We present a new version of the Central Limit Theorem for multivariate martingales.

We investigate the set of centers of completely and jointly mixable distributions. In addition to several results, we show that, for each n ≥ 2, there exist n standard Cauchy random variables adding up to a constant C if and only if |C| ≤ n*log(n − 1)/π