@predict_addict@sigmoid.social it's even worse. If you have just correlations the distribution will converge towards a gaussian if your portfolio is just large enough, and things are reasonably little correlated. But the thing is that returns may even be sampled from fat tailed distributions. Very extreme values are more likely than under a normal distribution. Then you get the case that whatever it converges to is actually a Levy stable distribution, itself with power law tails.