# What is the standard for upvoting an answer you don't know is correct?

Often, when I look at a question and answers, I come across a long, well-written answer to a difficult question. I may not have the expertise to verify the details of the answer, or I may have the knowledge but not the available time to do so. Recently, I have been going ahead and upvoting the answer, even though I am aware that, as far as I know, it is possible that the solution has a serious flaw in it. My only evidence for believing the solution to be valuable is what, if anything, I know about the particular poster and the empirical fact that most detailed responses to difficult questions that I am able to verify do, in fact, turn out to be correct (especially when considered modulo minor typos and such).

Is it appropriate to upvote such an answer? My rationale is that, in many cases, very few people will have both the relevant background and available time to check such a detailed answer, and it seems unfair that some of the very best answers to questions on the site receive very little credit, while a solution to something very basic (for the responder and many of the members of this site, though of course not necessarily so for the original poster of the question) may receive a flurry of upvotes immediately. I personally am much more interested in reading the long, well thought out replies to problems whose solution method is not widely known, so I would like to show my appreciation for such answers.

But I have enough concern about my practice to solicit community feedback and see whether others engage in a similar practice or whether some find this practice against the spirit of what an upvote should signify. And more broadly speaking, what are good criteria that people use in practice to determine what answers merit receiving an upvote?

• As a general rule I won’t upvote an answer that I’ve not worked through, though for answers from a few regulars in a couple of fields I may sometimes skim a bit. I’d be very uncomfortable upvoting an answer that I’d not checked. Frankly, this means that I rarely upvote answers that take significant effort to think through unless either I’ve worked on the question myself, or questions are slow that day. – Brian M. Scott Feb 25 '13 at 19:30
• So implicit in that approach is that the verification of the correctness of the answer is more important for voting than the overall value of the effort put into the answer. One consequence of this is that it ends up incentiziving people to focus on answering the easier questions than the harder, thought-provoking ones. That's not to diminish the value of answering "easier" questions, especially since really good explanations and hints can take a lot of work. But I do think that standard leads to the community valuing the more in-depth answers much less than what feels right to me. – Michael Joyce Feb 25 '13 at 20:56
• I guess a side question of mine which motivated this thread is "Should we be rewaring a certain type of community member more than we currently are?" And perhpas upvotes is not the way to do it, but I still think it is problematic when people get 6 or 7 upvotes for explaining how to solve a basic integration by parts problem from Calc II, while a detailed response on the literature of a (already solved) research problem that someone came across in their work gets 1 or 2 upvotes. Obviously, MSE reputation isn't the end all and be all of life, but it is part of what drives people to participate. – Michael Joyce Feb 25 '13 at 21:18
• So implicit ... answer: Yes, of course: the effort put into an answer has no intrinsic value to the querent or the site-as-archive. The value of an answer is determined mostly by its content and slightly by the mechanics of its presentation. That answers to harder questions get fewer upvotes is a predictable fact of life, and I’m not about to compromise my standards to try to compensate artificially. (Besides, I find that answering a hard question is its own reward.) – Brian M. Scott Feb 25 '13 at 23:10
• @Brian I don't disagree, but I notice this on stackoverflow, as complex answers get less upvotes. They require more work/reading to verify. You would hope it be the other way around. – AaronLS Feb 26 '13 at 23:31

Compared to the overall level of activity on Math.SE, voting is slow. SE allows up to $\mathbf{40}$ votes per day if you play your cards right. I average $\mathbf{32}$ votes/day, the most on Math.SE by a wide margin.

• gives a reasonably detailed derivation of a nontrivial result
• is coherently written, with logical flow within and between paragraphs
• uses thought-out notation and layout of formulas
• does not raise OP's objections ("that's not what I ask about")

then I see nothing wrong with upvoting it, if only to affirm that this is the sort of contribution that Math.SE needs more of. In my experience, careful reading of such answers (and I do read them) shows them to be essentially correct with very few exceptions. That said, I normally don't vote on answers outside of my areas of competence. I don't even see such answers (or questions), due to aggressive use of hide-ignored-tags feature. Sorry, this includes .

To compensate for the preceding sentence, I'll point out a way to quickly find potentially underappreciated answers in algebraic geometry: search for [algebraic-geometry] is:answer score:0..2. (Or some other score range). Adding isaccepted:yes can filter down to the answers approved by the OP, which generally indicates some degree of correctness*. This is not meant to batch-upvote answers, but to quickly get a list of answers which deserve another look, and likely another upvote.

The standard SE search does not take length into account. But the Data Explorer can sort by length. Someone with SQL knowledge may decide to write a query for low-score answers of substantial length (perhaps also with a given tag).

(*) I'm not considering the scenario of users posting homework questions they don't understand and subsequently accepting answers they don't understand.

• It's a bit unfair to say that you have the highest record. It's almost obvious that you were a member before this account, or at least observed the affairs very closely. Many people on that list were members for quite some time before they even started playing an active role in the community. I could easily start a new account and vote 40 votes a day, and let you know next week that I beat your average by a wide margin as well. – Asaf Karagila Feb 24 '13 at 20:14
• @AsafKaragila Excuses, excuses. While we're at it, the second user on the list (BenjaLim) participated in Math.SE since March 2011, but the current account is 6 months old. – user53153 Feb 24 '13 at 22:19
• BenjaLim had a split account and it was merged into the newer one. It's some sort of a discrepancy. I am fairly certain this is not the case with you. :-) – Asaf Karagila Feb 24 '13 at 22:31
• @AsafKaragila I see; division by the "age" of account may distort the picture. I should have quoted the number of votes in 2013 instead. – user53153 Feb 24 '13 at 22:38
• Yeah, that would probably have worked for the better. Sadly, J.M. took an absence of leave somewhere in mid-January, which explains the decrease in voting. :-) – Asaf Karagila Feb 24 '13 at 22:41
• @AsafKaragila To be more precise (since I'm observing the affairs very closely), he's on a break since late November. Visited the site once since then, on January 16. – user53153 Feb 24 '13 at 22:45
• True. I actually noted that after my previous comment (I checked his activity tab again). – Asaf Karagila Feb 24 '13 at 22:50

I wouldn't upvote an answer unless I have verified that most of the post is correct, or at least is less incorrect than other answers with as many upvotes.

People who answer harder questions know that they will earn fewer reputation points but choose to answer anyway, presumably because they value producing correct answers more than earning reputation points. So let's vote in a way that's consistent with those values and helps the correct answers rise to the top.

• +1 Although I find that helping "correct answers rise to the top" is hardly an issue in my area. The questions that I tend to look at are lucky if they get any answer at all. – user53153 Feb 24 '13 at 19:56
• @5pm Good point, so I should add that in the case of a unique answer, spurious upvotes may pressure the OP into accepting an answer that he/she doesn't understand or can't verify. – Trevor Wilson Feb 24 '13 at 20:41
• I have zero sympathy for the OP in that scenario. It's up to the OP to know if they understand the answer or not, and to react to unclear or insufficiently detailed answers with comments. – user53153 Feb 24 '13 at 22:23
• @5pm, Trevor: Sometimes there is only one answer, but it sucks. Even if it is long. In some cases I could tell that the content of the answer is gunk just by seeing that a well-voted question had a zero-voted answer by a certain user. Despite the question being very long and obviously a lot of time was spent, it was the wrong case to vote it. – Asaf Karagila Feb 24 '13 at 22:51
• @AsafKaragila Sure, no reason to vote up nonsense or an answer that misses the point. A more typical (for me) example: I just upvoted this answer without checking the correctness of every computation there. Shame on me, shame, shame. – user53153 Feb 24 '13 at 23:05