How does the system extrapolate from the existing userbase how many users there are likely to be with at least a certain rep score in forty days' time? I can think of a few possible algorithms none of which make much sense. Has the algorithm used ever been made public?
(I've seen this [bug] report from 2012, but I don't think this is a duplicate. The old meta post has a lot of guesses at possible algorithms, but nothing is confirmed.)