothing in the past can convincingly point to its possibility. Second, it carries an extreme impact. Third, human nature makes us concoct explanations for its occurrence after the fact.” Examples include the outbreak of the First World War in 1914, the market crash of 1987, the terrorist attack of September 11, 2001, the Pacific tsunami of December 2004, the hurricane Katrina. Other examples from the realm of science are the discovery of the theory of relativity and, from medicine, the spread of Aids and of BSE. There are also “positive” black swans, like the extraordinary growth in the power of computers and the rapid spread of the internet in recent times.

Human beings are not programmed by evolution to understand extremely rare events. At first we think they will not happen. Then we ascribe too much importance to them: for example, traffic accidents and cancer are far more significant killers than hurricanes and terrorist attack; yet alleviation of the effects of widely reported disasters diverts money from far more worthy and more worthwhile causes. The trouble is that no one gets rewarded simply for preventing bad things happening.

This book gives an excellent antidote to popular misconception. It explodes a number of widespread fallacies, such as belief in beginners’ luck among gamblers or fear of crime among the general public. But the main attack is directed at social scientists and economists who believe that the statistical “bell curve” is applicable to social, economic and commercial events. The relevant distribution is, in fact, a “fractal” distribution, which is so highly skewed that the averages can be disturbed by the addition or removal of a few individual observations. Examples are the distribution of personal wealth among US citizens and the distribution of sales of bestselling books. The causes of fractal distributions are strong positive feedback and the absence of physical barriers to growth. The rich get richer because they are rich, and they can do so without anyone actually shovelling gold out of the ground. Bestsellers sell better because they sell better, and it does not cost much to keep the presses rolling longer. The same applies to computer software, where costs of copying and distribution are now wholly negligible.

Such domains are called “scalable”, and Nassim Nicholas Taleb recommends his readers to engage in (or invest in) activities or professions with the scalable property because the potential returns are not limited by the amount of resource or effort expended in achieving them. So, after investing as much as you cannot afford to lose in the safest possible stocks, invest the remainder in a few speculative projects that are scalable. If the worst happens, cultivate the traditional calm of the philosophers of antiquity.

Among the most scalable investments today is investment in pure scientific research, especially if it has no predictable application. Experience shows how often improvement in pure knowledge and understanding leads to previously unimaginable benefits and innovations. Furthermore, it is only knowledge and understanding that can offer hope of protection from negative black swans. The tragedies of Aids and BSE could have been greatly mitigated if scientists had only known in time what they have found out since. Even the statistics of the bell curve have made enormous contribution to the progress of agriculture and medicine. Without them, many of the hungry would have starved and many of the sick died.

This is a marvellously readable book, if you can tolerate the racy conversational (and even jokey) style. You can simply dip into it wherever the section headings take your fancy. And they often do: “A new kind of ingratitude”, “Plato and the nerd”, “The expert problem”, “The tragedy of the empty suit”, “The grueness of emerald”, “How to look for bird poop”. Such dipping would not risk losing the thread because there is not much of one anyway.

Taleb is a man of wide learning and humane interests; he is an excellent raconteur; he has had an interesting life; he has made many interesting friends. And he is a skilled and experienced controversialist, with many pet hates: investment fund managers, social scientists, statisticians, philosophers (other than Karl Popper), Nobel prizewinners, dull high-school teachers and snotty European middlebrows (half the time, he even hates black swans themselves). I expect he hates book reviewers, too, but wisely keeps that fact to himself.

Tony Hoare is a principal researcher in computer science at Microsoft Research Ltd in Cambridge.