32

In the recent reputation recalculation resulting from weighting question upvotes back to 10 rep instead of 5 rep, some sites on the SE network underwent not only a reshuffle of the reputation leagues but a change in who's the #1 highest-rep user. These included:

  • Puzzling: I lost the top spot in May but got it back in the big November recalc. (I've posted 200+ questions and Gareth McCaughan none.)
  • Movies & TV: Paulie D (6 questions) had overtaken Walt (0 questions) this year to be the top user, but then Ankit Sharma (400+ questions) shot to the top in the big November recalc.

How many times did this happen on the whole network? How many sites got a new #1 rep user only due to the question rep change?

6
  • 2
    Considering the change is retroactive, a SEDE query isn't going to work here, is it?
    – Mast
    Nov 21, 2019 at 12:21
  • @Mast I'm not really sure how to find the answer, hence why I'm asking here :-) Nov 21, 2019 at 12:36
  • 10
    Writing has had one of our top 3 users delete his account over the Monica Cellio scandal, so that skews the balance. Nov 21, 2019 at 12:53
  • 3
    @GalastelsupportsGoFundMonica :-( Seems to happen too much there - I recall a previous top-rep user on Writing deleting his account too. Nov 21, 2019 at 12:55
  • I haven't seen any change to my repo on SO, is it rolling out slowly or something? Nov 21, 2019 at 18:17
  • 1
    @danielbeard It was all finished on (IIRC) 13 November. Nov 21, 2019 at 18:19

1 Answer 1

26

According to this SEDE query (which might not be entirely accurate, since it can't account for answer downvotes given by users), eleven:

The query compares the current situation (+10 reputation per question upvote) with the old one (+5). It does account for the daily reputation cap (except that suggested edits are ignored, but I doubt that would make much of a difference – these users have long reached the point where they can even edit tag wikis). Right now it's pretty accurate since SEDE was updated November 17th, and I've saved the current situation in the Wayback Machine.

The old reputation is kind of a 'virtual' situation, it will continue to count future upvotes with the old rules. So the query will eventually answer something different: which sites would have a different top user if we'd still have the old reputation system?

7
  • 2
    Thanks! I thought someone like you would be able to calculate this :-) Would it be possible to modify the query to include approximate rep scores as well? E.g. "old #1 user from 83k to 84k, new #1 user from 77k to 89k" or something like that. Nov 21, 2019 at 12:54
  • 2
    I thought that Stack Apps had 10 rep upvotes before this.
    – Laurel
    Nov 21, 2019 at 12:55
  • @Laurel thanks for catching that.
    – Glorfindel Mod
    Nov 21, 2019 at 12:56
  • 1
    @Laurel Or did they change question upvotes to 20 rep on Stack Apps? :-) If they decided to double it everywhere and forgot that special case. Nov 21, 2019 at 13:05
  • That is a nice computation. Did you include also the effect of bounties? (not that it will be so much). You could take your computed difference in new and old rep and subtract it from the rep of a user (then you do not need to iterate over all the bounties given or obtained). Nov 22, 2019 at 8:19
  • 1
    Thank you! Bounties don't need to be taken into account, since they don't affect the daily reputation cap, and the starting point of the computation is the actual reputation of the user.
    – Glorfindel Mod
    Nov 22, 2019 at 8:21
  • Ah, I missed that 'starting point' in the computation. (I will study your query a bit more. I have grown some interest in this stuff - getting information from databases - but I am not very proficient in it yet) Nov 22, 2019 at 8:22

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .