4
\$\begingroup\$

On a few other stacks, I've come across questions that people here would probably consider too broad, and closed accordingly. However, many of those were actually very helpful to me as a curious investigator, looking for solutions to my own problem. These questions don't typically ask for a list (like "What are all the cake recipes you know?"), but are typically fairly open-ended within some limited subspace. While I've seen a lot of these with bad answers, many have thoughtful, analytical solutions that either walk through finding the solution yourself, or explain several key possibilities.

Sometimes the same thing that worked for one situation might not work for another, and having multiple ideas or suggestions can help figure out what's going on. Other times, I'm not really sure what I need, and having a set of answers that fit a given criteria gives me things to review, and come to my own conclusion about what's best for my situation.

I'm furthermore interested in answering some of those types of questions. If someone were to ask "What can I multiclass my bard with in order to expand my spellcasting abilities?" it would be fitting to have an analysis about how class X, Y, and Z, all offer different strengths and synergies with the bard, and how to choose a class based on that. If someone asks "What 5e adventure can I run to try and scare my party with lots of undead?" an answer explaining that of the many 5e campaigns, only two really revolve around undead would be quite fitting.

I understand that they're off topic here, but I don't fully understand what makes them bad. Can I get a better explanation?

\$\endgroup\$
  • 3
  • 8
    \$\begingroup\$ It is worth noting that not every Stack Exchange site handles all issues in exactly the same way. In particular, there are Stacks that allow questions that we don’t here, that I personally think they shouldn’t allow there. There are Stack sites I avoid because I find them too “noisy” and welcoming of bad questions and/or bad answers. \$\endgroup\$ – KRyan Nov 14 '17 at 20:50
  • \$\begingroup\$ Right now the question is a bit mixed up because list and idea generation questions aren't necessarily the same thing and don't necessarily experience the same problems. Rather than asking why two separate types of questions are bad, and also rather than dealing in theoreticals, could I suggest that you: (1) roll this back to be about lists, (2) ask a separate question about idea generation questions and provide some concrete examples of questions you'd like to see stay open? \$\endgroup\$ – doppelgreener Nov 15 '17 at 13:26
  • \$\begingroup\$ You've defined "list questions" in your answer, and that definition is not something I'm interested in asking about, which is why I revised my question. Some mentions of "list question" remain, and I think we're getting caught up in mis-matched definitions. I don't care about open-ended lists, but I do care about questions that could have multiple "correct" answers. I think I'm really asking about what you're calling "idea generation" questions but within specific parameters, preventing them from being wholly open-ended. \$\endgroup\$ – inthemanual Nov 15 '17 at 17:00
  • 2
    \$\begingroup\$ @inthemanual Ok, that's fair enough. Generally huge question changes should not happen and should instead be new questions (hence my suggestion to roll this back then open a new question) but it's no big deal to clean up in this circumstance: I'll remove my answer as obsolete and I'm fine doing so since we've covered approximately the same on list questions elsewhere already, and nothing else gets affected. (There's already a new one from SSD covering your question. \o/) \$\endgroup\$ – doppelgreener Nov 15 '17 at 21:19
  • 3
    \$\begingroup\$ I'm super-tempted to edit this to read "What are some of the reasons we close questions which ask for lists of options?" \$\endgroup\$ – mattdm Nov 16 '17 at 23:52
14
\$\begingroup\$

I appreciate the value in having a “menu” of answers to read through and learn from. These can be really informative and a great way to self-learn.

However, a question generating that kind of set of answers has some inherent problems when asked here.

The problem

Generally speaking, the kind of question you're thinking about is considered incomplete here. “What can I multiclass my bard with in order to expand my spellcasting abilities?” could be a great question, but if we have to resort to guessing what might be the solution, that means the question is missing vital information for both answer-writers and voters to determine the best answer.

If voters have to guess when voting, the site's systems break: voters who refrain from voting out of doubt don't help raise up the real right answer(s), and voters who vote anyway are just contributing to what's effectively a popularity poll.

Lack of well-considered votes, and a surplus of popularity-votes, fails to fulfil the site's mission.

“But… Huh? Didn't you say those answers are valuable?”

I started by agreeing that those sets of varied answers are valuable. Why would we throw out that value just because they cause some seemingly minor problems for voting?

This does, despite appearances, come back to the same reason we reject list questions: SE wasn't designed to generate those sets of answers, while discussion forums are pitch-perfect for them. RPG.se exists in symbiosis with the other RPG help and resources sites that are our neighbours, and it throws sand in our site's gears when we try to duplicate badly what those other sites already do smoothly.

By rejecting questions that go against the system's intended purpose and fail to properly trigger its quality assurance features (voting, etc.), we ensure that this specialised site works well for the questions it was made for: the ones that don't work well elsewhere. That's our core service and mission.

So to serve our intended function and audience, we need to be a bit tough-love on questions that don't fit here, even if they seems like they'd be fun or interesting.

It also means that questions that would get better results elsewhere are more likely to be posted in those places. Those fun and interesting questions have a home on the Internet already, and should hurry to where they can flourish, instead of being stunted by our constraining mechanisms, like a square peg in a round hole.

The silver lining: We still generate these “menus” of answers, just differently and more slowly

We actually do still have a way of hosting menus of answers. Instead of being strung out under one incomplete question though, we host them by having fewer answers under a set of related, complete questions. By closing questions that are too broad — in the sense that they can only be answered by guessing at possible solutions — we encourage question-askers to add details to make it possible for voters to tell which submitted answers are actually correct.

A set of such questions about, e.g., rangers and spellcasting means that all the answers that might have been guess-submitted to a vaguer question will have homes under more specific questions. The system works at high efficiency with such questions and answers, fulfilling our mission and preventing noise from interfering with our core audience's use of the site. Browsing all our questions in a search for rangers and spellcasting will turn up the same set of answers — except that they will be well-vetted, quality answers with documented conditions under which they work and don't work.

This obviously takes longer, as it takes time for varied but related questions on a topic to be asked naturally. So the Internet is enriched by the different approaches: here there will be a curated set of quality answers that accumulates slowly, and meanwhile elsewhere on our neighbours' discussion forums there will be more raw-and-uncut discussions that happen more quickly.

\$\endgroup\$
  • \$\begingroup\$ Would a better solution than "guessing at the correct answer" or closing the question be to provide an answer that allows the querent to come to their own conclusion? An answer like "We can't know exactly, but here's how you can figure it out yourself?" Those are the kinds of answers to the questions I'm talking about that I've found helpful elsewhere. \$\endgroup\$ – inthemanual Nov 15 '17 at 19:58
  • \$\begingroup\$ I also feel like this solution provides some kind of canonical answer to the same types of questions, so that anytime someone asks "What can i multiclass X with" we can point them to a way to figure it out, instead of shutting them down and/or telling them to ask elsewhere. \$\endgroup\$ – inthemanual Nov 15 '17 at 19:59
  • 4
    \$\begingroup\$ @inthemanual Those kinds of answers are sometimes golden. We do like to see those. We have to weigh the possibility of a golden answer being submitted to a too broad question against the possibility of many bad guesses (i.e., “noise”) being submitted though, especially since in practice, the noise often drowns out and out-scores the golden answers. Sometimes a question will survive for those kinds of reasons, but often a Too Broad close is in response to existing proliferating bad answers, or because experience suggests the Q's wording has already doomed it. It's a balance. \$\endgroup\$ – SevenSidedDie Nov 15 '17 at 20:03
  • \$\begingroup\$ In practice then, it's still the responsibility of the question-writer to craft a question that attracts good answers rather than bad ones. Two questions can appear identical in purpose, but be written such that one stays open and gets those golden answers, while the other gets bad answers and close votes. \$\endgroup\$ – SevenSidedDie Nov 15 '17 at 20:07
  • \$\begingroup\$ We tend to ask for clarification or bounds on these questions, bringing them in from being too broad, but could it be proper to instead guide these questions towards "how can I find out" questions rather than "what accomplishes X" questions? \$\endgroup\$ – inthemanual Nov 15 '17 at 20:12
  • 2
    \$\begingroup\$ @inthemanual Yes, that can sometimes be a more natural revision of a question than adding more details. Especially when they don't have more details to add, “how to find out” might even be closer to their real question anyway. (Changing a question to “how can I…” is even our recommended course of revision for questions that start out as recommendation requests, which is a type of idea/list question that really didn't work well here.) \$\endgroup\$ – SevenSidedDie Nov 15 '17 at 20:16
  • \$\begingroup\$ @inthemanual No, you can't fix the question with an answer like that. You can fix a question by asking that instead, almost always! Instead of 'what class to multiclass with with Bard' ask about how to go about figuring out what class to multiclass with your bard. You'll still need to actually put the work into the question-- there are a lot of ways to figure that out after all, and you need to provide a critera by which options can be judged-- but it can work. \$\endgroup\$ – the dark wanderer Nov 16 '17 at 4:53
  • \$\begingroup\$ @inthemanual this case of a multiclass bard, that question might be another of the same variety, though, so really you want to ask "How do I go about figuring out how to go about figuring out what class to multiclass with Bard?", which is a pretty well-discussed matter of philosophy. Or, put another way, "What are the different major ways of figuring out what to multiclass my bard with and what do those differences mean?": an overview question. \$\endgroup\$ – the dark wanderer Nov 16 '17 at 4:57
  • 3
    \$\begingroup\$ There is also the formula our charop meta recommends: "I have a bard who is doing X. I want them to do XYZ. How can I build them to do that? Is there anything I should multiclass into?" (Multiclassing might be the opposite of a good idea for reaching a build they describe.) \$\endgroup\$ – doppelgreener Nov 16 '17 at 9:34

You must log in to answer this question.

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