# Deletion of on-topic answers only because they contain mathematical errors.

Are mathematical errors in an answer that is otherwise reasonable and on-topic ever a sufficient reason for users (other than the answer-author) to delete that answer?

For a recent answer, the author might still delete or repair it on their own or in response to comments, given enough time to realize the error and consider a correction. In this case it seems better that only the author should delete an answer, no matter how wrong the contents.

For old answers that are severely wrong, I expect there will be some debate but personally I do not see any additional sufficient reason. Maybe there could be some exceptional situations where the Q and A are from deleted users, the answer is unique or accepted, and other questions are closed as duplicates of the wrongly-answered question. For the generic case, the ability to downvote, comment and to post other answers seems (to me) a sufficient quality control mechanism.

Addendum. From earlier discussions it appears that

• moderators do not delete flagged answers only for being incorrect. 1, 2 , 3, 4, 5, 6, 7

• the Low Quality post review queues were never intended for answer deletions based (only) on the content being incorrect. This was the understanding from StackOverflow/Stackexchange and there has always been the same consensus on this meta. a, b, c, d, e

• Not the same question, but to some extent related: Is sometimes keeping wrong answer reasonable?. (Notice that the older question asks specifically about self-deletion by the answerer.) – Martin Sleziak May 25 '16 at 6:14
• I do not go around the site searching for answers I could vote to delete, however if some post I consider of negative value to the site shows up in the LQ queue I might vote "delete" and not "Looks OK." Do you recommend to vote "Looks OK" on very flawed material extremely unlikely ever to be fixed? – quid May 25 '16 at 7:18
• I would not mind discussing this topic further, but I have trouble understanding what changes/refinements you are suggesting to the current policy (if we have any). Are you saying that users should never vote to delete incorrect answers, if there are no other problems with them? Sounds a bit too sweeping to me. I do think that you have a more limited scope in mind. Please elaborate! – Jyrki Lahtonen May 26 '16 at 19:48
• IMHO we are handling incorrect answers adequately, when we base any action on the overall usefulness of the post. Answers containing errors can still be useful, but may also be totally useless. It seems to me that case-by-case judgement is called for. Blanket rules rarely work well as a general policy - we have seen occasions of that. – Jyrki Lahtonen May 26 '16 at 19:51
• Isn't erraneous answers what downvote is for? – mathreadler May 27 '16 at 20:49
• Wrong answers are not useless, and usefulness is a subjective standard that is best left to broad-based voting. Given that completely wrong material is celebrated as virtuous "work and effort" when it comes from an OP in a question, and is alleged to illuminate all sorts of things about the question, I do not understand why the same respect should not be shown to work and effort of answerers, who after all are the core of the site. @JyrkiLahtonen – zyx May 28 '16 at 18:39
• If "broad-based voting" qualifies an answer as useful it cannot be deleted by users' votes. – quid Jun 1 '16 at 19:39
• A recent discussion related to this: Should we vote to delete wrong answers? – Martin Sleziak Jul 3 '18 at 13:49

From the help-center (my emphasis):

Answers that do not fundamentally answer the question may be removed. This includes answers that are:

• commentary on the question or other answers
• asking another, different question
• “thanks!” or “me too!” responses
• exact duplicates of other answers
• barely more than a link to an external site
• not even a partial answer to the actual question

If the mathematical errors are so severe that the resulting answer is "not even a partial answer to the actual question" then "it may be removed."

I do not propose to delete incorrect answers as a matter of course, and almost always a vote to delete should not be a first step when dealing with such an answer. But, I do not see the problem with deleting the occasional utterly wrong and not useful answer where OP shows no intention to fix them.

• In the question, the words "otherwise reasonable and on-topic" can be taken to mean "if the errors could somehow be fixed, the answer would be a valid response to the question". (This is true even when we know it is logically impossible to fix particular errors, such as purported answers to questions requesting a proof of a known false theorem.). I don't see much overlap between that and the not-even-partial answer category from the Help Center. It seems the latter is mainly aimed at answers that are not even ostensibly on topic. – zyx May 25 '16 at 21:47
• The development of the general policy (in the sense of official SE guidelines) is in the direction opposite from the one you claim. It is relevant to recall that initially the option of voting to delete others' answers just did not exist for users. Even if it exists since about five years, the shift away from up/down only (except for spam, offensive) trickles down only very slowly. Thus, I thank you for having raised this issue, as it will allow to make this wider known. – quid May 25 '16 at 22:15
• Is it "opposite to what I claim" ? Distributing a capability to more people (usually) goes hand in hand with a dilution of the ability and restrictions on its use by the additional less skilled users. I don't know the early history of StackOverflow, but generally the evolution in matters of "community moderation" is from a few wise and experienced users granted total moderation power (say arbitrary deletion), to giving out lesser forms of that ability to more sub-moderators as the system grows, while also setting increasingly formal rules and limits on how the power can be used. – zyx May 25 '16 at 22:24
• Yes, it is opposite in that in earlier days the policies regarding answer deletion were more restrained than they are now. When only moderators were able to delete answers (excluding punitive flags) it was not feasible to have a policy of deleting (very) incorrect material, because the policy must not put moderators in a position where they are forced to make decisions on correctness, as they may lack the competence in the specific subject matter and for general workload. Abstractly mods do not judge correctness, because they cannot, and not because we wouldn't want them to. – quid May 25 '16 at 23:36
• No. The moderators have always been capable of judging material within their areas of expertise, which (in terms of number of postings) would have covered a large slice of the very-incorrect material. Nevertheless the idea from both the moderators and the users was that it is undesirable that this should be grounds for deletion of answers. Apparently all these people did not receive your memo that "we would want them to" do so. (By the way, who is this "we" that you speak for?). – zyx May 26 '16 at 6:37
• Incidentally, I think the much more commonly argued reason for moderators to avoid correctness judgements (even when expert and when they had the time to handle a particular post) was that those would effectively be handled by the "community moderation" using up/down votes. – zyx May 26 '16 at 7:14
• It appears you missed the reason "general workload." For my statement on reasons to be true it is not necessary that "we would want them to." As I am talking about general guidelines the "we" refers to an abstract notion of "an SE community" not this (or any) particular one; same for "mods." – quid May 26 '16 at 7:24
• I answered already the comments on "moderator workload". To repeat: even in cases where the moderators found themselves with time to delete some postings based on correctness judgements (and using the binding or permanent moderator votes) the expressed idea on the meta was that it is desirable to avoid that. There was additional debate on whether moderators should use the super-vote powers at all where it can be done by ordinary user votes, but not much support for the idea of moderators cleaning up wrong mathematics from the site. You obscure this with talk of requiring them to do that. – zyx Jun 1 '16 at 21:19
• @zyx this misses the point in that it puts aside the need or at least the desire for some consistency in moderator actions. Anyway, policies evolve over time. This thread is now the current one. My view is perfectly in line with the one expressed in this thread by a moderator (Jyrki). Thanks again for starting this thread and facilitating this overdue update on guidelines. – quid Jun 1 '16 at 21:41
• @quid Since you have been here only a few years, you missed the first few formative years of the site, If you browse this history on meta you will learn that we chose to do many things differently than the SE generic guidelines. – Bill Dubuque Jun 4 '16 at 14:13
• @BillDubuque not really relevant, as I said: guidelines evolve. (It's also not quite true, but it is irrelevant.) – quid Jun 4 '16 at 23:05
• @quid Truth is always relevant. – Bill Dubuque Jun 4 '16 at 23:39
• @BillDubuque I should maybe clarify that "not quite true" refers to my ignorance on the matter, not proclaimed difference of earlier guidelines from those of SE. It is true those existed and continue to exist. – quid Jun 4 '16 at 23:43

Why don't you edit the answer? I don't get it. StackExchange sites have an edit button under every question and answer for exactly that reason.

You may consider deleting the answer if it is irreparably flawed, but most of the time, an edit would be sufficient, right?

• I think the OP is not asking whether he should delete such answers. He already believes he shouldn't. Instead he is asking for ammunition against other people who delete answers. And trying to generate propaganda to convince other people not to delete answers. – GEdgar Jun 8 '16 at 15:05
• Edits aren't supposed to fundamentally change the answer. If the OP said, "the answer is A" you can't just change it to say "the answer is B". That's a fundamental change to the answer. Editing is more intended for making improvements to things like phrasing and formatting or possibly adding more detail that is consistent with the OP's intent. – AJ Henderson Sep 6 '17 at 16:30

Perhaps a special "Flawed" flag should be developed that adds a message template with words to the effect:

This answer contains valid and/or helpful content relating to the question but is flawed in the mathematical details. Readers should proceed with caution in any use of this answer.

• How would the mechanics of this go? $n$ users have to flag before this message shows up? People can just downvote instead and express their displeasure in comments if need be. – J. M. is a poor mathematician Jun 4 '16 at 15:10
• Based on the concerns of the OP for this meta question, which is what my answer addresses, it appears that the existing downvote system rates a question in an all-or-nothing way that sometimes throws the baby out with the bathwater or conversely forces the keeping of a really good but technically flawed answer that is a disservice to readers. I imagine the best way to implement this idea is similar to how we implement off-topic or duplicate flags. The exact details probably will require a community decision. – O.M.Y. Jun 4 '16 at 16:17
• We have a pretty good mechanism for pointing out flaws in an answer already, namely to comment on the answer pointing out the flaws, and downvoting an answer that (due to flaws or for other reasons) are not considered to be helpful answers. – Henning Makholm Jun 6 '16 at 18:29