How would you like to get a little smarter by listening to one of the funniest dick jokes ever told? I know. We all dream of such a thing. Or maybe that's only me. Either way, it sounds both glorious and implausible. But bear with me. Get to the end of this and you will be a) smarter and b) laughing your ass off.
, Phil and I discuss how superforecasters methodically unpack questions and pursue each sub-question step by careful step. As an illustration, we look at the legendary physicist Enrico Fermi, who made uncannily accurate estimates. He taught his methods to students by asking them to answer questions like this: “How many piano tuners are there in Chicago?”
No, you can’t Google the answer, or draw on any information source. You have to make an estimate based on nothing more than your own thoughts. So how do you make the most accurate estimate possible?
Most people will scratch the back of their head, roll their eyes, think vaguely about how many people there are in Chicago and how many common it is for people to own a piano and … Out will come a number. Ask them why they chose that number and not some other number and they will shrug. They don’t really know. It just feels like this one is in the ballpark. Maybe.
Fermi showed there is a much better way.
Unpack the question by asking “what would I have to know to be able to calculate the answer precisely?” List that. Of course you won’t actually have that information. But then you can further unpack each item on the list by again asking “what would I need to know to be able to calculate the answer precisely?” Keep doing that and you will, eventually, be forced to make guesses. But when you make those guesses, and combine them to come up with a final answer to the first question, your estimate will be more accurate than if you had followed the old scratch-my-head-and-guess method.
Why? When you use the scratch-my-head-and-guess method, what’s inside your brain is a big jumble of information, beliefs, assumptions, and biases — and somehow or other, your guess emerges out of that mess.
But when you methodically unpack the question as advised by Fermi, you are forced to take that jumble apart. That bit of information? Take it out and lay it on the table. That belief? Lay it out next to that information. That assumption? Don’t ignore it. Put it on the table.
With everything laid out neatly, you can look at it slowly and carefully. Mistakes are corrected. Biases are adjusted for. Assumptions are critically examined.
Net result: a more accurate estimate.
As important and illuminating as Fermi estimation is, it’s not exactly scintillating stuff. Piano tuners in Chicago? Enrico, you’re boring the audience! But I guess you don’t have to worry about such trivia when you’re a legendary physicist.
I am not a legendary physicist. Hence, you can imagine how excited I was when I discovered a comedy sketch in which computer engineers tackle a question Fermi-style — and simultaneously deliver one of the funniest dick jokes of all time.
I’m not kidding. It happened on the brilliant Mike Judge satire Silicon Valley
Of course my excitement lasted only for about the minute it took me to realize that as clever and entertaining as the sketch is, it is completely obscene. There's no way in hell I would ever be able to use it in a book or talk.
Which seems like such a waste.
So here it is
. A great illustration of methodical problem-solving. And a great dick joke.