# Downvoting to order?

I have stumbled with the following in a user profile.

I use downvotes to organize answers to a question. By voting up I bring up the answers that I consider more important, useful, or that I think should be read first. By voting down I also do the same. If I downvoted your answer, it might not be that I think it is bad, but that I think other answers should be on top of yours in the list.

I strongly disagree with this idea. What's more: I think it is inappropriate to "order" answers in any way. Of course, upvotes and downvotes surely order answers, but downvoting just to "order" a list of answer seems simply wrong. The order should be the consequence, not the cause to upvote or downvote. In any case, one should upvote the question one deems more useful than the other, but if an answer is less useful than another, that doesn't entitle to downvote it. Do you agree with the quoted phrase? Why? Why not?

If the above wasn't clear, Pete has read my mind:

The question is whether downvoting answers just because one finds other answers to be better is acceptable behavior on the site. If I understand and remember correctly, users have previously been suspended for such behavior, so at least in the past this behavior has been viewed as unacceptable. In case we are revisiting this issue again, I would say that it should not be acceptable to downvote answers for any reason other than a sincere belief that the answer is (in some way, at least) incorrect.

• So: user A prefers to order answers with votes up/down; user B prefers not to. There is no disagreement here. – 40 votes Jul 30 '13 at 3:47
• @40votes That is not what I am saying. I am saying a user downvotes an answer which is useful because he/she thinks it should be below another answer he/she deems more useful. – Pedro Tamaroff Jul 30 '13 at 3:48
• So, the user does a certain thing, which you would rather not do. That's still not a disagreement. A disagreement would arise if that user told you how to vote, and you did share his/her opinion. Fortunately, that isn't the case. – 40 votes Jul 30 '13 at 3:51
• @40 votes: The question is whether downvoting answers just because one finds other answers to be better is acceptable behavior on the site. If I understand and remember correctly, users have previously been suspended for such behavior, so at least in the past this behavior has been viewed as unacceptable. In case we are revisiting this issue again, I would say that it should not be acceptable to downvote answers for any reason other than a sincere belief that the answer is (in some way, at least) incorrect. – Pete L. Clark Jul 30 '13 at 3:53
• My attitude toward voting philosophies is that they are inherently subjective. I've yet to see a particularly constructive conversation about the subject on the SE network. I think if the voter thinks the post should be downvoted - whatever that means to them - they should downvote. If you think the answer score represents some underlying truth that is somehow contaminated by the behavior you've described above, you're taking this site way too seriously. Plenty of terrific answers have low scores and just as many trivial answers have several upvotes. – Macro Jul 30 '13 at 4:08
• @Macro I take what I do seriously, and thus what I post in the site. Again, I am not worried about me getting downvoted, but about the general idea of downvoting to order according to merits of other answers. I agree with all other things you say, though. However, as it has been pointed out "...downvoting answers just because one finds other answers to be better is acceptable behavior on the site... (...) ...has been viewed as unacceptable." And I agree with this. Don't you? – Pedro Tamaroff Jul 30 '13 at 4:11
• Well, at least we've found the source of the inexplicable downvotes – Emily Jul 30 '13 at 4:45
• @Arkamis The downvotes were not inexplicable, and I wasn't hiding. Actually in my profile it was explained my voting policy for all to see. It is a meditated act, I believe it to be beneficial, I have no interest in hiding it. Quite the opposite. – OR. Jul 30 '13 at 5:10
• The downvotes are functionally inexplicable, as the downvoter's identity is not revealed unless one does a brute force search through the user database. Furthermore, numerous opportunities availed themselves as to why downvotes occurred. I cannot think of a single instance where this explanation was proffered in the comments to a relevant post. – Emily Jul 30 '13 at 5:13
• They are functionally inexplicable, and yet the FAQ tries to gives them an explanation that goes beyond their actual effect, organizing answers to a question. This reminds me the greeks defining point as adimensional entities and what not ..., and then mathematics making the transcendental step of defining them as whatever behaves as a point does. The power of defining something just as it behaves have yet to get to the voting in this website. – OR. Jul 30 '13 at 5:36
• I do not do this, but I can see why someone would. Say user A has by far the best answer but they are a few hours late to the party. All the other answers have +10. User A is never going to catch up, no matter how good their answer is. Downvoting the other answers will mean the gap sits at 8 as opposed to 9. It takes 5 people to get a balance as opposed to 10. – user1729 Jul 30 '13 at 11:06
• @RGB downvoting correct answers that other people might consider useful is just plain trolling. In addition, it is very contraproductive since it might lead the OPs of the correct but downvoted answers think that there is something wrong with their posts, they did something wrong, etc ... I know that many people on MSO like to sort answers by downvotes too, and I personally consider such a behavior as very reprehensible. It undermines the good collegial and helpful atmosphere where people can trust each other on any site. – Dilaton Jul 30 '13 at 22:16
• @Dilaton: in my opinion the tone of your comment undermines the good collegial and helpful atmosphere quite a bit more than a downvote here or there. – quid Jul 30 '13 at 23:23
• @quid downvotes are like a slap in the face. So in my opinion they should certainly not be dished out to people who have basically done nothing wrong but posted a correct, legitimate answer (partial answers are legitimate too...). As others pointed out, sorted by vote is only one among other possible views available. If one wants to contribute to sort answers by votes, this can for example be achieved by downvoting bad/wrong things, leaving standard correct answers alone, and upvoting stellar and particularly interesting posts. I see absolutely no need to downvote correct (partial) answers. – Dilaton Jul 31 '13 at 14:41
• @quid hidden behind the accounts posting are still real people who put more or less effort into trying to help the OP of the question or contribute something of potential value for everybody who reads the question. In my opinion they can rightly so expect to not get downvoted for posting correct answers. Do you not agree that my alternative voting scheme, which avoids downvoting correct answers, leads to the effect of a sorted list of answers too, but it does not have the negarive side effect of giving OP's of correct answers the impression that there is something wrong with their – Dilaton Jul 31 '13 at 15:49

Let me present a different voice here (from the other moderators).

If you hover your mouse over the up/down-vote arrows for an answer, the caption that pops up is "This answer is useful" versus "This answer is not useful". And considering that we are a Q+A website, I am uncomfortable with a strict correctness-based approach to voting.

Let me expand a little bit.

While mathematics is indeed one of the few subjects lucky enough to have universal truths (Yes, Asaf and other logicians/set theorists, I know what you do but I'm explaining at a lay level here), the doing of mathematics is still a human endeavor. As a professional mathematician, when I read a proof I don't just think about whether it is logically sound, I also try to figure out its thematic purpose (how I can appropriate the technique and train-of-thought that led to the proof for other uses), and when doing so, I am very appreciative of authors who explain things clearly.

I feel that the same should be (allowed to be) extended, to a certain extent, to voting on Mathematics StackExchange. After all, Math.SE is a human activity. And one of the great things about Q+A websites is precisely the intangibles provided by the human element. Given the same question, one may give a machine checkable answer, or one may give a human readable sketch that skips some minor steps and outlines only the important ideas. The latter, while incomplete, I think would be generally more useful to readers than the first, strictly correct answer. And if a user chooses to use votes to voice his opinion of this fact, I don't see why it should be a problem.

(This is to say nothing of the inherent tax built into the system for downvotes: a downvote takes 1 reputation from the voter. If anything it is more self-regulating than the blatant inflation of reputation driven by the free upvotes.)

Lastly, let me repeat what I often say:

Voting (and usefulness) is subjective. And a user is never obligated to explain publicly their votes.

(Though I generally prefer it when people leave constructive criticism in the comments.) If a user has a really high bar for what is useful to him and what is not, and if he applies this consistently across the board (as opposed to targeted voting), I don't really see how it is any of our business how he chooses to use his votes, even if he explains rather poorly his motivation for doing so.

You'd note that I haven't really addressed the question yet.

My personal opinion of the use of votes for ordering is: why? Isn't that basically futile? A user is free to choose on which tab he reads the answers, and I, for example, read SE sites on "active" instead of "votes". And considering that this is a one user effort I really wonder more at the efficacy of this voting method than anything else.

• "And considering that we are a Q+A website, I am uncomfortable with a strict correctness-based approach to voting." May I ask you to elaborate? I guess I don't understand what it is that you're considering. About proofs that are technically correct but more suited to a machine than a human being: have you seen such proofs? I don't think I have. (I've seen proofs written in such human-unfriendly ways that I don't care to read them. But if I did read it and decide that it was correct, it would seem that I have some information that it would be beneficial to pass along by upvoting.) – Pete L. Clark Jul 30 '13 at 6:15
• By the way, I am not advocating a completely strict correctness-based approach to voting. I think you shouldn't downvote an answer that you perceive as correct. I also think that you should upvote every answer that adds value, and I think that at least in most cases the first correct answer adds value. When I do see an answer that I find correct but missing something that I would really like to see, I try to leave my own answer to the question (and whenever I do that I feel it is only good sportsmanship to upvote earlier correct answers). – Pete L. Clark Jul 30 '13 at 6:18
• Let me also say that I am not specifically advocating for users to get suspended; I was just recalling what has happened in the past. I agree that if someone quietly downvotes without targeting anyone as determined by the software, we have to give them the benefit of the doubt. Let me also say that in the one instance of someone suspended for downvoting that I remember, the user had a habit of downvoting other answers to a question that he had answered. That's a different and more problematic form of behavior, in my opinion. – Pete L. Clark Jul 30 '13 at 6:26
• @PeteL.Clark The several paragraphs after the sentence you quoted is the elaboration. I even wrote "Let me expand a little bit." The machine-readable example is hypothetical; I don't know of any answers written in Coq or any other proof assistant. But here you already illustrate my point: you would upvote a Coq proof on MSE (provided you verify it yourself), while I won't. What I find useful (for myself and for other human readers) is different from what you do, and I respect that difference. – Willie Wong Jul 30 '13 at 6:32
• In regards to this comment: I was wondering whether that was what you were thinking of. (I was not sure since in my mind that behaviour is quite different from the one under discussion.) – Willie Wong Jul 30 '13 at 6:36
• okay, thanks. I did read your whole answer, and I still didn't understand it completely, hence my question. Given that there have already been many thousands of answers on this site, stating policy in terms of totally hypothetical answers seems unnecessarily...well, hypothetical. But if we're going there: I would never read a proof written in any formal language.... – Pete L. Clark Jul 30 '13 at 6:36
• Also, I didn't suggest to other people that they upvote every correct answer: I can do that myself because I almost never read enough to run out of upvotes. But I think downvoting a correct answer written in Coq, say, would be a poor thing to do. Can you think of a good argument for it, or is your position entirely based on personal liberties? – Pete L. Clark Jul 30 '13 at 6:38
• @PeteL.Clark: my argument is two-fold. (a) Most readers of this site will not be able to make use of (or even understand) a proof written in Coq. If a user posts a proof in Coq with no embellishments and no explanations (and not even a caveat that the proof itself is written in the Coq language), I think the presence of that answer would be more confusing to readers than helpful. (b) In terms of style, there's a reason that we don't write formally verifiable proofs day in and day out. We'd be so bogged down in details that we cannot clearly demonstrate the important ideas behind the proof. – Willie Wong Jul 30 '13 at 6:43
• A slightly different example would be the difference between a 150 line derivation of an algebraic identity by brute force, and a two line combinatorial proof. On the other extreme, one can make a mathematically correct statement while giving too little details. If a question is asked about why two definite integrals are equal to each other, to just say "they are both equal to the hypergeometric function ${}_1F_2$ evaluated at so-and-so" is not incorrect, but hardly useful. In short, I feel that a useful answer on MSE should explain why and not just how. – Willie Wong Jul 30 '13 at 6:46
• Hmm, I don't find enough to disagree with in your last comments to continue the discussion, un/fortunately. :) (In particular, leaving some allusion which will convince the cognoscenti that you know how to answer the question certainly need not constitute a "correct answer".) – Pete L. Clark Jul 30 '13 at 6:51
• Also, I have certainly downvoted answers in which everything said there is correct, but what is missing is more essential than what is present. Maybe this is a (slightly condescending but) helpful way to say it: if you would give full credit to an answer when grading an exam with that answer given, you should not downvote the answer, IMO. – Pete L. Clark Jul 30 '13 at 6:56

This (downvoting answers you know to be correct because you think others are superior) is not an acceptable use of downvotes. The official guidelines for downvoting are located here in the privileges section of the FAQ. There is however one exception, and that is community wiki questions (as well as of meta posts of course). In fact, one of the main reasons questions are made CW is to allow users to use voting as an ordering.

Edit: Here is the text of the "When should I vote down?" section of the FAQ page:

Use your downvotes whenever you encounter an egregiously sloppy, no-effort-expended post, or an answer that is clearly and perhaps dangerously incorrect.

You have a limited number of votes per day, and answer down-votes cost you a tiny bit of reputation on top of that; use them wisely.

• In the link it says downvotes are used to indicate what answers are least useful. This is open to interpretation. One could argue in favor of many practices by claiming they are in service to this, including the behavior described above. – Macro Jul 30 '13 at 4:12
• @Macro It says "least useful". The relevant section is "When should I vote down?" which I will edit into my question. – Alex Becker Jul 30 '13 at 4:13
• The quoted text above is still inherently subjective. Clearly this is true given that not all users that encounter a given question will vote the same way. You're chasing your tail by trying to precisely define something inherently subjective. Good luck trying anyway. – Macro Jul 30 '13 at 4:19
• @Macro Is an answer that is completely correct, but not as in-depth as another answer, an "egregiously sloppy, no-effort-expended post, or an answer that is clearly and perhaps dangerously incorrect"? Clearly not. I'm not trying to define precisely how one votes, but I am trying to demarcate the boundaries. That is part of my job as a moderator. – Alex Becker Jul 30 '13 at 4:21
• @Macro You're trying to find the fifth foot of the cat, really. – Pedro Tamaroff Jul 30 '13 at 4:21
• @Macro no $\neq$ not enough – Alex Becker Jul 30 '13 at 4:35
• @Macro: thinking that not enough effort is expended is clearly not the same as "no effort expended". To me it also seems completely unreasonable to evaluate a correct answer as "not enough effort expended". Your comments amount to a defense of the idea that it is acceptable to downvote an answer because one regards other answers as superior. If the history of this site is any guide, this manner of "interpreting the link differently" could get you suspended in the future, so it is probably worth clarifying now. – Pete L. Clark Jul 30 '13 at 4:39
• I completely agree with what Alex said here. I have been exasperated with poorly thought-out "hints" that were not tried out and turned out to be, not even wrong, but dangerously useless. I would link to one specific example from this past weekend, but as I do not believe in public shaming, I will refrain. I will say that such an "answer" is a textbook example of something deserving of a downvote, and I surely supplied it. – Ron Gordon Jul 30 '13 at 4:55
• @Ron: I agree that bad hints can be more harmful than bad answers. Unfortunately I have found that it is hard to give good hints on a site like this: a truly good hint needs to be better thought out than a complete answer, and in the amount of time it takes to craft a good hint, someone else will almost certainly leave a complete answer, which the OP will usually like better. I think that we would need to somehow make hints part of the site mechanics to really fix this phenomenon. – Pete L. Clark Jul 30 '13 at 5:07
• @Ron: To be sure, I entirely support your method of answering questions. I'm just saying that there are some forces acting against it. When I answer questions that could be homework questions in undergraduate courses, I try to leave something for the OP to do, even if it's a quite straightforward connection of dots that I've laid out and numbered. But I very often find that within a few minutes on either side of leaving such an answer, someone else will leave an answer which precisely connects those dots. And the site mechanics encourage this... – Pete L. Clark Jul 30 '13 at 5:27
• @AndresCaicedo I find it unlikely that we would even be able to tell that a user was downvoting in this fashion. However, if a user repeatedly claimed to downvote posts which they thought were correct and helpful but note the most helpful, we would probably at least consider taking action. – Alex Becker Jul 30 '13 at 5:33
• @WillieWong Yours is a very reasonable position, thanks. – Andrés E. Caicedo Jul 30 '13 at 6:15
• (To be very precise 'this site' refers to the main site, not the meta site.) – quid Jul 30 '13 at 12:37
• @PeteL, please accept my apology if you found my lazy attempt at a joke disrespectful. That was not intended. Re: the discussion, statements like the one you concede - "if you feel that an answer leaves out key steps, then you do not sincerely believe that it is correct" - imply a flexibility that, I think, makes this thread pointless. Under that mentality, if there is a superior answer present, any answer less than it may be considered "incorrect". Therefore there exists a good-faith voting strategy that is equivalent to the one described here, so there's really nothing to be done about it. – Macro Jul 30 '13 at 14:29
• @quid: I consider yours by far the most natural reading of the statement in the FAQ. (I also ignore it: I don’t ever downvote on the main site, preferring to leave comments.) – Brian M. Scott Jul 30 '13 at 16:43

After I wrote the comment that Peter has quoted in his question, I rethought it a little bit, and now I think I said something subtly stronger than I meant to. Instead of

[I]t should not be acceptable to downvote answers for any reason other than a sincere belief that the answer is (in some way, at least) incorrect.

let me backpedal a bit to the following stance:

It should not be acceptable to downvote answers that one sincerely believes are correct.

The difference concerns answers that don't say anything incorrect but do not answer the question either. This is a gray area where I think the decision whether to downvote depends a lot on the context and in particular, the potential voter's evaluation of the amount of good faith exhibited by the answerer.

(Let me also reiterate that my personal policy is that one generally should downvote answers that one believes are incorrect. I don't hold that this should be a site policy -- anyway it would be impossible to enforce since who knows who is reading answers and evaluating them as incorrect? -- and other regular users of the site have disagreed with it. But I will say that if everyone followed this policy as well as the one advocated above then the meaning of downvotes would be completely clear, which would be a good thing. Every once in a while one receives a downvote for no apparent reason, and this is a bit frustrating.)

• Indeed, I agree with this new phrasing. Each user can do what they want, but they shouldn't generally downvote correct answers. On the other hand, it remains that the site would run better, and this sort of bad down voting wouldn't affect things so much, if people voted more in general too. Vote early, vote often, vote right. – davidlowryduda Jul 30 '13 at 4:21
• @mixedmath I would buy a pin with that last catchphrase. You'd need to add a smiling picture of yourself. – Pedro Tamaroff Jul 30 '13 at 4:23
• I agree completely with both Pete and @mixedmath. What I view as unacceptable behavior is exactly what Pete says, downvoting "answers that one sincerely believes are correct". – Alex Becker Jul 30 '13 at 4:23
• I would pedal even further. My stance is "It should not be acceptable to downvote answers that one sincerely believed to be useful." But that is almost tautologically impossible to be on the wrong side of. – Willie Wong Jul 30 '13 at 6:09
• @Willie: well, I can see and respect that that's your position, although I'm still trying to understand it. Although "correct" and "useful" are not a priori synonymous, I think at least 99% of the time they coincide. Since the former is objective and the latter is completely subjective, I will stick with "correctness-based voting", at least for now... – Pete L. Clark Jul 30 '13 at 6:31
• @WillieWong: to me the main question here is, and perhaps I will elaborate in an answer, is the usefulness of one answer to be decided solely on the merits of this one answer pretending all others are not present, or is it to be decided against the context of the (potentially) other answers that exist too. Because if it is the latter then, well, answers providing only information also present in others in a better form are not useful given the context. (This is more an abstract argument, to justify the pattern, I do not personally dv like this.) – quid Jul 30 '13 at 12:20
• @quid: On reflection, the question you posed is a much better one to think about than the one I answered (so yes, please post an answer to elaborate!) And I can only say that "sometimes I vote with the context factored in". For example, when shown two similar but subtly different solutions I will often vote up only the one that I feel gives the better presentation (though I've never voted down the other based just on it not being as good as the other). So I am afraid I cannot even say whether I have a strong personal opinion to the question you asked. – Willie Wong Jul 30 '13 at 13:34
• -1 because I think it is acceptable to downvote answers one sincerely believes are technically correct yet not helpful to the asker. – Henning Makholm Aug 3 '13 at 15:43
• @HenningMakholm I sincerely but respectfully disagree with this reason for downvoting correct answers. Even though an answer my not be directly helpful to the OP himself, it may be helpful to and appreciated by other readers of the question. The OP of the question is not the only person the answers can be useful to and who can benefit from them. I still see no good reason for downvoting correct answers and think it is rather harmful for reasons explained in other comments to this meta question. – Dilaton Aug 3 '13 at 23:36
• @Dilaton: As far as I'm concerned, being technically correct is a necessary but not sufficient condition for an answer to be useful to the question as asked. If, for example, someone answers a precalculus question with a technically correct dissertation about category-theoretic generalizations of arithmetic, then I reserve the right to deem the answer not useful and down it, no matter whether the information the asker looked for is technically present somewhere in all the abstract nonsense. – Henning Makholm Aug 4 '13 at 1:12
• The statement "it is acceptable to downvote answers one sincerely believes are technically correct yet not helpful to the asker" worries me. There are situations where an answer provides useful insight to an audience that does not include the OP (at least at the time of asking) yet is entirely relevant to OPs question. I frequently find such answers worthy of upvotes. Conversely, I think a technically incorrect answer can be useful (as long as it is not misleading). – anon Aug 4 '13 at 2:25
• @HenningMakholm downvotes of answers that put for example a simple question into a broader picture, as I understand what you say, belong to the in my opinion bad from a scientific point of view not justified, downvotes. Now I know that you probably contribute to the not justified dowvotes, and if I see them I always cancel them by an upvote and leave a corresponding comment that there is nothing wrong with this answer. Not tollerating answers from a deeper point of view than the OP asked aboult is quite narrow minded. I exactly agree with what the anon said above me. – Dilaton Aug 4 '13 at 7:32
• @Dilaton: I mostly disagree with the opinion Henning has expressed here. Especially, IMO thinking that an answer will not be useful to the OP is a good reason not to upvote it; in order to downvote it, you would need to think that it won't be useful to anyone, and it is very hard to see how a correct answer could not be useful to anyone.... – Pete L. Clark Aug 4 '13 at 9:39
• ...But I don't think you're being entirely fair to Henning: he didn't say that he downvoted answers according to this practice, only that he "reserved the right to do so". Also he didn't use the phrase "technical nonsense"; he said "abstract nonsense", which has a reasonably specific, non-derogatory technical meaning in mathematics. Note also that -- along with lots of people in this discussion -- he is describing a very extreme, presumably hypothetical situation. Or can someone exhibit a precalculus question which has been answered using category theory?? – Pete L. Clark Aug 4 '13 at 9:42
• @PeteL.Clark agreed, I probably let me carry away a bit by my disagreement with the downvoting practice described. An additional point why I disagree with the idea of Henning's practice is that if done like this, it could be interpretted as patronizing the OP by deciding what is useful for him. When I ask questions, mostly on Physics SE, I am well capable of deciding myself what I consider useful to me by accepting an answers and more often than not upvoting additional ones. As I always like and appreciate answers that give a deeper insight into the issue I asked about, – Dilaton Aug 4 '13 at 12:17

I am sort of new to this site but some might know me from a place close by. I find this discussion very interesting on abstarct grounds. It never is quite clear to me why there is such a, in my opinion, huge difference between upvotes and downvotes in peoples perception. Don't get me wrong, I do not downvote a lot myself, neither do I like to get downvotes either, but why such a strong reaction against downvotes.

The user whose text has spawned this debate in my opinion votes both ways, as many do, for up, namely to a considerable part also on the relative merits of an answer (against other answers).

Now, some, in particular Pete L. Clark seem to put forward an idea of some 'absolute' correctness based voting, which he seems to apply (upvote everything correct).

First, this raises the issue that what is prominently propagated as one of the sites principles that the 'best' answers raise to the top would not work well or at all anymore; two answers that are both correct while one is better would more or less inevitably get the same score (except one is given earlier perhaps, but then more often then not a better answer might come latter as mentioned by others so that it would even be worse and the better answer would always stay behind).

Second, as he also acknowledges, this is not viable for people that view a lot of question. And, third, it is certainly not what happens in practise that much I think, otherwise some very simple answers would have to have a lot higher scores. In addition, if this principle was really followed the simpler an answer the higher the score (as more people can judged its correctness), assuming comparable number of views, which is sort-of reasonable.

So, how do (many) people actually vote (I think), or at least how do I vote (up).

At least I typically vote on perceived merits or usefulness if you like. If I think an answer is good I vote it up, if it is just correct but sort of standard, I might still vote it up if its score is low, and if not then I just leave it. But here is the first point where my voting is not absolute in any way; whether or not several other people already upvoted the answer does not really change its usefulnes. Still I feel that there is something like a reasonable score for some type of answer (with all the vagueness attached to this). Observation of voting patterns confirms that this is not isolated (though there are some other effects to consider, namely that people follow others in voting; like, this looks good but I am not sure it is correct, but well 7 others thought so alread so it should be fine +1).

Now, there is a second point where my (up) voting is not absolute. Namely, if there are several answers I might only upvote the one (or the two or more) I consider as best and not all, in particular if the relative score distribution is not as I think it should be. And, even if I tried not to do this, it might even be hard to avoid. For if I see a question and I see one answer that only gives a partial answer then I might objectively find this more or less impressinve depending on whether there is another answer next to it that gives a full answer (in particular in a case where I had not known an answer myself). So, again, my up voting depends on the other existing answers.

Now, in OP it is claimed the order should be only a consequence of and not a cause for the votes (mentioning both up and down). At least implicitly postulating some voting based on the merits of the answer looked at in isolation. I doubt many people really vote (up) like this. Judging each answer on its own merits not taking other existing answers (so the relative quality) into account at all not even indirectly. Yet as soon as one does this, one already votes for the purpose of sorting. Again, I think for up-votes this type of voting is at least not uncommon.

Now what does this user do that is so questionable, in some people's opinion, they use up and down votes in more or less the same way, and not with quite different meanings as is the standard way it is used up vote everythng you like there are no very clear rules and hardly any questions asked (except for voting fraud) but downvote only for serious problems, like incorrect.

I know it is like this, but it is not quite clear why it must be like this (in the sense that somebody wishing to deviate from it faces quite severe opposition, including more or less threats to be suspended).

Sure, one looses points via downvootes, but also one does not get any by non-upvotes or gets unjustified ones by unjustified upvotes; and the stakes are higher for up-votes (2 against 10 as we are talking answers).

Therefore, my opinion is, if one does not also impose strict guidelines and norms when to upvote, I do not see why this must be so different for downvotes.

Again, sure the one is something positive and nice and the other not and so it is not quite the same, and perhaps it is soemhow good it is like this that downvotes are rare. All I am saying is that the reasons for it are quite vague and rather in the realm of emotions. Therefore, in my opinion, quite a bit of tolerance should be excercised towards those that for one reason or another deviate from the standard, since this standard is sort of arbitrary.

Analogies are always difficult, but if somebody downvotes a lot (but not in spiteful or ad hominem ways) the worst (I do not say one should do even this) I think one may think is this is a bit of a grumpy or unfriendly colleague; not unaccaptable let us fire(ban) them already (if they do not change their ways instantly), in particular not if in other ways this colleague is competent and supportive.

• About downvotes: I find it a little baffling that a lot of people defend very subjective or harsh downvoting philosophies and then clarify that they almost never downvote themselves. To be clear, I am not suggesting that "You should get suspended if you downvote in X way". I was (i) calling attention to the fact that in the past people have been suspended for downvoting practices, so 40-votes's early comment which suggested that this could never happen was counterfactual; and (ii) arguing for a certain code of conduct that would be beneficial to this site. – Pete L. Clark Jul 30 '13 at 19:15
• Fundamentally I am arguing for an increased meaning: there is a sorting into better and worse answers. There is also an evaluation of whether answers are correct and incorrect. The latter is in my opinion one of the great services that a site like this provides: user $X$ can post something quite technical; they may not know anyone nearby who is qualified to evaluate it. But somewhere in the world there will be others who can read and evaluate that answer for correctness. I think it would be great if the downvote were used and used for this purpose alone. – Pete L. Clark Jul 30 '13 at 19:18
• Thank you for the clarification. Yes, the advantage of chronologically first is a problem I see. The revised strategy you propose seems reasonable to me (though I have in addition some vague thresholds beyond which I do not vote standard things, or I do not vote at all on standard things, as often they are rather over-voted anyway relative to more specialised things). A main reason why I argue for not taking downvotes too seriously or for the flexibility of their usage is that I would find it healthier if people would not take them too seriously as on occassion they will happen. Also, – quid Jul 30 '13 at 19:45
• and this is in my opinion almost a design-error of the site at least in view of the community norms that evolved, since in some way the way up/down voting is presented via the interface is in complete symmetry. It is not that a downvote is presented as something so serious (as opposed to the flags, where one is really told do this only for this, and are you sure this is an instance of this and so on). By contrast a downvote, is technically like an upvote. And, I think to for indicating errors it is not sufficient to only downvote, you anyway need to say what is wrong. – quid Jul 30 '13 at 19:49
• I think I agree with everything in your last two comments, at least if your final "need" is read as "morally ought to" (which is the same way that I intend my abjuration of downvotes of correct answers to be read). – Pete L. Clark Jul 30 '13 at 19:56
• So, it is a priori not clear why the act "downvote" needs to be reserved for this purpose. If it is for incorrect, one could say it is expected there is an explanation in addition. If there is none one could then infer it was for whatever, and simply move on. I think this is simply the healthier way to proceed. As it avoids not inconsiderable turmoil that can happen over spurious downvotes that one can never fully exclude. So, I think they should be rare but not too restricted since if they are so restricted the get too powerful for creating conflicts. – quid Jul 30 '13 at 20:02
• My final comment was written before seeing yours. In fact this 'need' is meant even still weaker. What I mean is if you want the person to understand what is wrong, which seems geenrally desirable, it will often be necessary to explain it (to this end you need the comment) – quid Jul 30 '13 at 20:04

There is a lot of discussion going on here about whether or not it's so bad for someone to use their voting privileges outside the prescribed guidelines. And while there is something to be said of a user's ability to vote however they please (and the inability to stop them from doing so), I think that that voting outside the prescribed guidelines is a detriment to the site. And it's one that we can't really do much to prevent, as I'm sure many will be eager to remind me. But I do think it should be discouraged. Here is why.

The voting arrows, when rolled over, say the following:

A lot of the current discussion seems to focus on the idea of this being prescriptive for the voter. But in fact this (and the FAQ) reads more as being descriptive for the reader of the answer. In other words, imagine I'm the reader of a question. I see that it has 3 downvotes. I wonder "hmm...what does this downvote mean?" so I roll over it and see the above description. Moreover, I check the FAQ and see much the same description. So I will conclude that this is a bad answer, and is inaccurate or incomplete in some way.

This is where voting outside the prescribed guidelines becomes a problem. When it happens, the FAQ and rollover text less accurately describe the nature of votes. If a perfectly good answer is downvoted into oblivion because another answer was better, newer users who are familiar with the FAQ or who check the rollover text will be mislead to believe that the answer is simply not accurate.

Some have argued that people are more likely to simply read from top to bottom, rather than taking the votes into heavy consideration. But this will only be the case for users who have not familiarized themselves with the FAQ and the way the site works (which by the way, is something that seems to be extremely strongly encouraged here, so it's reasonable to assume that 90% of users indeed have).

So something important to consider is that the voting guidelines are not only prescriptive of expected practices, but also descriptive to readers of the meaning of votes. It's on us, the community, to make sure that this description is accurate as possible.

• Thank you for the replies on the other answer and the thoughtful answer. I thought I will comment and reply here. Let me start with a detail, really new users I think cannot see the downvotes and I think nowbody is proposing (as it also would be unnecessary for sorting) to vote things to negative score. And one always has to actively check to see if there were downvotes so for casual readers the risk seems not so high. Also, this goes a bit both ways. In the about it says very early on that the best answers show up first. [cont] – quid Jul 30 '13 at 18:58
• [cont] So, one could also say one wants to follow this and make sure the best is really first (because otherwise readers will be mislead in another way), which is precisly the point of the discussed voting strategy. Now for the replying part: I agree that for marking as accepted it is given the current software necessary to compare. But again this could be easily different. Why not let OP accept however many answers they want (about says the answer worked for OP, so if more than one worked why not let them accept more than one). [cont] – quid Jul 30 '13 at 18:59
• [cont] In general, from the description in the about and this is also acknowledged in OP here, a main purpose of the voting here (at least commulatively) is to find the best answers. And this is comparative, too. In OP and also in what you say, the idea is more still one needs to evaluate each answer solely on its own merits. One can hold this opinion, but as I said I see some issues here, and in particular it is somewhat at odds with the idea that still the total score should give some ordering of the answers. [cont] – quid Jul 30 '13 at 19:00
• If this is the main point, according to about, why not take it actively into account when voting (up). To reply to your point why a partial answer should become somehow worse if a complete one follows: for example it could have an effect on ones judgement on the difficulty of the problem or the taken approach. Say somebody asks a question and I do a lengthy calculation to arrive at a partial solution then somebody seeing this might think: wow, this is really a complicated problem, great that user achieved a partial solution [cont] – quid Jul 30 '13 at 19:02
• [cont] yet if then you come along and have a better idea and solve it completely in five lines, then afterwards maybe my intial attempt will not look so impressive anymore and I think this would be fair. One could also look at it in a different way: say a partial answer gets posted, a complete follows. You say the lerit of the partial one should remain unchanged. But, if this is so, could one also post a partial answer after a complete one (and assuming it is not a different approach but more the same idea only sufficiently different that it is clear it is not copied). [cont] – quid Jul 30 '13 at 19:04
• If merits stay strictly unchanged by existing answers why not. Sorry for so many comments! This got way longer than I thought. But, certainly, we agree that one should not downvote too much and also I agree it is very important that it is in general roughly clear what votes mean. – quid Jul 30 '13 at 19:05

I honestly don't see such a use of downvotes as a major problem worthy of a public trial of a specific user (which is what this thread is about to turn into). A few points that occurred to me while reading this discussion.

1. The FAQ may say whatever it says, but we didn't sign an affidavit to follow the letter of the FAQ when registering, did we?
2. The stated goal of this voting policy is to improve the relative pecking order of the answers. I fail to see how this is such a bad thing. One may argue that we don't know whether this user is actually following this policy, or whether is he using this as an excuse to cast downvotes left and right. As of today his upvote/downvote ratio is 5:2 (with less than 100 downvotes). That ratio is lower than with most, but not extremal (we have seen worse). It does sound like he is exercising this policy in moderation - not worthy of this kind of uproar IMHO.
3. There is so much variation in the voting policies of members anyway. If you look at the stats of the top contributors, you will find a lot of variation in the both the upvote/downvote and question/answer ratios as well as with the overall level of voting activity (say, per reputation points). This policy is uncommon, but I don't think we should try to enforce even a clear majority view on this issue. The input of a single user will be lost in the crowd noise anyway.
4. [I removed this point as it was making an uncalled for comparison. I was being a drama queen. I apologize to those offended by it.]
5. I kinda like the idea of this user putting his hard earned, if not money, then at least his reputation, where his mouth is. Instead of just whining here occasionally how relatively weak answers rake in a lot of upvotes simply because they could be followed by the masses, or some other such woes affecting/originating from the reputation game. Tomorrow I may regret including this point, and admire the gall less, but today is what it is.
6. I must add that I hope this user is not downvoting "competing" answers except in the most clear cut cases. IMHO that would be bad sportsmanship.

It is quite possible that I am underestimating the ill effects of a handful of downvotes. I am simply bad at psychology. Also, I wasn't here in the beginning of Math.SE, so I don't know, if it is written in the constitution that thou shalt not downvote a correct answer. My own observations have rather supported the point of view that people are free to vote according to their chosen policy. For the record: my upvote/downvote ratio is 70:1 or thereabouts. I am not considering adopting this kind of a policy myself, but if somebody wants to do it that way, so be it.

• I wonder why there is a vote to delete this, and wish there weren't. Point 4 is overdramatic. I agree generally with this answer. – Jonas Meyer Jul 30 '13 at 17:39
• I think that the talk of "suspension" is being taken out of proportion. As has been clarified in comments, the example that someone had in mind was a user who had been downvoting other answers for questions in order to improve the relative standing of his/her own answers, which is not what this overall discussion is about. – Carl Mummert Jul 30 '13 at 18:00
• It may be that I took something out of the proportion. But I am not the only one. We are having this discussion, because someone, as a consequence of his public policy, has cast 59 downvotes in a period of 16 months with hundreds of thousands of answers to pick from. – Jyrki Lahtonen Jul 30 '13 at 18:13
• I think it is worth discussing, and I think you're getting the impression that this discussion is about the particular user. It's more about the practice itself. If someone steals a penny from Fort Knox, would it be blowing it out of proportion to discuss, simply because all he took is a penny? It's the practice that we're discussion, not the volume to which it took place. – Ataraxia Jul 30 '13 at 18:18
• @CarlMummert: I do not think your description of the situation is accurate. It is true for some comments on OP, but please note commenting on Alex Becker's answer it was asked "Would downvoting in the problematic fashion being discussed be grounds for moderator actions, such as suspension?" the reply was after saying it unlikely to happen for purely practical reasons "However, if a user repeatedly claimed to downvote posts which they thought were correct and helpful but note the most helpful, we would probably at least consider taking action." – quid Jul 30 '13 at 18:21
• @Ataraxia: I see your point. But my point was that I wouldn't want the person who stole a penny to be tried by a lynching mob, nor to see him publically tried all over the media. A court of law is the right place, or in our case a team of trusted moderators. – Jyrki Lahtonen Jul 30 '13 at 18:45
• @Jyrki: While I like and agree with most of your answer, here are some points. (i) Please avoid comparing actions on an internet site to oppressive political regimes. I don't think this comparison is fair or even really respectful to either the users of this site or the inhabitants of North Korea. (ii) The post did not link to or otherwise explicitly identify the specific user. I for one don't know who that user is and don't much care. (iii) I don't think it is a bad thing at all to clarify site actions as being approved / tolerated / neutral / disapproved / possibly worthy of suspension. – Pete L. Clark Jul 30 '13 at 18:58
• In particular, I am not a moderator, and what I am advocating is a code of conduct, not that certain people get suspended. Since I am saying that I find casting downvotes in unacceptable according to certain internal cognitive processes of the user, this could only even possibly lead to a suspension if the user was extraordinarily vigorous and public in this practice. E.g. if someone cast 10 "ordering downvotes" a day and left repeated comments explaining that that's what they were doing, then perhaps that would turn out to be a problem. – Pete L. Clark Jul 30 '13 at 19:03
• The downvoting voting may happen. The amount depends on the answers I read and their nature. Explaining it constantly is what is very unlikely to happen. Well, I am sure I will not do it often. Probably almost never. That is why I put the short explanation in my profile and a more extended explanation here in my answer to this question. – OR. Jul 30 '13 at 19:10
• The fact that is only such an awkward condition (downvoting + explaining) what would make the action punishable should tell you there is a design conflict. Ability to downvote freely, the de facto role of voting, and the interpretation the FAQs gives, do not pull in the same directions. – OR. Jul 30 '13 at 19:21
• @PeteL.Clark: but in particular if the person explains what they are doing why should this be that much of a problem. Then it would be clear to everybody, no there is not mistake, no there is not spite, it is only I think this answer is better then the others. And commments to this extent happen a lot on this site as far as I see. (something like "I think this one should be accepted answer" or "I prefer this one over accepted answer" or "this answer seems better than that answer because of..."). And, also in CW it is considered alright to downvote, why? – quid Jul 30 '13 at 19:21
• @quid: You say there is no spite, but it is annoying to get a downvote on a correct answer, and the explanation "I like the other answer better" does not help: it is not a very gracious response to the time and effort someone put in to leave a correct answer. Comments, especially those which are phrased to be supportive of the other answer rather than negative of the answer one does not view as optimal, are much nicer. I also feel that never downvoting correct answers is a great way to keep one's ad hominem feelings in check. – Pete L. Clark Jul 30 '13 at 19:47
• About CW: that seems like an easy one. CW is for questions that are deemed nonobjective, so that answers are matters of opinion. The fact that there is no reputation change in CW questions shows that the site views them totally differently. Still, casting upvotes and downvotes for CW answers seems like getting twice as many votes as someone like you who doesn't like to downvote. I see much less clear purpose in downvoting CW answers, for what it's worth. – Pete L. Clark Jul 30 '13 at 19:51
• @PeteL.Clark: re ad hominem, I always feel the bigger issue regarding ad hominem influence on voting is up if one does not up-vote only because one does not like the poster it would 'cost' them 10 points, how to keep this under control seems more relevant and difficult; also upvotes mainly because of id of poster is a problem. Thus, I also consider ad hominem downvotes not so relevant as IMO there actual infuence is completly negligible compared to the other effects juste mentioned (I know I got some over time on the other site, but really I don't care.) [cont.] – quid Jul 30 '13 at 20:29
• @quid: Again, I fully agree with all you say. Especially: an unexplained downvote is worse than a downvote accompanied any reasonably courteous explanation. (This is especially true to me, since I interpret downvotes as expressing that there is something specifically wrong with the answer. I think almost anyone would worry though that that might be the reason for the downvote.) Thanks for a very thoughtful and nuanced discussion. – Pete L. Clark Jul 30 '13 at 20:59

I am not sure if this is useful and it does not directly answer the question: but my strategy to deal with downvotes that are in my opinion not justified is the following:

• When I detect such downvotes of posts other than mine, I counter them with a +1 and again leave a comment that I see no need for a downvote to reassure the OP that there is nothing wrong with his post in my opinion.

In particular leaving such corresponding comments may at the end even lead to a higher score of the downvoted post than it would have obtained without it, if people agree that the downvote was not justified ...

I will explain how I see it (slowly adding the ideas. I need to save because my browser has a glitch and I lose it if I don't save quickly).

small point $1$: There is a design problem (perhaps 'design conflict' sounds more accurate) if the software allows for voting down however one pleases, while at the same time it is wanted to give the votes a meaning that goes beyond "I want this answer to move up and/or this one to move down". Not all design conflicts have a solution at the level of software, of course.

Point 2: This is a website where answers to questions are given and stay here for further reference. As in all websites in which answers are given and rated, in the long run the answers start getting organized according to those ratings. What is important is that this order goes with the main goal of the site, which is to educate, to inform. It is a fact that we read from top to bottom (usually), so whatever is further up is read first. For that reason, it is of the interest of Math.SE to bring up more useful answers and down the less useful ones.

Point 3: I have been told about the 'written rules' (as opposed to the 'built-in rules', e.g. a user with reputation $1$ cannot downvote) on when to downvote. But I think that in the same way a user can be told (more like suggested) only to downvote in certain ways, a user can be told (and this is more easy to achieve) how to interpret a downvote. Afterall, see point $2$, the content and the form of that content that gets created in this site is of more importance than the game of reputations, or egos, or interpretations of votes.

Point $\pi$: Even if I stop using votes this way or prevented from doing it, votes and specifically downvotes that do not follow the expected prescription will occur. So still the pad down in the shoulder while saying "don't worry about the downvotes" is being necessary, and is being used anyway. The reason for this is point $1$, there is a design conflict.

Point 4: Solutions: and of course I don't have a polished finished big idea just some small ideas that the rest can take, leave, or adapt and enrich, ...

Point 4.1: Not do anything. After all, the effect of a single person (me) gets averaged out, in the long run. The effect in the long run (the ordering of the answers) is what I think is important, and I understand the effect of only one person is negligible, so I am at peace with my effect disappearing.

Point 4.2: Another solution could be to instead of frowning upon, or making some downvotes seen as bad, with the aim of giving guidelines on how to downvote (or vote in general), it is much easier to tell people how to interpret receiving a downvote. Currently the latter is being done. People are told that sometimes there are crazy downvoters. Of course because downvotes occur that may not be explained with the conflicting guideline of downvoting when there is a flaw.

Point 4.3: Another possibility is to not have negative downvotes. There maybe many ways to actually do this, but what is important is that the effect of downvoting is not putting an ugly negative sign next to an answer, but to move it down in the list. A negative sign is aggressive, punitive by nature. By just getting rid of it you can make the users of this site happier users. Simply moving answers down in the list has the desired effect of giving them less relevance.

Point 4.4:

Point 5: A quote.

"The downvotes are functionally inexplicable ..." (Arkamis).


They are functionally inexplicable, and yet the FAQ tries to gives them an explanation that goes beyond their actual effect, organizing answers to a question. This reminds me the greeks defining point as adimensional entities and what not ..., and then mathematics making the transcendental step of defining them as whatever behaves as a point does. The power of defining something just as it behaves have yet to get to the voting system in this website.

OK, done. Perhaps for the moment.

Wait ... Nevermind. hehe!

See, it is impossible to try to control how people downvote, if they are free to choose. But how to read downvotes, and what the actual effect is, is easy. The effect doesn't transcend the placement of this answer in the list of answers.

• When I see I a downvote, the first and only thing that comes to my mind is "Hmm... where have I gone wrong? Have I made a mistake somewhere?" That is how I -- and I guess others -- interpret downvotes. Of course, then we find the "you're just trying to engorge youself, take a -1" downvotes, which I don't disagree but neither fully agree with: the answer might be helpful to other users that might come across the question, and maybe not only the OP: of course, one should always put the OP first, but once one thinks that has been cleared out there is no (...) – Pedro Tamaroff Jul 30 '13 at 4:30
• harm in adding more advanced ideas or topics in one's answer. The canonical example, in my opinion, is Bill Dubuque. He provided awesome answers, and most of the times they were hard to get at the first time, but with some effort and practice one could make the best out of them. – Pedro Tamaroff Jul 30 '13 at 4:31
• " The is a design problem if the software allows for voting down however one pleases, while it is not wanted to happen such a sign. The problem may not have a solution at the level of software, perhaps." I cannot understand this. – Pedro Tamaroff Jul 30 '13 at 4:32
• While a user can be trained to interpret something in some way, they've already been trained to interpret it in a different way. A stark minority insisting on a complete renewal of how we interpret things is pretty severe, and typically turns out poorly. – Emily Jul 30 '13 at 4:49
• What I don't really understand is, if you're so keen on sorting answers, why isn't it enough to just upvote the good answers? Those will get sorted to the top obviously, and the ones that you refrained on received one less vote from you...so why isn't that enough to suit your sorting needs? – Ataraxia Jul 30 '13 at 6:25
• So in your version, if someone provides an answer, and later someone else provides a better answer, that first person deserves to lose 2 points of reputation because of this? I am aware that this is a quite irrelevant amount of reputation, but it seems a strange idea. – Tobias Kildetoft Jul 30 '13 at 6:39
• @Ataraxia Yours is the only reasonable argument I have seen so far. Good question. I am was being surprised I didn't get so far. In fact, it is enough to have an upvote per question to be able to express opinion on the ordering of the answers. Downvoting is tantamount to having two votes, one of which applies to all questions except one. If I am given that much power, why wouldn't I use it? – OR. Jul 30 '13 at 6:40
• How can go into a discussion of what is appropriate downvote behavior without first getting to know how downvotes work? Have you considered reading the FAQ? – Tobias Kildetoft Jul 30 '13 at 6:48
• Nah. I am not interested in reading FAQs, unless I need it. I am here mainly to teach and learn math. I didn't go into a discussion. I was taken here. I just gave the reasons for my voting because I was asked, and also people deserved to know, but that's it. The system allows me to vote however I like. If people ask, I am glad to answer why. Since I consider reputation as mostly unimportant it makes no difference to me what effect votes do on it, specially if it so small. – OR. Jul 30 '13 at 7:00
• So, you can't be bothered to read the FAQ and yet you refer to it in your point 5? – mrf Jul 30 '13 at 9:13
• I don't like the idea either. I mean, the combination of a rating that is expected to be only on the answer's own merits (absolute voting is being used in other answers) together with a voting that cannot make voters accountable. By functionality, rate is relative. In a hits list a very good music may be rated low in the list if in that year there are a lot of other productions that people like better. – OR. Jul 30 '13 at 14:07
• @quid Regardless of its partiality, is it a good answer? If yes, then it deserves an upvote. If an better answer is provided later on, does that affect the quality of the partial answer? No, it doesn't alter the content of the partial answer at all. So how does it change the quality of the answer? – Ataraxia Jul 30 '13 at 17:34
• Votes only affect the order if the viewer selects "votes" instead of "active" or "oldest" as their ordering preference. This happens to be the default, but not the only, choice. In any case, ordering the answers is not the only effect that votes have; they tell whether an answer is useful or not. That is what the hover message on the voting arrows indicates. – robjohn Jul 30 '13 at 22:48