# How aggressively should we tag a question as duplicate?

Some questions seem standard and very likely to have already been asked. For example, a question on Prove Binet's Formula would almost certainly have been asked, even without referring to the links by the side.

On the other hand, a question like restricted partitions of n was closed because I commented that it was similar (not the same) as another question asked earlier.

[I'm new to meta, so please edit the tags as need be]

• Cf. #1756. – Lord_Farin May 29 '13 at 7:16
• possible duplicate of Coping with *abstract* duplicate questions. – Micah May 29 '13 at 17:06
• Are these votes to close purely for reasons of irony? – Douglas S. Stones May 29 '13 at 22:34
• I have recently come across a number of duplicates which were either different questions (for example, the old question was on fields of characteristic $0$, the new one was more general), or questions which say "I want to do this question using this method", which was closed as a duplicate but the OP had to do some work to figure out how the duplicate answered his question - which is kinda unhelpful (or maybe very helpful, I dunno...). Also, I have seen "check my proof" questions closed as being duplicates! I think when you get to that point: Yes, we are being overly zealous... – user1729 May 30 '13 at 8:01
• @user1729 The problem to me seems to be not the overzealous duplicate suggestions, but close votes being cast after only a cursory glance (I'm looking at myself here, too). Furthermore, it might help to have some sort of accepted practice of adding explanatory comments upon voting to keep open (so as to inform others of one's conclusions). To me, the problem thus seems the suboptimal attention paid by the close vote reviewers. – Lord_Farin May 30 '13 at 8:08
• @Lord_Farin: I always leave a comment when I vote to leave open (which works, but only when there are not lots of comments already). I almost agree with your analysis of the problem: if all I can offer is a cursory glance I leave it and wait for someone with more time to do the voting. However, if they were never tagged as duplicates in the first place... – user1729 May 30 '13 at 8:16
• We need to make a meta meta math stack exchange so that we can argue about whether this question should be closed as a duplicate. – treble Jun 9 '13 at 3:40

Standard disclaimer: This really is a comment, but it's too long, so...

My gut feeling is that the efficiency of catching duplicates has been going down. When I started here, the site was about a year old, and the old-timers could largely rely on memory in spotting duplicates (exact or abstract), and move swiftly. I managed to build a similar personal smallish database, and join in the battle. I would often remember who had given a good answer to an essentially duplicate question, search his/her history of answers, and act.

Since then the number of questions has exploded, some of the prominent old-timers have disappeared. Recently a top user described his approach as: "This question is probably a duplicate, but retyping an answer is faster than locating the original, so..."

Such a comment could be taken as a sign of lazy attitude, but I want to make it clear that I'm not pointing fingers. As the size of the site grows, a point will be reached, where no one can remember for sure. In other words that comment gains more validity each day. Each and every one of us will eventually reach a point, where personal recollection will no longer help. The break even -point varies from person to person, but we all have one.

Younger members may have grown up with search engines, and know how to use them to optimal effect. Consequently their break even point may be way higher. If somebody wants to share effective search tips and pointers, I'm all ears.

BECAUSE I think that we should and could do a better job in spotting duplicates than we do currently.

TL; DR; The problem is somewhat caused by the deadly sin of sloth. On the part of the askers, also on the part of the FGITW answerers,... But also by understandable human limitations.

CAVEAT: I largely ignore and , so I may not have seen the worst of it. The ultra-high volume tags are likely more difficult to manage in terms of catching (abstract) duplicates than my favorite ones.

Addition on July 27, 2016 (the only part of this post written after I was elected)

There's no end to the sloth. I don't see any alternatives to downvoting answers by veteran users to questions that they surely know have been handled on the site many times. If not in exact same form then essentially so. I am more forgiving to relative newcomers, but the ultra high rep users should IMHO be role models here.

I am not talking about the "accidental" duplicates, some question that may have been asked twice or thrice. This is about the oft recurring questions, all parts of which have reappeared on the site several times. Questions that a veteran cannot fail to think is a dup.

• Beside the valid points you address, there is also the observation that it requires a lot of work (in the order of magnitude of a blog post, if not more) to provide a properly general abstract QA pair (which in my view should be augmented with some worked examples -- the ones instigating the abstraction should do fine). This probably deters people; perhaps we could have an answer to #1868 filled with requests for generalizations (along the lines "I observe this type of question but I'm not in a position to write an abstract duplicate")? – Lord_Farin May 30 '13 at 8:09
• One doesn't have to rely entirely on memory to catch duplicates. Some questions are just so standard that you know that someone has asked them before, even if you have no memory of having seen them here before. The related question down the side of the page are the first place to look for the duplicate. If that doesn't work, pick out a keyphrase and type it into Google with site:math.stackexchange.com. Also, an ounce of prevention is worth a pound of cure: let's be more aggressive about re-titling questions with uninformative titles. That'll make finding duplicates easier. – Gerry Myerson May 30 '13 at 8:52
• You wrote: If somebody wants to share effective search tips and pointers, I'm all ears. I have asked this in a separate question: How do you search for duplicates. Other questions tagged (search) might be of interest, too. – Martin Sleziak May 30 '13 at 12:34
• Thanks, Martin. I should've remembered that one :-) – Jyrki Lahtonen May 30 '13 at 12:49
• Regarding the (calculus) and (algebra-precalculus) tags: Should we be more lenient with duplicate questions with these tags? The people asking these questions might not be mathematically mature enough to be helped by "abstract" duplicate questions. Even if the question is formally the same as another, an answer tailored to the details of the question is probably more helpful. – Potato May 30 '13 at 19:13
• I have in mind here people who are just wrapping their heads around integration or high school algebra. I think the computational details involved in "translating" more abstract duplicates might be prohibitively difficult, since at that level the questions are mostly about errors in computations anyway. – Potato May 30 '13 at 19:14
• I see your point, Potato. It has a lot of merit. It does open the natural question: where exactly should we draw the line? If freshmen are given a free pass, then what about sophomore/junior non-math majors? How can we tell? Looks like flexibility and case-by-case judgement is (again!!) called for. I'm not very comfortable with the concept of an abstract duplicate actually. – Jyrki Lahtonen May 30 '13 at 19:17
• "Since then the number of questions has exploded, some of the prominent old-timers have disappeared." - I have been feeling rather lonely lately, not only in the dupe-finding business... – J. M. is a poor mathematician May 30 '13 at 20:37
• I mean, if we don't use abstract duplicate as grounds for closure in calculus, it is kinda pointless to use it at all. – Jyrki Lahtonen Jun 1 '13 at 7:24
• @Jyrki: I strongly suspect that abstract duplicates are in general more trouble than they’re worth: the questions that come up most often are, as Potato suggested, very frequently asked by people who have already failed to make sense of what’s in their text or lecture notes, which very likely qualifies as an abstract duplicate. – Brian M. Scott Jun 7 '13 at 22:31
• Since this just became active a personal request: could you show me a place where my deleted answer on this question could have been posted instead? The one where it would fit nicely is closed as a dupe, incorrectly in my opinion (by you among others). Someone pointed me to about dozen questions to show it is a dupe, but neither did I find an alternative place nor did I see an answer that fully subsumes mine. (Of course in one way or another the thing came up numerous times.) – quid Jul 27 '16 at 18:11
• A strong $({\bf {-}1})$ is well-earned for the advocated destructive downvoting (and also for insulting users ("sloths") who do not share your personal opinion on the topic of "duplicates", not to mention digging up very old posts). Very poor form for a "moderator". – Bill Dubuque Jul 27 '16 at 18:35
• Re: If somebody wants to share effective search tips and pointers, I'm all ears. (2nd time - since the question has been bumped.) I doubt that I will mention something you don't already know, but I have posted answers with tips for searching here. Some additional links are collected here. – Martin Sleziak Jul 31 '16 at 7:31