by das monde
Tue Sep 25th, 2007 at 11:15:23 AM EST
Recently I picked up a random book at the Schiphol airport, it was "Fooled by Randomness" by Nassim Nicholas Taleb. Still doubting whether I need to know more about random things, I paid for the book.
The author is a no-nonsense practitioner and educator of uncertainty, hunting at Wall Street for upshots of risk ignorance. The book considers probability not as an engineering or computational discipline; the chapter on human misconceptions about randomness is acknowledged as the least original part. Instead, probability is introduced as a method of dealing with our ignorance and lack of certainty.
Most illustratively, the author develops stories of three Wall Street traders (Nero, John, Carlos) whose fates were vigorously shifted by the 1998 summer crises. You can look very smart for some time, but the same "skills" may blow you up at some rare circumstances. And vice versa, great achievements (or mere survival) may be preceded by long periods of
"underperformance":
[Think] of someone involved in scientific research. Day after day, he will engage in dissecting mice in his laboratory, away from the rest of the world. He could try and try for years and years without anything to show for it. His significant other may loose patience with the looser who comes home every night smelling of mice urine. Until bingo, one day he comes up with a result. Someone observing the time series of his occupation would see absolutely no gain, while every day would bring him closer in probability to the end result.
Reasonably successful and cautious people may have difficult social times living in a neighborhood of "better performers", even if hotter performance requires taking pure chances. Stock trading or company management may be very rewarding, but they are not as skillful as dentistry, cooking, or playing Rachmaninov.
CEOs take a small number of large decisions, more like a person walking into the casino with a single million dollar bet. External factors [play] a considerably larger role than with the cook. The link between the skill of the CEO and the results of the company are tenuous. [The CEO] may be subjected to the monkey-on-the-typewriter problem. There are so many companies doing all kinds of things that some of them are bound to make "the right decision".
What does there to say about investors paying enormous salaries to CEOs for perhaps 90% random performance? The stock market game could be like the Russian roulette only with much more empty chambers. Imagine the oligarch Abramowitch offering you $10 million if you stay alive after one shot. Odds are good, only the worst case is terrible.
An extreme case of an asymmetric bet is known as Pascal's wagger: you rather believe in God, because you have way to much to loose if he does exist. Are we not in the same situation regarding global warming or peak oil?
A perfect model (absent in the book) of luck determining fat rewards is modern poker tournaments. The last few World Series of Poker championships were won by relative novices, often getting to the main event by winning small satellite tournaments online or in a local Indian casino. Then you outlast the other few thousand players in an aggressive fest of "all-in" moves. Daring to go "all-in" for near 50-50 chances is a necessary skill of tournament poker, negating many other skills. The latest winner, Laos born South Californian Jerry Yang has mathematical and psychologist background, but undoubtedly quite a number of players were using the same skills and strategies but were eliminated by luck.
Human emotions can hardly acknowledge random factors and evolutions, especially when political and military leadership, or corporate fortunes are at stake. Sensationalist media is doing its best to overwhelm and exaggerate the senses. Are we helpless against our emotions and mind making connections whenever conceivable?
Go to the airport and ask travelers en route to some remote destination how much they would pay for an insurance policy paying [say, $1 million] if they died during the trip (for any reason). Then ask another collection of travelers how much they would pay for an insurance that pays the same in the event of death from a terrorist act (and only from a terrorist act). Guess which one would command a higher price? Odds are that people would rather pay for the second policy (although the former includes death from terrorism).
The author is content to confront natural psychological biases against randomness at work, and gladly enjoys its playful deceits in arts. It is quite an achievement to meet randomness with a clear mind. People may even have some intuitive tools to deal with uncertainty after all, if only they are not fooled by modern "certainties", such as (in my opinion) progress by greed?
I would not agree with everything in the book. Say, generation of "truly" random number sequences is a serious mathematical problem, with implications for the Monte Carlo simulations favored by the author. The particular division between Utopian and Tragic Visions of humankind is debatable: falsificationism of Karl Popper is not particularly related to homo sapiens (being counter-intuitive above that) but is an issue for any cognitive subject. (To my view, Popper is presented very much in Feyerabend's way.) Secondly, Milton Friedman's "distrust" of governments might soon become the greatest utopia under implementation till now. Human knowledge will never be complete, and humble realization of that is very healthy. But it is still usually valuable to know more, especially when there are probably some fellow humans bound to shock and overwhelm you with confusion and uncertainty.
But I have to say, I enjoyed new understanding of randomness from the book.