@xortdsc, I’m not a math person, but for a higher level perspective, what you describe would always generate evenly distributed encodings. I remember having a related discussion on this thread. That was related to a bit different implementation than what you are describing, but I think the points @blue2 raised there still apply.