19

TL;DR I'm proposing the following:

When someone posts an answer and the corresponding question is closed within a certain amount of time after that, then, for certain close reasons, all rep gained from that answer should be reversed.


First off, there are two questions here on Meta.SE that already deal with "should closed questions award no rep". But:

  1. This question is not a dupe, as I am only interested in what happens to answers.
  2. Both questions are more than 3 years old, and I think times have changed since then.

Now, I assume most of us know the FGITW problem, which for bad questions turns into "repwhoring" - for those who don't here's a situation I run into frequently on Stack Overflow with and questions:

  1. A bad question is asked (typo, off-site recommendation, opinion, ...).
  2. I see it, vote to close, and maybe leave a comment, explaining why this type of question is unwelcome on that site, or for typo questions, point out the typo.
  3. A bunch of bad answers appear, usually low-quality (code-only, link-only or very little text), for typo questions just pointing out the typo (basically the same as my comment), or sometimes even asking for clarification/not even answering the question.

The answers that don't help the asker usually get downvoted, but those that do technically help (in pointing out the typo, or providing a link) are upvoted.

According to the Meta consensus, this is bad, and such answers should be downvoted.

To some point I agree with that, but I think this is also bad, because

  • The information they provide is technically correct, so downvoting seems inappropriate.
  • You lose rep when downvoting them, which is going to stop a lot of people.
  • They get 5x the amount of rep from a single upvote that they lose from a downvote, so unless the down:up ratio is 5:1 or higher, they still make rep overall.

Now, why are those answers bad in the first place, if they are technically correct?

  1. Because they encourage questions that get closed, i.e. questions that are not ok. This encouragement is bad, so in order to discourage bad questions being asked, we should also discourage the answering of bad questions.
  2. Because when voting to close the question, you are trying to enforce a quality policy. People who ignore that policy get rewarded. That feels wrong.

In the two questions that deal with rep from questions, the biggest issue is of course:

Duplicates

If the question has already been asked, that does not mean the it is bad or off-topic, only that the asker didn't do enough research.
I therefore think that the "duplicate" close reason should be excluded from rep reversal.
The same holds true for questions that are migrated away - they're not bad, they just belong elsewhere.

Now a few other potential problems with this:

Taken from Sampson's answer on the first linked question:

Reputation in general is a reflection on your status with the community. The events that take place prior to the closing of a thread reflect your participation in the community, and is therefore not subject to change because somebody else broke the rules[.]

Absolutely true, only that by making someone feel like it's ok to break the rules, you harm the site. Closing doesn't change this, it only confirms it.

And from Robert Harvey's answer:

Closure is the state in which we put questions that need improvement. Closing a question prevents further answers, but if someone posts an answer that is good enough to attract upvotes, I don't see why we should penalize them for that.

Because it still encourages bad questions. Which is bad.
And nobody should be penalised, just not rewarded for it.

Also from the same answer:

If you don't want users to gain rep from bad questions, be diligent about casting close votes, so that the question gets closed before anyone can post an answer to it.

That is one of the things we're trying to do in the SOCVR chatroom, however:

Then, another problem is the time span between the posting of an answer and the question being closed.
Currently, deleting a post older than 60 days does not reverse rep in most cases - I think the same should apply if closing caused rep reversal, but with a shorter window, I propose 3 days.

Then of course, if a question is reopened, that rep should probably get re-awarded.
Now wait - that might drive people to vote to reopen bad/off-topic questions, right?
Totally true, but

  1. The number of people who answer bad questions and have enough rep to vote to close/reopen a question is not that high - usually those answers come from low-rep users.
  2. If that happens, you can handle it the same way as any other privilege abuse:
    Flag for mod attention.

So, my full request (with request for comments) is:

When someone posts an answer and the corresponding question is closed within X days after that, and if the question has not been closed as a duplicate or migrated away, then all rep gained from that answer should be reversed.

For "X" I suggest 3, but that is just based on my gut feeling.

  • Looks like it's more about battling rep-whoring over off-topic questions than FGITW though. – Mathieu Guindon Nov 17 '15 at 22:56
  • @Mat'sMug True, edited that in. – Siguza Nov 17 '15 at 22:57
  • 6
    Am I the only one who first parsed "combat repwhoring" as a noun phrase? The mental image is... interesting. – Ilmari Karonen Nov 18 '15 at 0:37
  • 1
    I think "reversed" is a slightly poor choice of wording; you mean "reset back as though it had not occurred", and for that reversed is not exactly wrong, but when referring to numeric changes, reversal sounds more like flipping the sign: subtracting positive values instead of adding them and adding negative instead of subtracting. That, of course, is not what's intended at all. – Nathan Tuggy Nov 18 '15 at 0:44
  • 2
    Ps. +1, but even with the TL;DR summary at the top, this proposal is pretty hard to read. Some trimming and copyediting might be useful. – Ilmari Karonen Nov 18 '15 at 2:57
  • 4
    @NathanTuggy That's how the voting system calls it. But yes, I mean "undone". – Siguza Nov 18 '15 at 16:38

You must log in to answer this question.

Browse other questions tagged .