[amelius]

I didn't want to imply that there is a loss of information. Yes, it's one level up in the hyperoperation chain, I suppose, but what I meant is that it's not a typical way of doing statistics, especially the higher moments.