Mediocristan
Taleb calls a domain Mediocristan when, in a sufficiently large sample, no single observation can dominate the total. Human weight illustrates the idea: an unusually heavy person can be striking individually while contributing little to the combined weight of a large group.
THE IDEA
TRAVELS INSIDE MEDIOCRISTAN
This scalable/nonscalable distinction allows us to make a clear-cut differentiation between two varieties of uncertainties, two types of randomness.
Let's play the following thought experiment. Assume that you round up a thousand people randomly selected from the general population and have them stand next to one another in a stadium. You can even include Frenchmen (but please, not too many out of consideration for the others in the group), Mafia members, non-Mafia members, and vegetarians.
Imagine the heaviest person you can think of and add him to that sample. Assuming he weighs three times the average, between four hundred and five hundred pounds, he will rarely represent more than a very small fraction of the weight of the entire population (in this case, about a half of a percent).
You can get even more aggressive. If you picked the heaviest biologically possible human on the planet (who yet can still be called a human), he would not represent more than, say, 0.6 percent of the total, a very negligible increase. And if you had ten thousand persons, his contribution would be vanishingly small.
In the utopian province of Mediocristan, particular events don't contribute much individually—only collectively. I can state the supreme law of Mediocristan as follows: When your sample is large, no single instance will significantly change the aggregate or the total. The largest observation will remain impressive, but eventually insignificant, to the sum.
I'll borrow another example from my friend Bruce Goldberg: your caloric consumption. Look at how much you consume per year—if you are classified as human, close to eight hundred thousand calories. No single day, not even Thanksgiving at your great-aunt's, will represent a large share of that. Even if you tried to kill yourself by eating, that day's calories would not seriously affect your yearly consumption.
Now, if I told you that it is possible to run into someone who weighs several thousand tons, or stands several hundred miles tall, you would be perfectly justified in having my frontal lobe examined, or in suggesting that I switch to science-fiction writing. But you cannot so easily rule out extreme variations with a different brand of quantities, to which we turn next.
When to use it
Forecasting average product returns in retail where outliers are bounded.
Setting insurance premiums for typical health claims (excluding catastrophic events).
Estimating average commute time in a city where extreme delays are rare and bounded.
Managing inventory for items with steady demand and limited variation.
FROM THEORY TO EVERYDAY LIFE
In practice
Weight of the heaviest person in a sample
In a group of 1000 people, Taleb illustrates adding an unusually heavy person, with 500 lbs as an example weight. The point of the estimate below 0.6% is that a single person still contributes little to the group’s total weight.
Annual caloric consumption
Even a day of extreme overeating (e.g., Thanksgiving) does not significantly affect yearly caloric intake of about 800,000 calories.
A MOMENT TO REFLECT
Put it to work.
What are examples of quantities in your daily life that behave as if they belong to Mediocristan?
How would decision-making change if you mistakenly treated an Extremistan variable as if it were from Mediocristan?
Why is it important to distinguish between scalable and nonscalable randomness in risk management?
Often confused with
CONTINUE THE CONVERSATION
Source texts
- The Black Swan The Impact of the Highly Improbable With a new section On Robustness and Fragility (Nassim Nicholas Taleb)