A method for measuring uncertainty

Here's one simple technique for measuring uncertainty. It's called the equivalent bet.

Say you have $100 and you're going to bet it on one of these two propositions:

1. I give you a spinning dial (think “Wheel of Fortune” or the board game Life) the face of which is completely green, except for one red wedge that takes up 20% of the dial. You bet $100 that when you spin that dial, it will come up on green the first time.

2. You bet $100 that George Washington was born between 1700 and 1710. (Using only what you know right now — no Googling!)

Got it?

Okay, which bet would you take?

…

…

…

If you prefer the second bet, about George Washington, that's a pretty good indication that you're more than 80% sure about this range of dates for Washington's birthday.

And if you prefer the first bet, with the spinning dial, it must mean that you’re less than 80% sure about the Washington dates.

Congratulations. You've just measured uncertainty.

What's the point?

Obviously you can look up George Washington's date of birth on your phone.

And it doesn't have much value for your business decisions anyway.

But there are plenty of questions that do affect your business decisions, for which you can't just look up a correct answer. You have to estimate.

You can use the equivalent bet technique to measure the uncertainty, or to adjust your estimates to a measurable level of uncertainty.

Don't know anything about American history? You could expand the range of dates until it feels like greater certainty. (I mean, surely we can be 100% certain that Washington was born between the years 1000 and 2023, right?)

Here's the thing:

Disregarding an estimate because it's "just an estimate" is as unnecessary — and risky — as treating it like a guarantee.

The hard part is knowing where it lies, somewhere between “useless” and “gospel.”

But there are ways to figure that out. And they're worth using.

All the best,
A.