Black Swans, Broken Windows, and Magicians

Magicians can make many things disappear. But not broken windows or black swans.

When I was young I was fascinated by magicians. With a bit of waving of their hands they could make people disappear, shrunk in half, or turned into pigeons. I still sometimes wish I could do that. But the only thing I can make disappear is money, not people.

Of course, as a kid I already knew it was all show and illusions. In fact, the First Law of Thermodynamics states it is impossible to make things disappear. Mass and energy are constant. It has to go somewhere. (And when it has the form of money it often goes into the pockets of the magician.)

Broken Windows

Some people have tried something similar with the Broken Windows theory. It is the idea that, when an environment deteriorates, people’s behavior deteriorates along with it. The “magicians” have claimed that the broken windows effect doesn’t exist, and they pointed victoriously at the New York crime rate example, famously described as a broken windows effect in Malcolm Gladwell’s book The Tipping Point. It turned out that the New York crime rate statistics were a bad example. And, while pointing out that the broken windows effect didn’t exist there, the “magicians” tried to convince us it didn’t exist anywhere else either.

But that was only an illusion.

The broken windows effect has been proven empirically by researchers in The Netherlands. It also follows logically from Lewin’s Equation (people’s behavior depends on their context) and memetics (people copy each other’s norms). Common sense dictates that discrediting one example of the broken windows effect only shows that the example is bad. But an attempt to use this to discredit the whole concept of broken windows is a Fallacy of Hasty Generalization.

Black Swans

Black-swan Similarly, some magicians are trying to make Black Swans disappear. This is the idea that some events are hard to predict, have a low probability of occurring, but have a high impact when they do occur. The magicians suggest that “black swans do not exist”, while pointing victoriously at the examples of the financial crisis (used by Nicholas Nassim Taleb) and the war in Iraq (used by Donald Rumsfeld). They say plenty of people predicted the financial crisis and the trouble in Iraq, and therefore “there are no black swans”.

Again, this argument is merely an illusion.

Unlikely and unpredictable events with big effects were suggested by academics, complexity researchers and systems thinkers long before Taleb named them swans and painted them black. The existence of such events follows logically from Gödel's Incompleteness Theorems (no description of the world can be complete), Heisenberg’s Uncertainty Principle (no observer can know everything), and the Butterfly Effect (even the smallest events can have big consequences). It also follows empirically because nobody (including the magicians) had predicted that such low-predictability-high-impact events would someday be called “black swans”. And now the metaphor of black swans is used all over the world, and having a significant impact on our discussions.


Some magicians can put up a great show. They sometimes make it hard to trust common sense. And I don’t blame you. I even like a bit of magic myself every now and then. But please, try to enjoy the show for what it is: an illusion. When determining whether something exists or not, rely on the words of the scientists, not the hands of the magicians.

  • Many Preconceptions about Scrum (Dutch article)
  • 7 Questions for Philippe Kruchten
Related Posts
free book
“How to Change the World”
  • Andrefaria

    I think it should be a principle for every software development team to never allow a window to be broken.

  • Glen B. Alleman

    Some definitions of a Black Swan:
    1. The disproportionate role of high-impact, hard to predict, and rare events that are beyond the realm of normal expectations in history, science, finance and technology
    2. The non-computability of the probability of the consequential rare events using scientific methods (owing to the very nature of small probabilities)
    3. The psychological biases that make people individually and collectively blind to uncertainty and unaware of the massive role of the rare event in historical affairs.
    It’s the last one that occurs many times in commercial software development. It’s the first two that are used in domain, where probabilistic risk dominates the conversion – nuke power, manned spaceflight and the software intensive systems that control them.
    It may be useful for the commercial domain to look at the space and defense domain it see how to detect Black Swans and decide if that process adds value.
    In all three case the Black Swans are there – the observer just may not be able see them and conclude that they aren’t there. It’s that old saying “what I don’t know (or choose not to see) can’t hurt me.”

  • Wtpayne

    I really, really, really like that aphorism.

How to Change the World - free Workout - free