Advertisement
Articles
Advertisement

CICIORA’S CORNER: Black Swans

Wed, 09/30/2009 - 8:25pm
Walt S. Ciciora, Expert on Cable and Consumer Electronics Issues

It is always important to be aware of fat tails and the Black Swan

Being “semi-retired” means I have time for reading. So this summer, among the books I read, two were especially noteworthy. Both are by Nassim Nicholas Taleb, originally from Lebanon. Walter S. CicioraThe books are “The Black Swan” and “Fooled by Randomness.”

Taleb’s family became relatively well off in Lebanon by hard work over several generations. The reasonable expectation would be for more success. But due to political instability and fighting, the family lost it all and had to flee. This experience caused him to look at life in a different way that can be insightful to us in our professional and personal lives.

A “Black Swan” is defined as a highly improbable consequential event.

It’s a surprise, and usually not a pleasant one. This concept derives from the experience of generations of people only seeing white swans. After becoming convinced that swans could only be white, a species of black swans was discovered in Australia. This gave rise to the little bit of wisdom, “absence of evidence is not evidence of absence.” It takes just one counter-example to prove wrong a theory of “there aren’t any.”

Taleb gives a fine example of a Black Swan event from the perspective of a turkey. The turkey’s experience leads him to believe that he is fortunate to be in the care of a very nice family. Every day they bring him plenty of his favorite food and clean cold water, and they talk affectionately to him. Life could not be better. Then one November day, the Black Swan event occurs. From his experience, this was highly improbable; it did not fit with any of his past experience. Yet it was certainly a consequential event.

There’s another aspect to this story. People, and turkeys, fail to learn from history because many of history’s consequential events occurred before they were born. They have no direct experience with these events, and so believe them to be highly improbable. Anyone born in the mid-1980s won’t remember the double-digit inflation of the early 1980s and may consider its repetition to be highly improbable. This is a rather dangerous assumption, especially in these times.

Taleb’s intellectual curiosity focuses on randomness. He spent his early career on Wall Street and became convinced that a great deal of what goes on there is just plain luck. The books are filled with examples and insights that will reward your time spent reading.

He gives an interesting coin toss example. If we start out with 1,000 people who toss a coin – heads winners, tails losers – we expect, on average, for there to be 500 heads. Eliminate the losers and do it again and there will be about 250 heads, and again 125, then 62, and 31, and 15. A crowd begins to form to cheer on the talented ones, the ones who keep flipping heads. Bets are placed. Next comes eight winners, then four, then two, and finally one.

What secret does the winner have? How did he manage to flip 10 consecutive heads? It seems impossible. Can we get him to flip the coin one more time so we could study his technique closely? Maybe a high-speed filming? Likely if he agrees to flip again, he has a 50 percent chance of being embarrassed. How many heroes on Wall Street and elsewhere in life just happened to be lucky and then fail because they believe too much that it was talent and not luck? With enough participants in the game of luck, isn’t there going to be a winner?

My interest in these subjects has its roots in a course I took in college, “Lectures on Random Signals and Probability Theory.” (We jokingly turned that title around into “Random Lectures on Probably Signal Theory.”) The course dealt with the problem of recovering desired signals obscured by random noise. While signals and noise were of interest in analog cable, the benefits of this theory blossomed in the digital realm. It turns out that to make the mathematics of the theory tractable, a critical assumption had to be made. It was necessary to assume the noise was normal Gaussian white noise. That is, the voltages of the noise were distributed according to the well-known Bell Curve. That assumption made the mathematics workable.

But assuming a Bell Curve distribution is a dangerous assumption. It frequently doesn’t work. So I learned decades ago to be leery of the “Gaussian” assumption. Taleb’s Chapter 15 in “The Black Swan” goes so far as to call the Bell Curve a “great intellectual fraud.” His book is filled with clever turns of phrase, and quite fun insults, against those he finds insufferable.

In fact, strictly applying the Gaussian Bell Curve yields predictions of something bad happening once in a thousand years. It is responsible for two consecutive years of hundred-year floods. The actual distribution has “fat tails” compared with the idealized Bell Curve.

As we consider what is likely and unlikely to happen in our professional and personal lives, it’s important to be aware of fat tails and the Black Swan. I recommend those books.

wciciora@ieee.org

Advertisement

Share This Story

X
You may login with either your assigned username or your e-mail address.
The password field is case sensitive.
Loading