Why waiters (or their guests) can’t predict the weather

Apr 22, 2012 Humour, Strategy, Sunday Nation

I was sitting by my favourite ocean (there is only one) the other day (I was on a break, remember) and I noticed some ominous-looking dark clouds over the ocean. I asked a passing waiter whether he thought it might rain.

He looked at the sky, and said with gratifying certainty: “No chance. Those clouds are not coming here.” Reassured, I re-immersed myself in my book, ignoring the fact that the sky was darkening and the wind was becoming more insistent. Half an hour later, I was sitting in the hotel lounge soaked to the skin, having just managed to protect my book.

The book itself truly deserved saving: it is called Fooled by Randomness, and it is written by that very entertaining and erudite iconoclast, Nassim Nicholas Taleb. I love iconoclasts of all shapes and sizes; their being entertaining is an additional bonus.

Taleb tells you that if you place an infinite number of monkeys in front of typewriters and let them clatter away, it is CERTAIN that one of them will produce an exact version of the Iliad. This is not very interesting, because the probability of any one monkey doing it is ridiculously low.

But here’s the thing: should one monkey pull it off and present you with the Iliad, would you now be willing to consider the possibility that he could next write the Odyssey? No? What if you knew nothing about this experiment with infinite monkeys, and were simply presented with this hero-monkey who had just (verifiably) written the Iliad: would you not be rather disposed to expecting another great work from this remarkable primate?

The waiter was confident because his last couple of predictions about the rain had been correct. He was on a run, and thought he had exceptional skill in weather forecasting. The point though, is this: pretty much every waiter I’ve ever met on any beach thinks the same thing. They all think they’ve got the weather figured out; none of them actually do.

This is because the underlying process by which it rains on a particular beach at a particular time is pretty random. It is influenced by humidity, cloud formation and winds in ways that is very difficult for even a professional weatherman with proper measuring equipment to forecast. So the correct answer to the question, “Is it about to rain on this beach in the next half-hour?” is, “Who knows? Take precautions in case it does.” That is the answer almost no one gives you.

As you may have worked out, this column isn’t really about waiters or monkeys. It’s about you. Don’t you have the same irrational confidence in your forecasts – about the stock market, business performance, exchange rates? Don’t you feel that a string of successful predictions make you a forecasting guru – when in all likelihood that string was just a run of luck?

If the underlying process creating the result is stochastic, you have no reason to be confident. And many more results – share price performance, currency rates, business and personal success – have strongly random underlying generators than most of us imagine.

So abandon your bravado and get real. You may know much less than you think. Think more deeply, and get other opinions. Don’t make big bets on your (or anyone else’s) predictions. Be prepared for forecast failure, and be prepared for the possible consequences. Stay humble, and recognize the huge role of luck in your life. We are quick to extol the virtues of a businessperson who has had a successful run, and quick to forget the many others who failed by doing the exact same things as our “winner.”

My waiter friend learned from his forecasting error, and refused to attempt a prediction the following day. He’s wiser than many overconfident business leaders I know.