This Area 51 discussion suggests that posting brainteasers as questions should be allowed.

I think these could be one of the most, accessible, interesting possible posts to keep the community engaged. (We're here because we like puzzles, right?)

Eventually, we'll probably need to think as a group about guidelines for these.

For now, we should probably follow the normal guideline:

Don't worry too hard about restricting or regulating anything that isn't turning into a problem yet.

Eventually, there may be two potential concerns:

  1. It's important that any non-original specific brainteasers posted as questions are properly attributed whenever possible, (obvious, but less emphasized with puzzles than some fields), and
  2. We may eventually need some guidance on how to define duplicates, to prevent numerous variants of a problem type, where the variants don't introduce new complexities or make the puzzle more interesting in some way.
  • 5
    $\begingroup$ And if so, should solutions be enclosed in spoiler tags? $\endgroup$ – Michael Myers May 14 '14 at 20:59
  • 2
    $\begingroup$ @MichaelMyers, that's a great point - I think they probably should, but only for posts that are actual, specific brainteasers (As opposed to, "What are the best techniques to approach these types of problems"). $\endgroup$ – Jaydles May 14 '14 at 21:01
  • $\begingroup$ For your "100 prisoners" puzzle in particular, you linked to the solution in the question. What value do you expect to get from answers on this site? The answers to most non-original brainteasers are readily available online, so this will be a common theme. $\endgroup$ – Michael Myers May 14 '14 at 21:05
  • 1
    $\begingroup$ @MichaelMyers, I was thinking more of the benefit I'd get from someone else posting one of these that I haven't seen. $\endgroup$ – Jaydles May 14 '14 at 21:15
  • 1
    $\begingroup$ We're currently still collecting close votes on specific-puzzle questions, despite the +15 on this post. Any thoughts on how we can reverse this trend? $\endgroup$ – WendiKidd May 17 '14 at 1:44
  • $\begingroup$ @WendiKidd There is no contradiction between the high score of this question and the close votes on specific-puzzle questions. I upvoted this question because I think it's one of the most important meta issues. But my thought on the answer to this question is more and more that “solve this puzzle” should be off-topic. $\endgroup$ – Gilles 'SO- stop being evil' May 19 '14 at 14:53

I had thought from Jaydles post that this question would be on topic, but it rapidly got downvotes:

Which switch goes to which bulb?

My opinion is that this is a specific brainteaser and it should be on topic.

  • 1
    $\begingroup$ I'm with you. If the community agrees with my take above, that question should be solid (I up voted, as I remember enjoying it the first time I heard it.) If folks disagree with my take above, that's cool, too, but it'd be more helpful to hear their take here on meta. $\endgroup$ – Jaydles May 14 '14 at 21:48
  • $\begingroup$ It got three upvotes and one downvote. It may have rapidly gotten "a downvote", but certainly it doesn't seem to be frowned upon in general. $\endgroup$ – Joe Z. May 15 '14 at 0:18
  • $\begingroup$ It was more downvoted, but some definitely got changed. Looks healthier now, but did seem to be a kneejerk at first. $\endgroup$ – Rory Alsop May 15 '14 at 0:21
  • $\begingroup$ I see. Well, nothing on you, then. $\endgroup$ – Joe Z. May 15 '14 at 0:22
  • $\begingroup$ Question (slightly off-topic, perhaps) - how do you view the number of up/down votes a question has had? $\endgroup$ – Xynariz May 20 '14 at 22:53
  • $\begingroup$ Click in between the buttons for up voting and down voting $\endgroup$ – Rory Alsop May 21 '14 at 6:53
  • 1
    $\begingroup$ The reason for the down votes for this question are: 1) this is a well known question/answer already on this site 2) the user answered their own question 3) the user has a history of answering their own question 4) the user has a history of asking well known questions with answers already on this site $\endgroup$ – Trenin Sep 15 '14 at 16:54

This is kind of a meta-answer.

I don't know what the right policy should be, but the policy needs to be consistent

I don't see a difference between the two puzzles, except everyone knows what a sudoku is and not everyone has seen the first puzzle. But that is not a reason, they are both fundamentally just me posing a puzzle to the community. Whichever way we go, we must pick an answer to this question and run with it!

I also think we should try to give canonical, generic answers to challenge questions; and the question asker should themselves try to ask the generic version of traditional puzzles. I have done this twice:

This will keep the quality of content on the site high, if people not only have to be able to answer questions, but also be able to answer them for any input numbers rather than just the given ones.

  • 3
    $\begingroup$ I can see how that seeming inconsistency is frustrating. I do see one difference, although it's hard to define, which you're right is key. One of the key dangers of allowing any individual puzzles is that they could overrun everything else, which is a particular danger if they do so in a way where each one adds (or appears to add) relatively little new. It's easy to envision infinite crossword and soduku puzzles being posted, with little to differentiate each from another. You're definitely right that clarity and an open mind to changing as we go will be key. $\endgroup$ – Jaydles May 19 '14 at 15:22
  • 1
    $\begingroup$ In my opinion, the reason that the first was kept open was that the answers were discussing strategy, and how you could deductively prove the answer. The other had no answers, and no strategy comments. I do not think this is a function of the problems themselves, rather, I think this is a problem of us, as readers of the questions. Both are specific questions with only one specific answer, but we view the Sudoku as "boring", because we've done similar before. The other one, though, makes us think. Note, I'm not saying I agree with the closing, I'm saying that's why I think it happened. $\endgroup$ – Xynariz May 20 '14 at 21:24
  • 1
    $\begingroup$ I'm not sure if that is necessarily inconsistent though. While I think that specific brain teasers that are unique should be good questions. I'm not sure that a specific sudoku problem is. You can explain what the logic behind a solution to a particular brain teaser is, but a particular sudoku problem isn't significantly different from any other. The point of asking a specific puzzle is not simply to give a challenge, but to understand the solution. $\endgroup$ – AJ Henderson May 21 '14 at 18:49

I feel that puzzle questions are helpful to the community, provided the answers contain explanations to the solution. In such cases, I personally feel upvotes should not be awarded for correct answers, but for the best explanations of those correct answers.

No one benefits much from "just answers", such as "The amoeba has four hats."

On the other hand, if I am reading through an old question on Puzzling SE, even if I have no time or interest in sitting down and solving the puzzle, I can still benefit from reading the solutions with explanations; learning new problem solving methods from the explanations will help me when I try to solve other puzzles and logic problems in the future.

One thing I love is going through a question with five different answers and seeing five different ways of coming to the same conclusion, some "bruteforcing" using math or decision trees, and others with pure logic.

Anther type of answer I have seen on some Mathematics puzzles are not answers, but hints to help someone who is stuck on the puzzle; I am in full support of these as well.


So far I've seen far more puzzles posted than I had anticipated. I think we can clearly divide the questions on this site into A) questions about puzzles and B) questions that are puzzles.

While B questions will certainly keep the site interesting, if they become dominant they will drown out the more meaty A questions.

Looking into the future and extrapolating based on what I see happening to Stack Overflow and other large sites, if we focus too strongly on B questions, we will eventually see a large influx of poorly asked puzzles from new users, many of which will be copied from other places on the internet.

I think such questions should definitely be on topic, but they shouldn't be allowed to take over the site. It's still too early to say what will happen, and I see no reason to take action at this point. We will have to wait and see what happens.

  • $\begingroup$ Would you consider this question a question about puzzles or a question that is a puzzle? I think it's the former, but you might have a different opinion. $\endgroup$ – Joe Z. May 17 '14 at 14:19
  • $\begingroup$ These arguments are potentially very strong for requiring some sort of mathematical or theoretical basis to an answer, but I am not sure how much of a problem it will be. $\endgroup$ – user20 May 17 '14 at 15:02
  • $\begingroup$ @JoeZ. Yes, that question is about puzzles. It could be considered a puzzle (in the same way as you could consider most of SE's content 'puzzles'), but it is about another puzzle. $\endgroup$ – Kendall Frey May 17 '14 at 16:05
  • $\begingroup$ @Emracool 'B' puzzles can probably be divided into riddles and puzzles. Puzzles have one or a few definitive answers, whereas riddles have no logical answer. Riddles probably pose more of a threat to the site, unless they are clearly defined. $\endgroup$ – Kendall Frey May 17 '14 at 16:10

I think that posting concrete puzzles is going to harm the site in the long run. The internet doesn't need another puzzle sharing and solving site, there's plenty of forums for that around.

What does not exist is a puzzling Q&A site, and swamping this site with trivial "What's the solution to this puzzle?"-type questions is going to detract from that more useful function.

For sure, some questions about concrete puzzles should be on topic, e.g.:

  • Someone has tried and failed to solve a puzzle, and needs help with it.
  • Someone managed to solve a puzzle, but is not satisified with the solution path.

But I believe that we should avoid making this site a collection of puzzles.

  • 4
    $\begingroup$ you may be right, but it's hard to know yet. My take is that our best bet is to be watchful for actual problems. It may well be the case that specific problems overrun the site and get less interesting and weedy. In which case, we'll need to consider restricting them or making them OT. But we should start with, "If puzzlers enjoy a type of post that we can use the engine effectively for, let's let them post them, and see what works and what doesn't." $\endgroup$ – Jaydles May 15 '14 at 15:19
  • $\begingroup$ I guess it depends on whether we want this to attract experts or not. If it's just meant to be a friendly puzzle exchange thing, then sure, no harm done. But with the current list of active questions, there's a definite danger of losing puzzling experts. $\endgroup$ – user121 May 15 '14 at 16:00
  • 1
    $\begingroup$ good point. We always want to err on the side of expert appeal, since they attract the enthusiasts. In this case, I still think these may do more good than harm for two reasons: 1) Even experts in puzzling really enjoy sharing and seeing new puzzles, which we can't easily allow for without also allowing existing ones without a lot of arguing over what's "new", and 2) I'm not sure we'd have anywhere near sustainable activity without some of the specific puzzles. But I'm concerned, too - I think we'll need to keep an eye on this, and see what happens, esp. as the old classics "run out". $\endgroup$ – Jaydles May 19 '14 at 15:18
  • $\begingroup$ I think @Jaydles just touched on the key point - we will run out of these types of puzzles eventually. I agree that specific iterations of the same puzzle are duplicate, but the first question of each type should not be duplicate. As mentioned in other answers to this meta question, the puzzles should be generalized, if possible, to allow for more general answers and/or strategies. Experts on puzzling would undoubtedly have insights about how to solve these problems, and over time, I have no doubt that experts will learn such strategies from each other here. $\endgroup$ – Xynariz May 20 '14 at 20:38

I think that similar puzzles (that is, puzzles where the logic to arrive to the answer is the same) should be marked as duplicated, since they do not really add content. It would be nice that as long as it is possible the puzzles are written in a generic way (so, with N hats...) but I understand that sometimes this is annoying.


After seeing a lot of “here's a puzzle” questions being posted here, my take is that:

  • Simply posting a puzzle is not a question. A question should, y'know, ask a question.
  • Solving puzzles is of course on-topic here. The question should not be the puzzle, it should be a question about solving it.

This reminds me a lot of homework questions on many Stack Exchange sites where the asker just copy-pastes an assignment. These are often bad questions because the exercises are made-up problems which require an ad hoc series of steps to solve. A good homework question would ask for how to overcome a specific hurdle while solving a problem. That's the difference between a dump and a question: a dump is no-effort (which in itself is a reason to downvote but not to close), and does not call for a genuinely helpful answer. An answer that states “the solution is 3” may be technically correct, and answers the dump, but isn't actually useful (except to cheat on a take-home exam). On the contrary, if the question asks “Here's the exercise/puzzle: <problem description>. I tried <technique> but it didn't work because <whatever>. Is there a way to overcome this obstable?” — now we have a real question which calls for generally-applicable answers that tackle the principle that the exercise/puzzle demonstrates.

If you ask a question containing a puzzle, you should explain what you expect answers to be. If your expectation is just a solution, then the question should be declared off-topic for this site: this is a site for questions and answers, not a site for puzzle solving challenges. If your expectation is an explanation of a solution, then you should motivate the question — what is preventing you from solving the puzzle? What techniques have you already tried? If you've seen a solution, what part of it don't you understand?

These rules solve two problems that have been acute during the private beta:

  • The chestnuts problem — being the 1,000,001st site to contain a repository of solutions to well-known puzzles. Questions should be oriented towards solving techniques, not towards solution dumps.
  • Overly specific puzzles — change a few numbers and you have a new question. Again, questions should be oriented towards solving techniques — if the techniques to solve two puzzles are identical, then the questions are the same.

Don't worry too hard about restricting or regulating anything that isn't turning into a problem yet.

What was a minor nuisance (a constant stream of guess-what-I'm-thinking questions) turned into a major nuisance (an overwhelming stream of guess-what-I'm-thinking questions). Witness a large part of the meta debates in November. I wish we'd nipped this in the bud!


My opinion is similar to Mau's and Rory's, that puzzles should be allowed, but that multiple specific puzzles from the same category (e.g. Sudoku or River-Crossing) should not be allowed.

Currently, if I were to flag a question as being in the same family as another question, the only realistic option I have in the list of flags is This question has been asked before and already has an answer. I recognize that an answer can be tips on how to solve the puzzle, as opposed to a solution to the puzzle, but other users may not.

I recommend that we either clarify that text, or add a new option. For example, we could try creating a new flag type: The puzzle in this question belongs to a family of questions that already has a general answer. Then the flagger could select the question. This would allow users whose questions are flagged to know that we recognize their question, but we will not solve their specific sudoku/river-crossing-puzzle/etc. At the same time, we are trying to give them tips in the right direction.

However, this leads to another interesting problem. Some puzzles (i.e. End of the world on Sunday, or my bulb question) have one specific answer, and cannot really be generalized. On the other hand, other questions (i.e. dealing with things like sudoku and balls and scale) can be generalized. So, which ones should be generalized?

My opinion is that if a question can be generalized, it should be. Then others of the same type can be marked as duplicate. However, if the type of problem is too narrow to be generalized (such as the bulb puzzle), then one specific example should do.

Also, as others have said, let's make sure not to rush to conclusions/policies here. Yes, this question (should specific brain-teasers be allowed) is valid, but the policy we set will have a long-term effect on the site.

Also, in the comments to the question, a valid question was raised: "Should an answer be hidden in spoiler tags?" There is a good discussion about that here. (Thanks to durron597 for pointing that out.)


I think the fundamental distinction for a particular brainteaser to be on topic should be if there is benefit in an explanation to the answer and if the thought process behind the answer is unique.

A brain teaser or puzzle by itself is nice for a challenge, but not of much benefit beyond the initial challenge. Understanding how the problem is solved however, is of a lot of benefit of better understanding techniques for solving puzzles.

Thus, if a question presents a challenge that hasn't been covered yet and the answer can be explained to the unique problem, then it should be worth while. On the other hand, if it is effectively a different version or a restatement or another puzzle, then it is not of particular value and should be off topic.

There are two ways I see this working out. One is similar problems with different narrative. A problem with cats and dogs trying to cross a river is the same as cannibals and missionaries if you only change the names. The essence of the question is the same and so one should be a dupe of the other.

The other is the case of specific instances of a type of problem. For example, while one sudoku problem may be particularly difficult, it is still solved effectively the same way as any other sudoku problem if you have a comprehensive method of solving Sudoku problems, thus a specific sudoku problem would be off-topic, or perhaps more accurately a duplicate of a comprehensive "How do you solve sudoku?" question.


You must log in to answer this question.

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