
Nassim Nicholas Taleb @nntaleb
@Stefano_Peron @financequant Actually, Pareto is already multiplicative. If you maximize entropy under constraints E[x}, V[x], you get a normal. Under constrainsts E[Log[x]], V[log[x^2]], you get a lognormal, and under just E[Log[x]], you get Pareto. It is sort of a super Lognormal. — PolitiTweet.org