2
$\begingroup$

Wouldn't a "Reputation per Number of Posts" stat on our profiles, and a possible badge for some minimum amount of the same, discourage low-quality and/or spam posts?

The reason I ask for this is that in my admittedly little experience on this site, I have come across a few high Rep users who seem to be prone to posting hasty answers.

The penalty for a post which simply did not happen to be well-recieved is not much either- for every upvote we receive $10$ Reputation, whereas for every post that did not receive any attention, the denominator will increase by just $1$. This means that we would not be discouraging users from posting good answers or questions that they aren't completely sure of.

However, this might incentivise high-quality answers as well as questions. Of course, for the proposed badge, it is necessary that some minimum Reputation be a pre-requisite, as it is much easier for a user with $1$ post which got, say, $10$ upvotes, to gain a $100$ Rep per Post, compared to a user with a thousand posts.

So, why don't we do this? Am I missing any negative implication? What does the community here think of this?

$\endgroup$
13
  • 1
    $\begingroup$ What about highly downvoted question and/or answers that are deleted, or that users delete themselves.? Many factors to consider here. If what you ask for happens, we can expect users who received no upvotes to delete posts to misrepresent their action Total rep/Total posts. $\endgroup$
    – amWhy
    Jul 14 at 21:31
  • $\begingroup$ Well, highly downvoted questions and answers getting deleted is a good thing, isn't it? $\endgroup$ Jul 14 at 21:39
  • 1
    $\begingroup$ Think a bit, please. Of course it is a good thing to weed out poor posts. But it is incompatible with getting truly representative ratios of (rep earned or lost from all posts, including deleted posts)/(total number of all posts (current and deleted)). $\endgroup$
    – amWhy
    Jul 14 at 22:14
  • 5
    $\begingroup$ This should be asked, if asked anywhere, on Meta Stack Exchange, because it would require developers to modify the entire network to do that, all websites, not just math.se. $\endgroup$
    – amWhy
    Jul 14 at 22:18
  • 4
    $\begingroup$ This suggestion on Meta Stack Exchange seems a bit similar: New badge suggestion for high average scores on answers. $\endgroup$ Jul 14 at 22:35
  • 1
    $\begingroup$ An older post on this meta which asks how to find average score for one user: Average Q/A score? $\endgroup$ Jul 15 at 4:19
  • $\begingroup$ @AmWhy no I did think about that. Even with deleted answers, I feel it would still be representative of the average quality of work, because no one will want to delete any answer which received one or more upvotes. So, an insightful answer with many upvotes will still increase the score more than a mediocre one. $\endgroup$ Jul 15 at 5:15
  • $\begingroup$ @Martin Sleziak, oh ok, so it has been suggested before. Thanks for sharing. $\endgroup$ Jul 15 at 5:16
  • $\begingroup$ Well, do we do this for average score per question, and separately average score per answer? Yes, in both cases, there are users who chronically ask low quality questions, that get downvoted, often deleted. So a user showing 3 questions with a score of 60, who deleted or had others delete,6 questions, with scores of 0, -3. -1, -7, -2, -4, their actual net rep per post would show up 20, where in fact, it ought to be >-20. Yeah, it makes a difference. Ditto for anwerers who have a habit of answering very low quality question which get deleted, and hence, so do their answers. Don't minimize. $\endgroup$
    – amWhy
    Jul 15 at 17:58
  • $\begingroup$ So Ritam can we close this as a duplicate given by @MartinSleziak $\endgroup$ Jul 16 at 5:35
  • $\begingroup$ @JitendraSingh If you mean this question I would not say they are exactly duplicates - for example, there is no suggestion for a badge. If you mean the other link I posted, that one is on another site - so it cannot serve as a duplicate target. $\endgroup$ Jul 16 at 6:13
  • $\begingroup$ One problem with this idea that hasn't been accounted for yet is that the number of votes a post gets is strongly correlated with the topic. Posts about high school math or undergrad calculus tend to be highly voted. This is probably a result of there simply being more users who understand/are interested in such topics and will read these questions. However this means that the metric you propose does not necessarily reflect the quality of an answer, but to an extent also what topics a user engages with. $\endgroup$
    – Elliot Yu
    Jul 19 at 20:08
  • $\begingroup$ I expressed my thoughts about this sort of reconsideration of rep here: math.meta.stackexchange.com/questions/20212/… $\endgroup$
    – Ron Gordon
    Jul 22 at 19:55

You must log in to answer this question.

Browse other questions tagged .