I've become acutely aware recently that a significant number of users rush to post an answer so that they are top of the list of answers. These answers are often very brief, lack explanation and, by their very nature, have not been thought about for very long.

This behaviour is encouraged by several factors:

  1. People are more likely to visit newer questions.
  2. Answers with equal votes are listed chronologically.
  3. People tend only to read the first couple of answers.

The result of this is that the quicker you can reply then the more traffic your answer sees and the less competition your answer faces. That means your answer picks up initial votes very easily. Once you have a few votes then your answer is right at the top of the list and so is one of the few, if not the only, answers that is read once the questions starts to age.

Someone could spend a great deal of time writing a thoughtful and well-explained answer, only to find that when they post it, it's at the back of a queue of several two or three line "tweets". Despite the fact that the answer is probably of higher quality than those above, due to lack of exposure, the answer sits at the bottom of the line collecting no up-votes.

Moreover, it seems that many OPs will read the replies one-by-one until they find one that answers their question and then accept it. In other words, they don't select the best answer, but they select the first answer. I'm sure you've all found yourself on the receiving end of this quick-draw contest.

I would be interested to hear your thoughts on the matter, and to hear if you have any ideas about how me might move away from encouraging quick-fire replies.

  • 15
    $\begingroup$ I think that answers with equal numbers of votes are shown in random order. $\endgroup$ Aug 27 '13 at 19:29
  • 1
    $\begingroup$ @JonasMeyer That is only true if you sort by votes. Also, part of the problem is that voting has started before the lengthy answers get posted. $\endgroup$
    – user1729
    Aug 27 '13 at 19:37
  • $\begingroup$ @user1729: I guessed that sorting by votes was the context of Fly-by-Night's point about the order of equally voted things. For the other two methods, either late or early posts may be at an advantage. (I more frequently sort by "active" than "oldest," for example.) Edit: But "active" counts editing, too, so it might end up giving yet another advantage to an early post lately edited. $\endgroup$ Aug 27 '13 at 19:42
  • $\begingroup$ @JonasMeyer Sorry, you are right about Fly by Night's point. $\endgroup$
    – user1729
    Aug 27 '13 at 19:44
  • $\begingroup$ @JonasMeyer Your edit is one of the reasons I sort by oldest. It gives you the "truest" picture of the answers, and allows you to see if there is an answer which just got there first and so has a high vote count. It kinda forces you to look at all the answers individually. $\endgroup$
    – user1729
    Aug 27 '13 at 19:52
  • 2
    $\begingroup$ Here's a feature request to have "newest" sort order. $\endgroup$ Aug 27 '13 at 19:57
  • $\begingroup$ @JonasMeyer My personal experience is that posts with the same number of votes are listed in chronological order with respect to their last edit. For example, if myself and another user each had one up-vote and I was listed first, but then I edited by answer, I would be second. That's the way it appears to me with my browser and my set-up options. $\endgroup$ Aug 27 '13 at 20:21
  • 1
    $\begingroup$ On the SE networks this is known as the Fastest Gun In The West (FTGITW) problem. As you can see from the MSO discussions, there have been many proposals to solve the problem, but most of which suffers from one downside or another that makes its net benefit questionable. $\endgroup$ Aug 27 '13 at 20:36
  • 1
    $\begingroup$ meta.stackexchange.com/questions/6917/… $\endgroup$ Aug 27 '13 at 20:44
  • 2
    $\begingroup$ I've had similar concerns: meta.math.stackexchange.com/questions/9229/… $\endgroup$
    – Ron Gordon
    Aug 27 '13 at 21:43
  • 20
    $\begingroup$ It's so frustrating when answers that I would be ashamed of writing get rushed out and even accepted while I'm writing (what I consider to be) a thought out response. Consequently, I often wait for the dust to settle on a question before even attempting to answer it. Yes, I might spend considerable effort for 0 rep, but that's not the point. $\endgroup$ Aug 28 '13 at 0:49
  • 6
    $\begingroup$ Personally, I enjoy watching multiple users trying to do calculus II the fastest. I don't think it's just about votes. $\endgroup$ Aug 28 '13 at 1:36
  • 4
    $\begingroup$ I think it's hilarious that this particular question was posted by a user named Fly by Night. $\endgroup$
    – dfeuer
    Aug 28 '13 at 2:55
  • 2
    $\begingroup$ Ah, unsurprisingly that is one of the oldest suggested ideas to deal with the problem, according to the links above. $\endgroup$
    – rschwieb
    Aug 28 '13 at 15:16
  • 2
    $\begingroup$ Just want to chime in that I have this problem. I'm not necessarily fast (and I do make mistakes -- grrr) but I've posted some answers that I thought were the best but wound up ignored at the bottom because I wasn't fast. $\endgroup$
    – Betty Mock
    Sep 3 '13 at 3:08

I have heard at least one (quality) high-rep user say that they give a quick (but correct) answer to put down some sort of "ownage" on the question. They then edit this answer until it is full and complete. I see nothing wrong with this, and it is a strategy which I occasionally adopt.

On the other hand, it annoys me greatly when a high-rep user posts an incorrect, unhelpful, or even just irrelevant answer to a question. I saw this happen yesterday - a 18k user posted a correct, but entirely unhelpful answer*. However, I commented on the answer asking "How does this help the OP?" and it got removed relatively quickly (else I would have downvoted, 'cause I'm mean like that).

I believe that this is our best weapon: comments. When your typical user101 sees an answer, no matter the rep of the answerer or the early-ness of the post, and the first comment is "Sorry, but you cannot do..." then they are not going to upvote it (or at least they will think harder about upvoting it). If the answer has garnered some upvotes, then I would also consider a down-vote (because users are jaded by existing reputation).

*If you are interested, the question is here, and the answer went along the lines of "The Klein 4-group is isomorphic to $\mathbb{Z}_2\times\mathbb{Z}_2$, and hence is abelian." Although correct, this isn't really very helpful. At all.

EDIT: I should say that I would have posted this as a comment, but I couldn't resist getting in with the first answer. +1 already, to Asaf's 0. Score!

  • 21
    $\begingroup$ That strategy is exactly the thing that I don't like. I also notice a nasty dynamic where the quick-fire answers that get "ownership" of a question repeatedly fill out with the best parts of all of the other answers. In my humble opinion this is equivalent to graffiti-ing followed by plagiarism. That's not right. You must agree? (It is acceptable if one's primary goal is to artificially inflate one's reputation, as opposed to supply high-quality answers to OPs.) $\endgroup$ Aug 27 '13 at 20:14
  • $\begingroup$ @SamiBenRomdhane As I said, I did not downvote your post. When I do downvote I comment asking for clarification or whatever. If nothing happens, only then will I downvote (unless the post has a few upvotes, then I am more generous with my downvotes). If your post had remained for a day then I would have downvoted it. Also, downvotes can be rescinded if the post is edited. $\endgroup$
    – user1729
    Aug 28 '13 at 8:04
  • $\begingroup$ As to your answer, it was not helpful because the OP give the Klein group in a specific way, while you ignored this and used an isomorphism which the OP didn't know (else the problem would have been trivial). Proving this isomorphism would have solved the question (as you show commutativity when you prove the isomorphism). So you were using the result implicitly. $\endgroup$
    – user1729
    Aug 28 '13 at 8:05
  • $\begingroup$ @FlybyNight Perhaps I have explained this badly (or perhaps not and we just disagree!). When I say they place a "quick (but correct) answer" I do not mean an incomplete one. Rather, I mean one that can be built upon and improved. Then editing takes place to build upon and improve it. So at no point was a sub-standard answer posted. This means that if you were going to post a long answer then you still get a quick post in. It becomes long in time. $\endgroup$
    – user1729
    Aug 28 '13 at 8:12
  • 1
    $\begingroup$ Also, the "ownage" sounds silly and petty, but when you sit down to write a long answer and you discover that someone better than yourself is doing the same then why waste your time? $\endgroup$
    – user1729
    Aug 28 '13 at 8:12
  • $\begingroup$ To your edit, bam wham +4 vs. your +3. $\endgroup$
    – Asaf Karagila Mod
    Aug 28 '13 at 10:09
  • $\begingroup$ @AsafKaragila +7 vs. your +5. Just sayin'. $\endgroup$
    – user1729
    Aug 28 '13 at 16:00
  • 7
    $\begingroup$ If I’m answering a question that’s just been posted, I quite often post a correct answer and then flesh it out. This isn’t so much to lay claim to the question as it is (a) to give a useful response as quickly as possible in case I can catch the asker still on-line, and (b) because when I do this, I often think of more to say or better ways to say it as soon as I’ve posted. When I’m answering a question that’s been around for a few hours, on the other hand, I tend to flesh out my answer completely before posting anything. $\endgroup$ Aug 28 '13 at 22:08
  • 1
    $\begingroup$ My approach is similar to Brian's (as is my underlying reasoning). I do see the worry that this may sometimes look like "reputation gaming", if the later additions are similar to the points mentioned in other answers. @FlybyNight: I am sure that when Brian does that he did not copy those extras from the other answers. The same applies to many other members - we do not add something to our answers we did not come up with ourselves. At least not without a reference. That is just basic academic integrity. I try to be aware of this danger as the rep game cannot be ignored either, but fails happen. $\endgroup$ Aug 29 '13 at 11:11
  • 2
    $\begingroup$ I completely agree with this. Here is a case in point: math.stackexchange.com/a/453289/53268 The answerer here, in response to a question about evaluating an integral, introduced a parameter, evaluated the general case in Maple, then set the parameter to zero. No work or effort shown. Answer of course correct. But the "technique" alluded to (again, without any evidence of trying it himself) leads to a result unique up to an integration constant, which is...the original integral! Pointing this out was unproductive, although to be fair the answerer eventually posted another solution. $\endgroup$
    – Ron Gordon
    Sep 3 '13 at 18:44
  • $\begingroup$ +1 for your editorial comment and another +1 for its irony. And did you notice that only in certain structures of StackExchange Sites, 1+x=1 ? $\endgroup$ Sep 6 '13 at 23:25

If you want to do something about it, perhaps consider posting a bounty on a question, when you see an exceptional answer. Even a small bounty, 50 points or so.

This will bump the question to the top, and if you mention which answer you are going to award this bounty to then you will invariably cause people to take interest in that answer. Moreover, when I look around a question which has been inactive for a while, I am more likely to focus on an answer with a bounty than on the accepted answer (I will skim through the accepted answer, certainly, but it's probably the case that most of my time will be spent with the bounty given answer).

If enough people do this, then the "fastest gun in the west" becomes less of an issue.

  • 12
    $\begingroup$ My point is about a systemic problem. I would bankrupt myself trying to place bounties on every question. $\endgroup$ Aug 27 '13 at 20:28
  • 5
    $\begingroup$ You do know that change often starts with one person, right? I didn't say that you should do that for every question. Do it consistently, from time to time. And urge others to follow you. $\endgroup$
    – Asaf Karagila Mod
    Aug 27 '13 at 21:49
  • 5
    $\begingroup$ I did that today, at math.stackexchange.com/questions/475255/…. I don't think it attracted much attention, but at least I got what I think is the best answer up to near the number of rep points as the others. $\endgroup$
    – dfeuer
    Aug 28 '13 at 2:58

It seems as if badges are offered for doing everything on this site, why not make some sort of badge which encourages others to go through old questions and upvote better answers?

  • $\begingroup$ But how will the system that awards badges determine if you upvoted a "better" answer? $\endgroup$ Feb 18 '16 at 0:22
  • $\begingroup$ For the cases in which it is difficult to determine which is the better answer, that should tell you that both answers are of good quality and it really doesn't matter. But there are plenty of cases where one answer is clearly better than the other. $\endgroup$ Feb 18 '16 at 0:40
  • $\begingroup$ And how would the system (a computer program) determine those "plenty of cases where one answer is clearly better than the other"? $\endgroup$ Feb 18 '16 at 3:49
  • $\begingroup$ I would say it wouldn't be the system's job, but the badge would be made available to those with sufficiently high reputation. $\endgroup$ Feb 18 '16 at 4:28

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .