l5-the-black-swan_market-concentration

According to Taleb's analysis, the ten most extreme days in fifty years of market data represent half of total returns. Removing those days changes the picture of stock market performance beyond recognition, showing that rare events dominate outcomes.

Financial returns are scalable and fat-tailed, not normally distributed. Dozens of papers show the bell curve's inadequacy, yet practitioners revert to Gaussian tools because the numbers feel better than nothing, even when they are actively harmful.

Extremistan is Taleb's term for domains where single extreme observations dominate the total. The concentration of half of market returns in ten days is presented as proof that markets belong to Extremistan rather than the mild randomness of Mediocristan.

Avoid Gaussian tools whenever analyzing scalable phenomena like financial returns, where extreme events carry most of the impact. The skill warns that domain-dependent thinking—accepting the critique in theory but reverting to Gaussian tools in practice—is the most dangerous failure mode.