What Is Mediocristan? Nassim Taleb's Definition
Mediocristan is the statistical domain where outcomes are dominated by many small variations, with no single event dominating the total. Most variations cancel out. Averages are meaningful. Past predictions are useful.
Taleb's example: human height is Mediocristan. The tallest person in the world is maybe 3x the height of the shortest. No single person is 1,000 times taller. The average height of a population is a useful statistic that predicts individuals reasonably well.
Characteristics of Mediocristan
- Many small variations with no one dominating
- The average is representative of the typical case
- Rare events are rare
- Past distributions are useful for predicting future events
- Linear causality: more effort produces more results proportionally
- Error correction happens naturally through feedback
Examples: income for a taxi driver, test scores across students, daily variation in temperature, restaurant revenue.
Taleb's Contrast: Extremistan
Mediocristan is contrasted with Extremistan, where rare events dominate.
In Extremistan (wealth distribution, book sales, financial returns), the top 1% accounts for more than half of all outcomes. No single person being 3x wealthier than another is possible — they're 1 million times wealthier.
In Mediocristan (height, daily weather), the above is impossible.
This distinction matters enormously for prediction and risk management.
The Practical Implication
In Mediocristan-type professions (steady jobs, traditional employment, skilled trades), the advice "show up, work hard, get better, get promoted" works reasonably well. Linear effort produces linear results.
In Extremistan-type domains (creative fields, entrepreneurship, finance, technology), this advice fails. Linear effort produces unpredictable, power-law distributed returns. You might work hard and achieve nothing, or luck into something that transforms everything.
Understanding which domain you're in changes what strategies you should pursue.
Go deeper:
For the full breakdown of Mediocristan and Extremistan, read Mediocristan vs. Extremistan: Two Types of Randomness.