Mediocristan and Extremistan

There are two statistical worlds in which we live. Nassim Taleb has described them with names Mediocristan and Extremistan. The first world is a normal distribution world and the second world has extreme distributions. In mediocristan, two persons that have a combined length of 4.00 meters are close to each other. For example, they have 2.01meters and 1.99 meters in length. In extremistan, there are two persons whose combined wealth is 10 million. It is most likely that one of them has wealth close to that of 10 million and the other has wealth about 100,000. As you can see, these worlds are totally different from each other. In other words, one figure does not change statistics much in Mediocristan, but one figure can change everything in Extremistan.

Mediocristan

Mediocristan is the world where the majority of the people think we live in. This applies even to the majority of the most educated persons. Success equals effort and skills in this world. Mediocristan applies to most biological effects on humans. Genes work in ways that produce mostly results that can be found the normal distribution. The results that genes produce have high predictability. You can define probabilities to them and they do not change much. Here are some percentages how much some biological or psychological attributes of humans are explained by the genes they inherit:

• Height 70%, Weight 80%
• Reading disability 60%, Verbal ability 60%
• Face remembrance 60%, Spacial ability 70%
• General intelligence 50 %, Personality 40%

As you likely know, your biological parents´ genes affect your biological and psychological traits. If they have some extreme traits, likely, these traits are not so extreme in you. In a world with normal distributions, return to averages happen fast. If you have a height of 2.10 meters, your son´s height will likely be closer to average.

Extremistan

Extremistan is the place where a minority of the people think we live in. Extremistan applies to most social effects on humans. Success is non-linear with effort in this world. Social contagions produce results that produce extreme distributions. Most of them are from Extremistan. Some social attributes that are from Extremistan are book sales per author, wealth, and sizes of companies. There is no predictability in Extremistan. It is almost impossible to predict extreme statistical distributions like changes in stock prices. Even figuring out the mean of a simple 80/20 Pareto distribution requires a sample size of a hundred billionfold compared to normal distribution according to Taleb. Changes in stock prices have much larger extreme distribution. Therefore, they cannot be predicted in any modern computer or human brain. This means is that it is much easier to prove that someone is wrong than what the reality is. Today´s world suffers from the domination of high-effect, low-probability events. Most scientific breakthroughs happen when they are not expected, instead of deadlines put to researchers.

Events and effects that combine both worlds

Some events happen in both worlds. For example, some economic effects from predictable catastrophes like earthquakes have extreme distributions. An earthquake that has twice the power than the other can cause tenfold economic effects on people. The earthquake in the same magnitude as 100 years before can also cause manyfold effects today than it did then. These kinds of socio-economic effects grow in magnitude when more people live in big cities, and we use more of nature´s resources.

The modern world grows larger and larger winner-take-all effects. Even a little bit more skillful athletes can earn ten times more money than athletes a few percentage points below their skill levels. The same effects can apply to authors, musicians, and other artists. But this does not always apply. Even a little bit luckier author with the same skills can make a hundredfold more money than others.

Nothing to add,

-TT