7
\$\begingroup\$

This question exists because of this answer, and a discussion in chat that followed it. I'll get to that later.

This question is about a unique category of answers on Stack Exchange: specious answers.

Specious answers are incorrect and bad, yet look legitimate and good.

By an incorrect/bad answer, I don't mean an answer that has some inaccuracies. I mean an answer which is fundamentally and irrecoverably bad. An answer that, for instance, proposes a Fighter build that builds completely on the assumption you can use some particular abilities together - but you cannot use those abilities together, and the entire build falls apart as a result. An inaccurate answer can be corrected, but an answer like this has its problems at its foundation, and the only way to make it correct via editing is to hit Ctrl+A and then Delete.

My basic question is this:

How do we respond to answers like this on RPG.SE?

Earlier in chat, KRyan raised concerns about the monk answer. It was frequently incorrect, yet he has no means to express why. He cannot use the comment system, because the comment system is not for feedback. With the problems with the answer being subtle and frequent, an onlooker could think the answer's pretty good and upvote it.

Other Stack Exchange sites like GameDev experience these sorts of answers regularly. For instance, someone may ask "How do I program this enemy to do so-and-so?" Someone responds with some code which does the job wonderfully, and everyone upvotes it, but one person notices that in a (fairly common) situation, the algorithm will go into an infinite loop and freeze the game. The algorithm itself is not repairable, it simply needs to be thrown out and replaced with a completely different approach.

On GameDev, the typical response is for someone to just say so in a comment: "-1 This algorithm goes into an infinite loop and crashes the game when (common situation)." Readers will see the comment, go "Oh yeah", upvote the comment and downvote the answer. The day is saved.

Here, we can't leave a comment like that. The comment system is not for that. So what do we do?

  • Leave an opposing answer is a common mod suggestion. What do we do, leave an answer saying "@ThatOtherGuy's algorithm is completely wrong"? You could leave a better algorithm, and that's fine, but that won't address the fact a bad algorithm is getting upvotes. The problem here is not the lack of a contrary answer, it's conveying the incorrectness of the answer.
  • Downvote the answer? The question will get one downvote, and a comment will be made asking "Why the downvote? This answer is great!" (which will be deleted, because it's an unacceptable use of comments here)

So, how do we respond to such situations?

For the record, I don't know why the monk answer is incorrect. I am only concerned by the fact that @KRyan has no avenue to express why it is.

\$\endgroup\$

2 Answers 2

9
\$\begingroup\$

Allow negative feedback comments.

Permit us to leave comments explaining why an answer may be bad, without them getting deleted by the mods. In GameDev, doing so is an extremely valuable and acceptable usage of the comment system, and on a case by case basis, we should be allowed to do that too.


Addendum 2017: This answer was originally posted in 2013, and I don't have this stance any longer. I think it's useful if I can actually suggest something actionable, but not other cases. A non-actionable criticism comment tends to spawn arguments from community members or the post author, and those arguments inevitably need to get deleted (including the original criticism, since it'll likely just start more argument). It's better off not leaving them to begin with; choosing to not post them saves everyone a lot of time.

\$\endgroup\$
3
  • 12
    \$\begingroup\$ That's a strange interpretation of the ruling on comments. Go read it again - not just the bold headers, all the words. Comments are indeed for feedback. Downvote and put a comment explaining the issues with the answer. Just don't comment war or take up a whole page worth of comments. "-1, your fighter section is entirely based on a faulty interpretation of X and your monk section disregards X and Y." No one has a problem with that. And that's not what was happening when @KRyan was so cruelly singled out for abuse and harassment or whatever. \$\endgroup\$
    – mxyzplk
    Commented Jan 8, 2013 at 5:29
  • 1
    \$\begingroup\$ The thing about your addendum: not everything that only exists temporarily is counterproductive. Look at how holds work, for instance. Or clarifying comments that get removed once the clarification is provided. So, in the moment, the comment provides feedback. The answer author can change or delete their answer if they're convinced. Voters can see it and adjust their votes accordingly. If people argue with it instead and then the big argument blob gets deleted eventually, comment and all? It has still served its purpose in the moment. \$\endgroup\$
    – Alex P
    Commented May 4, 2017 at 18:37
  • \$\begingroup\$ (Also, fwiw, there are also plenty of smaller comment exchanges that haven't snowballed and gotten deleted on this site. Maybe that's just an oversight from a policy perspective, but realistically it does sorta imply they were harmless and not every comment chain is doomed.) \$\endgroup\$
    – Alex P
    Commented May 4, 2017 at 18:42
8
\$\begingroup\$

That answer was bad because it didn't answer the question.

A comment inviting discussion and asking for improvement is a good first step, especially to someone who has just joined the site.

My comment, for the record (as worked up in chat):

Welcome to the site! I'm finding it a little hard to see how your answer shows us how to build a better monk. While your comparative ratings are excellent, perhaps we could trim your answer down a bit to where you provide specific build advice based on cited rules? Feel free to join us in chat to workshop it.

Once the answer is specific to the question and has specific errors besides "you like monks argbughughughudgfhughd" then a single comment exploring those errors (with citations) may be appropriate.

\$\endgroup\$

You must log in to answer this question.

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