This is a bit of a nit-pick, but I don't let my students get away with this, so it seems worth mentioning. In the reputation section of a user's page it might say something like:

  • Reputation
  • top 0.90% this year

It seems that in the last line, either the decimal point or the percentage sign should be removed, since I don't think that this indicates that the user is in the top 9/1000 of users for the year. (Well, they are, but that's a rather poor lower bound!)

  • $\begingroup$ I think you're just really that good: the corresponding number in my profile is a plain 8%. Perhaps on some of the larger sites most of the active users are in the top few percent, so they find the added precision useful. $\endgroup$
    – Ben Barber
    Commented Nov 13, 2015 at 19:43
  • $\begingroup$ Alternative comment: top 0.9% is in fact quite a good lower bound, unlike 0.9th percentile. $\endgroup$
    – Ben Barber
    Commented Nov 13, 2015 at 19:51
  • 8
    $\begingroup$ Only users with at least 200 points are counted here. -- You are presently ranked 28th among 3108 ranked users, and $28/3108 = 0.0090 \dots$. This is where the "top 0.90% this year" come from, as far as I understand. $\endgroup$
    – Stefan Kohl Mod
    Commented Nov 13, 2015 at 20:27
  • 2
    $\begingroup$ The way this is intended to read is that if that if number of users is $X$ this particular user is among the $0.009 X$ best users, or is the $0.009 X$ best user, up to rounding. Thus the smaller the number the better. JDH is top 0.03% overall, while I am only top 2%. (The same for year.) I am not clear on what the perceived bug is. Could you please clarify this more. $\endgroup$
    – user9072
    Commented Nov 13, 2015 at 20:29
  • 7
    $\begingroup$ Okay, thanks to all, I was mis-interpreting. For some reason, I was taking "top 90%" to mean "better than 90% of the users", instead of meaning "better than 10%" of the users. Since the question reflects my total lack of understanding, maybe I should just delete it. $\endgroup$ Commented Nov 13, 2015 at 20:46
  • $\begingroup$ This question seems a bit related: An error in calculations of percents in reputations? $\endgroup$ Commented Sep 8, 2019 at 15:40


You must log in to answer this question.

Browse other questions tagged .