-14

Note sure where could I ask this... I would've expected reputation to roughly adhere to power law but instead I see very odd clustering where a pair of people have identical or near identical reputation:

travel SE rank

Was there any research made into the reputation distribution and the potential reasons behind such?

Edit for the downvoters: perhaps I should've posted to math SE? Could you provide a better place to ask?

Edit2: yes, M.SE reputation distribution is a (partial) answer to it.

5
  • No idea what you mean. Distribution based on what, exactly? Name? Age? Location? Also, in Stack Exchange it is called "reputation", not karma. Commented May 3, 2016 at 11:09
  • en.wikipedia.org/wiki/Pareto_distribution here's an example.
    – chx
    Commented May 3, 2016 at 11:12
  • So you want something like this, just for all sites? Commented May 3, 2016 at 11:27
  • 1
    Looks like the list you have is sort ordered by some metric
    – random
    Commented May 3, 2016 at 12:01
  • @random No, that is just the Travel.SE all time ranking. Commented May 3, 2016 at 13:41

3 Answers 3

7

I think you'll actually find that there is nothing weird here.

The only thing actually happening here is the human mind's propensity for finding patterns.

If I search through any of the reputation lists, I can just about guarantee I'll find pairs every now and then. This isn't coincidence. It's fairly inevitable as the numbers of users increase.

The graphs some sites have created (eg Maths) show that the curve is as you'd expect, with the extra spike brought in by the 100 participation bonus.

2

Having nearly the same reputation is just a matter of coincidence and there isn't really clustering.

Karma is the wrong word too: it is the total reputation earned from votes received on questions and answers. The distribution of it doesn't adhere to years of membership or something like that, just the usefulness of their contribution in time.

At most, there can be a relationship between the number of posts and reputation, but that doesn't work either since some users need less posts to have the same reputation someone else has with a lot of posts.

3
  • Sigh. How is this even an answer to my question? You basically presume I do not know what reputation is. Please check stackexchange.com/users/118083/chx?tab=accounts -- I know.
    – chx
    Commented May 3, 2016 at 11:18
  • 8
    Clearly you don't. Else you wouldn't have called it Karma in the first place. Commented May 3, 2016 at 11:24
  • @PatrickHofman That doesn't change his point that this is in no way answering the question. This should just be a comment at most.
    – Servy
    Commented May 3, 2016 at 13:34
0

This is largely because of the reputation cap. When looking at the top users of a site, they're all earning the reputation cap every day, and while they won't necessarily be earning exactly the same amount of reputation as some activities don't count towards it, they'll all be earning very close to the same amount of reputation each day.

8
  • The reputation cap has nothing to do with it. The screenshot is from a group of users with 11K. With just 11K you can't say the rep cap has influence on those users being packed together. Commented May 3, 2016 at 13:36
  • And looking to one of those users, the reputation cap is never met. Commented May 3, 2016 at 13:38
  • @PatrickHofman He's not showing the SO rankings in that screenshot; it's the 30th ranked rep user in that screenshot. That said, the answer applies to the question, not just the example screenshot. The various sites don't follow a power distribution, and the reason they don't (and can't, really) is because of the rep cap.
    – Servy
    Commented May 3, 2016 at 13:38
  • I didn't mention SO, since it would be very weird to have just 11K on SO and be in the top 30... Commented May 3, 2016 at 13:40
  • Your answer might be true for some sites, primarily SO I think, but not for other sites and definitely not the one OP is talking about. Commented May 3, 2016 at 13:40
  • @PatrickHofman The OP didn't mention any sites. He asked why SE sites, in general tended to not follow a power distribution. You, and only you have so much as mentioned travel.SE. Odd that you'd downvote an answer while also stating that that it's correct (and also posting a non-answer that doesn't even attempt to answer the question).
    – Servy
    Commented May 3, 2016 at 13:42
  • OP uses his screenshot to make a point: "I see very odd clustering where a pair of people have identical or near identical reputation" and then presents 'proof' of that. Your answer doesn't answer that case at all. Commented May 3, 2016 at 13:43
  • @PatrickHofman That wasn't the actual question that he asked though. That's just a single observation made in the question. The question is why do SE sites have the reputation distribution that they do.
    – Servy
    Commented May 3, 2016 at 13:44

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