Selected success alone is not evidence of skill. Judge the process across histories that could have happened, not only the winner that survived.
Equal variance does not imply equal tail risk: in Extremistan, a few observations can dominate the total and defeat a fence fitted to short history.
At the same average shock, curvature—not prediction—decides whether more disorder harms or helps.
Here, crossing 20% of starting wealth is an absorbing failure: an exposure removed from the game cannot average its way back.
Sharing downside realigns choices; it does not guarantee competence. One lucky path can still reward the wrong contract.
An operational reading of Taleb’s aphoristic metaphor: the model becomes tidy by clipping and quantizing what refuses to fit.