As I'm sure everyone on Meta has read, there's been quite a lot of confusion about whether or not puzzles are on-topic on Puzzling.SE. The latest information right now is from Robert Cartaino saying that everything's OK as it is - don't panic and the rest of this post assumes that things sort themselves out and everyone just needs to calm down and be happy :-)

Please have a quick look at the State of the Puzzling.SE rules thread as it summarises a number of other things which also need discussion. I'm just focussing on 3 very basic, core things we need to discuss and decide.

The really positive thing that has come out of the busy day today is that there's been a lot of discussion on some key principles that I think really should be discussed, agreed and formalised ASAP - there are a few really important decisions we need to make as a community.

1. What is our mission statement?

We need something clear that we're working towards. If we can't agree on the point of the site, we're only going to disagree on everything else. I personally really like Bmyguest's phrasing:

We help puzzlers to become better. Better at building puzzles, better at understanding puzzles, and better at solving puzzles.

He's since posted this slightly tweaked version:

We help puzzlers to become better through practice, critique, and discussion. Better at building puzzles, better at understanding puzzles, and better at solving puzzles.

This may need a little adjusting to also include people who want to discuss puzzle theory, but whatever we go with we really need to be absolutely clear on what we're trying to do.

2. What is a good (or bad) puzzle?

I think that Travis Kindred came up with a very well-written classification of puzzles which I'll paraphrase here:

All puzzles fall into three categories, calculation, interpretation, and assumption.

The first category consists only of puzzles with nothing left to interpretation. There is exactly one answer and there can only ever be one right answer. The author is not needed to confirm it, because the answer is obviously the only possible solution once discovered.

The next category is interpretation. People are often going to guess this kind of question wrong, but sometimes in a way that is constructive and fun. A great incorrect answer will fit the majority of criterion, but never all. The correct answer will prove to perfectly fit the criterion presented with few red herrings if any. Good riddles will almost always fall into this category.

The final category, assumption, can be defined as a puzzle with a low clue to crap ratio or a puzzle whose clues do not lead to a definite answer

If we formalise these terms for puzzles (Calculation, Interpretation, Assumption) we can make Assumption puzzles off-topic, which Calculation and Interpretation are on-topic.

There is also the question of is it a puzzle or just math - I won't repeat my post, just click the link for my opinion.

3. What is a good (or bad) answer?

Obviously an answer should ... well, answer the puzzle. However, since we're not just interested in the answer, I believe we should require an explanation of the answer.

Just saying "It's 42" is not an answer.
Saying "It's 42 because when you add all the digits together, they total 42" is a much better answer.

The best answer (in my opinion) is this (not quoted because it messes up formatting):

How do I work out the answer?

The question says "xyz" which means you should add all the digits together

Right! So that means the answer is...


...although I'm fully aware that not everyone agrees. A policy on how we should (or probably more importantly, how we should not) be spoilering is important. One of the hot topics is "full-post spoilers" which are generally agreed to be bad - does anyone have any good reasons why we should continue to allow them?

  • 2
    $\begingroup$ Just a quick note. A mission statement of: "We help puzzlers to become better. Better at building puzzles, better at understanding puzzles, and better at solving puzzles." is, in my opinion, a little dangerous. HOW are you doing this? If the site will have puzzles inviting multiple answers from multiple people, is that actually making anyone better at puzzle solving? Or is simply a fun platform for puzzle solving? The New York Times wouldn't argue their crossword makes puzzlers better - rather, it just gives them enjoyment. Just something to consider... $\endgroup$ Dec 3, 2014 at 19:12
  • $\begingroup$ @AndrewMartin I believe it ties in with "What's a good answer?" - if we decide that a good answer must show its working out, then people who don't understand the puzzle can read through the methodology used to solve it. The solver doesn't necessarily get better themselves, but other people browsing the puzzle can learn from them $\endgroup$
    – Joe
    Dec 3, 2014 at 19:14
  • 1
    $\begingroup$ @Andrew Perhaps "We help puzzlers to become better through practice, critique, and discussion. Better at building puzzles, better at understanding puzzles, and better at solving puzzles." $\endgroup$ Dec 3, 2014 at 19:19
  • $\begingroup$ Hey Joe, can you break your "here's my proposal" part in to your own answer to the problem so we can vote on that as one possible solution? +1 for rightfully pointing out that the #1 take away needs to be figuring out the quality issues. Burn it all down isn't the answer, but that doesn't mean there doesn't need to be fixes made. $\endgroup$ Dec 3, 2014 at 20:00
  • $\begingroup$ @AJHenderson good thinking - that's done :-) $\endgroup$
    – Joe
    Dec 3, 2014 at 20:03
  • $\begingroup$ Do the moderators intend to set policy based on meta feedback? I ask because on any board I've ever moderated, life isn't a democracy. New rules come in. Mods enforce rules. Non-mods smile or gnash teeth, but new rules persist despite either of these reactions. I approve of the proposed mission statement, FWIW. $\endgroup$
    – COTO
    Dec 4, 2014 at 0:13
  • $\begingroup$ We don't "define" rules/law here by democracy, but "shaping" them is the democratic process. If we arrive at something reasonable and propose it, I think moderators (and community) will accept it. $\endgroup$
    – BmyGuest
    Dec 4, 2014 at 12:30
  • $\begingroup$ @COTO see theory of moderation $\endgroup$ Dec 4, 2014 at 16:42
  • $\begingroup$ @COTO Moderators are primarily enforcers of existing policy. One part of that policy comes from Stack Exchange (like the "Be Nice" policy and the "too broad" close reason), but the other part comes from the community. Stack Exchange is primarily self-moderated, so the brunt of policy shaping comes from you all, the community. $\endgroup$
    – user20
    Dec 4, 2014 at 18:07

4 Answers 4


Here's my proposal for how we should go about this - feel free to agree and disagree with and and all points :-)

Mission Statement:

I don't think I can actually improve on this very much. It emphasises that we're targeting puzzle creators, solvers and would-be solvers. It's not complicated, but it doesn't need to be; it just needs to give us a target we can all agree on.

Edit: Travis has commented with a slightly tweaked version of this - I'm including this here instead.

We help puzzlers to become better through practice, critique, and discussion. Better at building puzzles, better at understanding puzzles, and better at solving puzzles.

What's a good question?

There are two main kinds of question - puzzles as challenges, and discussion about puzzles.

Discussion about puzzles are on topic unless they fall foul of a traditional off-topic rule (such as being spam).

Puzzles as challenges must adhere to the following rules:

I would suggest we formalise Calculation, Interpretation and Assumption puzzles, then outlaw Assumption puzzles.

  • Any kind of puzzle is on-topic (including riddles) so long as it is not an Assumption riddle.
  • Puzzles must be presented as a puzzle (a pure math question is off-topic but that same math question presented as a puzzle is on-topic).
  • Puzzles must not be trivial - this is somewhat subjective, and would need to be enforced responsibly by community voting on a per-question basis.

I'd suggest something along the lines of "must not be answerable by a link to Wikipedia" as a very general rule of thumb for the last point, but this would come down to community moderation.

What's a good answer?

As before, the following conditions only apply to answers on puzzles as challenges. Questions about theory may have any answers (so long as they don't fall foul of traditional rules).

An answer must show its methodology. If you don't describe your thought process and how you solved the puzzle, your answer is not an answer.

You may clearly mark an answer as a "work-in-progress" if you're working through a complex puzzle and wish to share your progress with others.

An answer must not be accepted if it doesn't contain the actual answer to the question.

An answer may use spoiler tags, and you are encouraged to hide any outright answers in spoiler tags - use the normal plaintext for showing your working out. Examples of this can be found here and here.

An answer must not consist solely of a spoiler tag, nor something like "The answer is" followed by a spoiler tag.

  • 2
    $\begingroup$ If "getting better at puzzle creation" is a goal, then Writing may serve as an example. Critique and analysis are the subjects of questions, not left to the side channels of comments, votes, and edits. $\endgroup$
    – jscs
    Dec 3, 2014 at 20:30
  • $\begingroup$ @JoshCaswell I'm not quite sure I get what you're driving at - could you rephrase it for me please? It's getting late over here and it's been a long day xD $\endgroup$
    – Joe
    Dec 3, 2014 at 20:38
  • 1
    $\begingroup$ I will try, sure. If members of the site want the crafting of original puzzles to be a part of the site's content, I suggest that it might be best to post about that craft directly, as the thing-the-question-is-about, which is what happens on Writers. This is as opposed to posting a puzzle, expecting a solution in an answer and discussion of the puzzle's merits via secondary routes. Is that any clearer? $\endgroup$
    – jscs
    Dec 3, 2014 at 20:44
  • 2
    $\begingroup$ So one might ask "I intend to put a dead end into this triple cipher puzzle; what steps should I take to ensure that solvers understand that it's a dead end?" -- or something along those lines -- rather than posting the cipher itself and getting peripheral feedback on the trick. $\endgroup$
    – jscs
    Dec 3, 2014 at 20:48
  • $\begingroup$ Hm. I'm not sure I agree - by posing it as a challenge, you can gauge how difficult/good it is by reading through the responses (which should contain methodology) to see how people are approaching it, and whether they're going in the right direction. Then comments can give you the direct "opinions on the puzzle" feedback. I don't think limiting it to be one or the other is very helpful for an author $\endgroup$
    – Joe
    Dec 3, 2014 at 20:48
  • $\begingroup$ With your specific example, yes absolutely - that's something that is considered on-topic as a non-challenge post. It's interesting that I completely overlooked those in my proposal though... I will edit them in. They should be there $\endgroup$
    – Joe
    Dec 3, 2014 at 20:49
  • 1
    $\begingroup$ That kind of indirect feedback would indeed be very valuable for an author; my concern is that it doesn't leave behind a good artifact for posterity. It leaves a forum-like thread of you-had-to-be-there posts. $\endgroup$
    – jscs
    Dec 3, 2014 at 20:51
  • 2
    $\begingroup$ I think if you're going to post a puzzle as a challenge, it's got to be as perfect as you can possibly make it. That doesn't mean it will be perfect, and you'll get feedback in comments and whatnot regardless. But I think that a work-in-progress post is just going to leave behind a mess that won't serve the site. $\endgroup$
    – jscs
    Dec 3, 2014 at 20:55
  • $\begingroup$ @Josh Agreed on the notion of work-in-progress being bad, but only if it's left that way. A completed work can also be improved. One might ask about how to create a good puzzle or offer a puzzle they've created. Either can lead to understanding. $\endgroup$ Dec 3, 2014 at 21:17
  • $\begingroup$ @BmyGuest I'm not sure I agree - if the answer was so simple that it can only be fully spoilered then the chances are high that the question was too trivial. I'm still open to debate on that point, but I'll leave it as required just for now $\endgroup$
    – Joe
    Dec 3, 2014 at 21:33
  • $\begingroup$ @JoshCaswell I think just tagging as "theory" or something if you want the feedback as your answer should be enough. You can also make it clear in your question that you're not looking for solves. Tagging as "challenge" or similar would invite solvers, and people can still comment on ways the question can be improved. I don't think making them mutually exclusive fits the content, even if it may not perfectly fit the way SE works $\endgroup$
    – Joe
    Dec 3, 2014 at 21:36
  • $\begingroup$ @BmyGuest yes - the point of this is to give people a concrete thing we can point to and say "This post is bad - you can learn how to improve it by going here", and not spawn meta discussion of "I don't think it's that bad" etc. $\endgroup$
    – Joe
    Dec 3, 2014 at 21:38
  • 1
    $\begingroup$ @Joe The only thing I disagree with is your sample answer. My thought is the answer itself should NOT be a spoiler and should be the first thing in the post and the method behind arriving at that answer should be what is put in a spoiler tag. $\endgroup$ Dec 3, 2014 at 21:41
  • 1
    $\begingroup$ @Warlord099 that then stops someone who doesn't know the answer from working through the puzzle with their hand being held - it seriously inhibits how much they can learn because they will be looking for ways to make the answer fit the puzzle, not approaching it organically. I believe that's quite important to make the challenges valuable to other users - even if they're unable to solve it completely themselves, they can dip in part-way and then continue, having seen what they missed $\endgroup$
    – Joe
    Dec 3, 2014 at 21:43
  • $\begingroup$ @Joe But if everyone answers that way you are opening the door for many repeated answers because people failed to check through all the spoilers. And since the method behind reaching the answer is the true challenge, just seeing that Joe answered 42 is not going to spoil the problem for me. And if I'm failing at solving the puzzle on my own I would see seeing the answer as a benefit because it would allow me to potentially reverse engineer Joe's reasoning without any additional help. And then if that is not enough I can still go through all the steps in the spoilers... $\endgroup$ Dec 3, 2014 at 21:59

In order to keep the "comments" a bit shorter, I posting this as a separate answer, but in general I think we can "edit" Joe's answer into shape.

  • Could we "accept" Joe's answer to indicate that's what we are working on? (i.e. the overall consensus-(attempted)-answer?

  • As stated above, I think all "rules" should be moderated with a bit of common sense. One can nearly always construct exceptions to a rule - but humans are generally good at spotting those, so let's use our brain and don't worry too much about formalizing rules to the letter.

  • "trivial" and other "forbidden" puzzles/postings: I think we should generally use the rules/guidelines to heavily down vote unwelcomed contributions (plus giving the according comment and reference to our 'rules'.) "Closure" and "On-Hold" should really be reserved for the extreme cases.

  • Work in progress: I agree with Josh Caswell here. If one intends to put a puzzle online to have it "finished later", one might as well state that up-front and have a line "Can you help me complete this XY..." in front. One can still put the rest of the puzzle as a challenge. We might consider a "work-in-progress" tag, but that's a different discussion for later. For "ordinary" puzzle-challenge-posts it should be an overall requirement that they are "as complete as possible".

  • $\begingroup$ Just in case you're getting mixed up on the WIP part - I mentioned answers that are part-complete (such as my post here which wasn't complete, but helped everyone else jump in halfway through). You may not be, that's fine too :-) $\endgroup$
    – Joe
    Dec 3, 2014 at 21:39
  • $\begingroup$ The importance of using downvotes to mark stuff you think isn't good for the site cannot be overstated. They are just as important as upvotes. Commenting about your votes is also extremely helpful. $\endgroup$
    – jscs
    Dec 3, 2014 at 21:47
  • 1
    $\begingroup$ Unfortunately, I think downvoting is just not effective enough. If a question is +20/-15, it's probably absolutely awful, since empirically people downvote much less than they upvote. But it appears simply as +5 and gives its poster (+5)*20 + (-2)*15 = 70 rep. That's not a message that this isn't content we want. And anything that gets onto HNQ just gets loads of drive-by upvotes. So, I think closures are needed. $\endgroup$
    – xnor
    Dec 4, 2014 at 5:01
  • $\begingroup$ @xnor Well I guess you speak out of experience (and you have more than I). Then at least let's try to establish a culture of proper voting at this site. $\endgroup$
    – BmyGuest
    Dec 4, 2014 at 6:16

As my previous classification was about riddles specifically and I've had time to reflect, allow me to slightly rephrase my reasonable puzzle rules. I'm also going to add ideas from David Mulder that I've adapted. I disagree with much of what he says, but the root of his ideas may be adapted to serve the puzzle challengers and the puzzle theorists.

The three virtues of puzzling: Brevity, Mastery, and Certainty.

Brevity is simple. A puzzle must be no longer than it need be. Puzzle content can consist of only three things: The core of the puzzle itself, the minimum required narrative (with allowances made for tremendous quality), and limited misdirection. Nothing else is allowed. In answers, only the answer, the reasoning, and suggestions for improvement are allowed.

Mastery is about the overall quality of puzzling.SE, creating the best content possible. When creating a puzzle, it must be expected that improvements will be suggested. When those improvements truly better the content, the content must be amended. To help others, it is strongly recommended that answerers suggest improvements to the puzzle.

Certainty is ensuring that puzzles make sense. The intended answer to a puzzle must be obvious once discovered and explained, though some well crafted answers may rival the intended one. Because of this, any question whose eventual solution cannot be derived with mathematical certainty must not be abandoned. Hints must be given periodically and a moderator can request an answer from the puzzle creator once enough time has passed without an accepted answer. Any question supersufficiently broad, with a great many answers being of equal validity, is unacceptable. An answer that's a huge stretch or a use of lateral thinking is also unacceptable. The classifications are as follows:

  • Premise - Though not a puzzle, this is presented as one. It may be a story, a poor attempt at a riddle, an unintentionally unsolvable mathematical puzzle, or any other such nonsense. A premise does not belong within the premises of Puzzling SE.

  • Calculation Puzzle - A puzzle whose basis relies upon the laws of the natural universe, especially math or science, requiring the answerer to use their understanding of the facts of and interaction between universal constants and semi-constants to solve the puzzle. Though the puzzle creator has the right to select a correct answer, the true answer relies more upon the facts of life. Answers to calculation questions are either right or wrong.

  • Interpretation Puzzle - A puzzle reliant on the knowledge of humanity, such as language, art, and philosophy, requiring pattern recognition, inference, and deduction. This type of puzzle will have answers that vary in how much they match the criterion of the puzzle. Because of the subjective nature of this type of puzzle, the puzzle creator is the final authority on the answer. This status comes with the responsibility of nurturing the answers. Until an answer is selected, the puzzle creator must provide hints occasionally to help potential answerers achieve their goal.

  • Challenge Question - The bulk of the site's content as of the time this post was written, these are individual puzzles with the answers being revealed by clever members of the site who are not the puzzle creator. The key here is that it is a puzzle presented as a challenge.

  • Theory Question - A question about puzzles themselves and how to create puzzles well.

Apart from clarity, the main reason I present the new classifications and add David's ideas is simple. There are a number of great members of the site that feel betrayed by the shift from Theory Questions to Challenge questions. They feel like they've lost their community. By leaving Challenge Questions what they are, we respect and preserve the new way. By altering how answers work to include improvements, we honor the old guard wanting the site to remain as the beta suggested.


1. What is our mission statement?

We already have a mission statements. All Stack Exchange sites are created with one. It's the first paragraph on the tour page:

Puzzling Stack Exchange is a question and answer site for those who study the creation and solving of puzzles. It's built and run by you as part of the Stack Exchange network of Q&A sites. With your help, we're working together to build a library of detailed answers to every question about the creation and solving of puzzles.

I'm quite satisfied with this mission statement. It's a pity that we've deviated so much from it.

  • $\begingroup$ I think your answer (and its current upvotes) show part of the problem here. Its certainly not the case that the whole community wants one thing or another. It's clear that there are wildly different views on what is desired. There seems to be a lot of Meta discussion about this, with the consensus seemingly "we all have different opinions". Whilst this is obviously healthy, I really am curious to see what direction this site will now take over the next few weeks and months given the tremendous divergence in opinions. $\endgroup$ Dec 4, 2014 at 0:07
  • $\begingroup$ @gilles I think it is the last sentence of this statement where the divergence arises. Yesterday showed that 'having puzzles on site' is a (wanted by community & backed by SE) key-aspect of P.SE. This is an attempt to anchor this in the mission $\endgroup$
    – BmyGuest
    Dec 4, 2014 at 6:24
  • 1
    $\begingroup$ To phrase it differently: The existing missing statement didn't match the current state of the site. This produced a clash resulting in an attempt to bring the site back to it - and split off the (larger) rest. This was deemed (by SE) as unreasonable, so looking forward, we should come up with a concept that support "status quo" combined with a very serious attempt to improve site quality. Simply "let's go back" is not a solution ('checked'). And Simply "ignore all, keep going" is neither. ('checked') $\endgroup$
    – BmyGuest
    Dec 4, 2014 at 7:07
  • 2
    $\begingroup$ I don't think you can look at PSE today and honestly say that that mission statement is reflective of where the site has grown to - it is certainly a part of the community that we don't want to lose, but it doesn't address a sizeable chunk of the posts and people on here. Simply declaring them as off-topic caused quite the commotion yesterday, so given that we can't change the people, I think altering the MS to be more in line with what the site is used for makes sense. We can tilt it towards a target that we aim at, sure, but let's be realistic about life $\endgroup$
    – Joe
    Dec 4, 2014 at 10:07
  • $\begingroup$ I'm giving a +1 here because at least someone is representing the strict interpretation of SE. And doing so rationally. $\endgroup$ Dec 4, 2014 at 13:03
  • $\begingroup$ My -1 is not to indicate I dislike the post. It is meant as a 'vote' that I do not think the mission statement is correct (anymore). Yesterday, the situation was different. (And the conclusion was/has been to split-off.) This was revoked. $\endgroup$
    – BmyGuest
    Dec 4, 2014 at 13:11
  • $\begingroup$ @BmyGuest What do you mean in that comment? Seems I have missed out on even more than I realized~ $\endgroup$ Dec 4, 2014 at 16:44
  • $\begingroup$ @DavidMulder, if you haven't read this thread then you definitely want to do so (regardless of where on the spectrum your opinion lies, really). $\endgroup$
    – A E
    Dec 4, 2014 at 17:20
  • $\begingroup$ @AE I did, I did, I meant the part of "(And the conclusion was/has been to split-off.) This was revoked." $\endgroup$ Dec 4, 2014 at 17:34
  • $\begingroup$ @DavidMulder, I'm not following... which part did you miss? Maybe what you missed was this: chat.stackexchange.com/rooms/19142/another-home-for-challenges and these puzzling.xyz puzzling.joechilds.co.uk - but now it seems they're not necessary because puzzles-as-questions aren't banned from p.se after all. $\endgroup$
    – A E
    Dec 4, 2014 at 17:40
  • $\begingroup$ @DavidMulder I think AE has answered it for you? $\endgroup$
    – BmyGuest
    Dec 4, 2014 at 20:00

You must log in to answer this question.

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