# Deleting an answer and posting another one

I saw a question about the smallest effect of gravity ever measured, and posted what I believed to be an interesting answer about the forces in an AFM. When I came back a few hours later, I saw that my answer got a single down-vote, and zero up-votes. Further, some other answers had appeared. Based on the comments that had also appeared (to the question and other answers), I realized that the OP was asking for something altogether different, and deleted my answer.

Still, it bothered me that existing questions were not quite getting the point. So the following day I wrote a completely new answer. Not a minor edit of the previously deleted one - it was as though I was answering a different question.

This got me a comment from QMechanic:

In the future, please edit a downvoted answer rather than deleting and adding a second answer. It may be seen as gaming the reputation system to avoid a downvote.

I thought the point of up/downvotes was to separate the wheat from the chaff. I had clearly misinterpreted the intention of the OP, so I deleted the irrelevant answer. Good riddance. Then I set about writing a better one.

I'd like to hear others' opinions: does one need to keep a bad answer around? Should you delete the entire body of your answer and write the new one "in place of the old one"? Or is what I did OK - just rip up the old one and start again.

Do you consider this "gaming the system"? I hope I don't sound like I'm whining - I just wanted to calibrate myself.

For reference, the timeline:

Nov 1, 20:19 - posted original answer
Nov 2,  1:41 - deleted answer
Nov 3,  2:38 - posted new answer


In case you're interested, here is the link to the deleted answer

• I can't see any difference between your current answer and the answer you've apparently deleted. But nevertheless, I don't consider that to be gaming the system, especially when it's a rare occurrence. Gaming the system is claiming a question is a duplicate to prevent other people from answering it. Or only answering hot metas. Or colluding to downvote. And when somebody answers a whole lot of questions IMHO it's usually fairly obvious that they're one of the good guys. – John Duffield Nov 10 '15 at 14:10
• @JohnDuffield: You can't see the deleted answer because one needs >10k reputation for that. – ACuriousMind Nov 10 '15 at 19:14
• Why would someone downvote this? It's a good question. A user was told by a moderator not to do something which was seemingly positive. We definitely should discuss issues like this. – DanielSank Nov 17 '15 at 7:23

I see nothing wrong with deleting an answer and posting another one that is completely unrelated to it. I mean, one can even post multiple answers to the same question if they are unrelated, so deleting one and posting a different approach is not inherently seen as bad by the system.

If you delete an answer and only post a fleshed out version of it anew, I'd say that's questionable, but if your new answer really hasn't much to do with the one you deleted - as I'd argue is the case here - I don't think it should be considered "gaming the system".

• What does "if they are unrelated" mean in regards to posting multiple answers to the same question (line 2)? – sammy gerbil Jan 29 '18 at 12:25
• @sammygerbil: By unrelated I mean that they are truly different answers instead of one being a modification of the other - e.g if one could potentially be an edit of the other then they are not unrelated. – ACuriousMind Jan 29 '18 at 16:04

Generally speaking, there might be cases, where it is difficult to decide whether a new answer should be considered related or unrelated.

(Moreover, it is a general principle on SE that moderators should not evaluate the physics content of a post. This should be done by user votes and user reviews. In particular, it should be up to the down voter to decide whether to retract his downvote if it is no longer relevant for the edited answer.)

Therefore, to keep the system simple and robust, it is better to not distinguish between related or unrelated answers. Which brings us to the main guideline:

Generally speaking, the guideline is, that one should edit an answer rather than deleting and creating a second answer in the same thread.

Note that the chronology plays a role here. It is allowed to post multiple different/unrelated answers. What is frown upon, is, if one first deletes a downvoted answer, only to later post a new answer, be it related or unrelated.

In the present case, I posted a comment below a deleted answer of Floris (which was meant only for Floris to see, but can also be viewed by 10k+ users, if they scroll down). I had not read Floris's answer (among other things because I didn't bother to open his link), and hence did not check on beforehand if it was related or not. I should stress, that I at no point had disciplinary action in mind with my comment. It was purely meant as an advice.

• I wouldn't go that far. Personally (i.e. not speaking as a moderator), I think deleting an answer to post another one that is completely unrelated should be fine. The issue is, of course, distinguishing between related and unrelated answers - but I would think that much of the time the new answer is obviously unrelated, and in those cases I'd favor just allowing it. I would rather not punish what should be (IMO) a perfectly valid strategy for retracting a bad answer, just because it can be misused to skirt downvotes. – David Z Nov 10 '15 at 5:13
• Thanks @DavidZ. Out of curiosity (and since you said "I would rather not punish") - how would a moderator "punish" in a case that merits it? – Floris Nov 10 '15 at 11:31
• The above meta answer is meant as a guidance to good net etiquette. I should stress that breaking the above guideline does not automatically lead to a suspension. In practice, there would be no suspension for a single incident that stands alone. There has to be repeated pattern or combined with other issues to lead to a suspension. Finally, let me add that a suspension is in practice only issued if all moderators agree with it. – Qmechanic Nov 10 '15 at 11:36
• @Floris you mean a case where someone deletes a downvoted answer and then posts another one with roughly the same content? Off the top of my head, probably delete the new answer and leave a comment advising the poster that they shouldn't do that, and should edit+undelete their original answer instead if they want to keep the content around. – David Z Nov 10 '15 at 11:53
• OK thanks both of you - I think this answer and the associated comments clarify the issue. – Floris Nov 10 '15 at 13:32
• I strongly disagree with this answer. If I post a bad answer that is voted down, those votes refer to that answer. If I then learn what's wrong with my answer and write a better one, that new answer should not acquire the downvotes of the old answer. The score of an answer is supposed to reflect its quality, not the quality of previous answers written by the same person. I find it very worrying that a moderator is talking about banning people for deleting bad answers and posting good ones. – David Richerby Nov 15 '15 at 11:42
• @DavidRicherby Re: "The score of an answer is supposed to reflect its quality, not the quality of previous answers written by the same person." That's a great point and I mostly agree. However, votes are connected with reputation which is precisely a measure of the quality of previous answers written by a person. This connection between an answer's score and its author's reputation add some complexity to the issue, don't you think? – DanielSank Nov 17 '15 at 7:29
• @DanielSank People being more likely to up-vote an answer by a high-rep user is a second-order effect and it doesn't alter what the score of an answer is supposed to reflect. To put it another way, why should somebody's answer to a question start at, say, -3 instead of 0, just because their previous answer to the same question was bad? – David Richerby Nov 17 '15 at 8:10
• @DavidRicherby Oh, I agree with that, for sure. The issue arises when folks try to game the system via deletion. Actually, perhaps we should not use that as an argument in any direction without data showing that such gaming exists and is a problem. – DanielSank Nov 17 '15 at 8:18
• @DanielSank If somebody deletes a downvoted answer and reposts something substantially similar, that's abuse in most cases. But I think it's quite clear I wasn't talking about that case. – David Richerby Nov 17 '15 at 8:23
• @DavidRicherby - Up until your last comment, you and I are in agreement. "If somebody deletes a downvoted answer and reposts something substantially similar, that's abuse in most cases." The new answer will stand on it's own. If it's still a bad answer, it will be downvoted again. If it's now a good answer, even if "similar" it may get upvotes. Also a (bad) answer that was downvoted, then edited (now good), and subsequently deleted and reposted, may appear to be deleting and reposting near identical content. ... – Kevin Fegan Nov 21 '15 at 18:55
• ... You would have to carefully review the edit history (and possibly have a good understanding of the differing content) to fairly judge that, and I don't think that detailed of a review would happen that often. The bottom line is, if what might be considered "possible abuse" results in better answers. That's a good thing. Otherwise, what incentive does a user have to improve their answer if their then "good" answer will (as you mentioned) start at "-3", and their alternative is only to delete the bad answer and move on without contributing anything to the question. – Kevin Fegan Nov 21 '15 at 19:04
• "Note that the chronology plays a role here." OK, so lets say, that someone writes a "bad" answer. It could be just plain "awful", or it could have been based on a misconception, or a previous version of an edited question, and it then (rightfully) gets downvotes. Then that user posts a new (different) answer that appears to be good (or is at least no longer "bad"), and it may even get some upvotes. The user has 2 (different) answers to that question. Then, sometime later (say 6 months) they delete the "bad" answer. Is this OK? In practice, how is it different from repost/immediate delete? – Kevin Fegan Nov 21 '15 at 19:22
• @KevinFegan Just to be clear, when I talk about "deleting an answer and reposting a substantially similar one", I mean reposting one that essentially fixes none of what was wrong with the original. I'd have no objections if, for example, somebody posted a generally good answer that got downvoted because a small part of it was really bad, and then deleted and reposted with the bad bit fixed or removed. Sure, most of the new answer would be identical to the old one but the reason for the downvotes has been fixed. – David Richerby Nov 21 '15 at 19:34
• @DavidRicherby - We are still in basic agreement. But consider that if the reposted answer was still bad (particularly if in the same way as the deleted answer), then it will likely get downvoted as well so there is no incentive for anyone to do this. I would say that the practice should (always) be allowed and just let the users upvote and downvote the good/bad answers... unless someone does this with nearly all of their answers. – Kevin Fegan Nov 21 '15 at 19:50