And one thing to remember is that (a) actual random distributions are lumpy, not evenly distributed and (b) there is pretty much no equality of outcome in whatever outcomes or input categories you choose. Anywhere. And very often the distributions in different places also look different (though there do seem to be some patterns).