How do we improve the quality of content produced on Puzzling.SE?

I feel like it's time to open this discussion in its most general form. Countless recent meta discussions (okay, fine, they're countable) keep coming back to a fundamental issue with this site: our content quality is rapidly declining. Not all of these discussions have strictly been about the quality of content on the site, but the discussions have invariably lead to concerns about the quality and direction of the site.

In private beta, our site was very different. We generated comparably complex, expert questions that generated theoretical, interesting, and intriguing answers. We still have those, but we've recently been flooded by "solve this puzzle" questions.

In the days of yore, we discussed whether specific puzzles were on-topic at all, and came to the general conclusion that specific puzzles are on-topic, though they need answers with complete explanations. We've seen answers nowadays that are simply "Is it >! [answer] ?", and these are terrible answers. But this isn't our only problem. We later asked whether riddles are on-topic, and at this time, we seemed to be heading toward 'no,' or at least 'yes, with limitations.'

Later, Jon Ericson, a community manager, asked the question of whether chestnut puzzles are on-topic here, and this entire question and answer are worth reading. Other questions worth reading:

In short, our stance on puzzles as questions and pretty much every other quality issue can be summarized in one word: "conflicted."

I think it's safe to say that we were not the least bit prepared to handle the sudden and terrifying influx of traffic. There wasn't strong enough consensus about what is and isn't on-topic, what constitutes good answers and questions, and what voting means to us. I doubt our traffic will be spiking fifteen-fold like it did over the past few weeks anytime soon, and this is our chance to work through some of these problems.

However, we absolutely need to address the quality of questions and answers on our site. We need to promote high quality content, and turn away low quality content. We want to attract experts, and experts are being driven away.

Let's keep our precedent firmly in mind, then take it with a grain of salt. Precedent is great, but we have issues to address, and our site has changed in many ways since much of this precedent was set. Let's focus on developing solutions that are directly appropriate to the quality concerns we have right now. We can make any decisions we need to about what's on topic and what isn't, what is good quality and what isn't, but whatever we decide here needs to drastically improve site quality. If that means disallowing questions-as-puzzles, then let's do it. If that means aggressively deleting low-quality answers, let's do it. But we do need to do something, so let's figure out what that something is.

What are the first steps to raising the quality of questions and answers on our site? What specific actions do we need to take to push us in the right direction? We need to figure out what's most important, and how we hold each other accountable for upholding our new policies. What needs to happen?

• Thank you for not beating around the bush. These are questions that need to be answered for the site to survive. – xnor Nov 17 '14 at 6:55
• It's high time the community firmly sets down the rules for how to act on this site, in both importing content as well as contributing content. The past few months people have been beating around the bush, unwilling to face the idea that this site has major issues needing to be addressed. – yuritsuki Nov 17 '14 at 22:56

We must define what is on-topic in a clear unambiguous fashion. I haven't been around all that long, however every discussion I've seen about what's on-topic has been very ambiguous and I'm still not remotely clear about what the story is. Indeed, the consensus in meta is often internally inconsistent. I argue that any definition of quality will need the guidance of what's on-topic.

We are talking about 2 disparate sites. You are talking about something like CSTHEORY for puzzles ($Site1$). And others are talking a PPCG with riddles as challenges ($Site2$) (the answer that fits all the clues best (this is another matter) wins).

You say experts are leaving. The one you mentioned was a lover of $Site1$. Well, I retort that experts are coming... for $Site2$ (Mathematicians, Physicists, and Computer Scientists). This is really a debate about which of the 2 we want to be. Frankly, I want $Site2$ - however the union of the 2 would be fine for me - but I'm seeing that many people are not ok with this union; they want $Site1$ and only $Site1$. It seems clear that there is no reason for them to be leaving on this basis anyway - there are filters available for the 'riddle' tag. Regardless, the community must decide where it stands on this - not a couple of people whining about what they want. I see many people entering the discussion who don't even have a single question or answer on the site and wanting to dictate what the community should be. But, I apologise as this is besides the point.

There are people for both, some will like one and some the other. I say we should be both, some just want one. This is what the topic is about.

I, for one, will leave if $Site2$ is locked out from on-topicness. I will give all my rep to avigrail and rand al'thor over a set of bounties and then the riddlers will forever reign supreme!!!

All in all, to answer your question clearly. We need to decide which of the 2 we are, and define on-topicness in a clear unambiguous way and make it clear to new comers (as I was not long ago) so that they can enforce it and decide if they want to participate.

Let's talk about e.g. riddles being on-topic. The issue of quality is complex and requires its own specific question (i.e. when is a riddle question good quality?) - and let us please avoid "I'm on a diet therefore you can't have a cookie".

ADDENDUM: I'm not saying that all challenges that are riddles or riddle-like are gold. Nor am I saying that anything goes and I'm definitely not saying that PPCG just accepts any challenge. They (and we!) must make sure the challenges are interesting. (I love spaghetti as much as the next person but too much will kill you)

• Yes, but before we define what's good we need to define what's on topic - or in defining what's on-topic surely we'd also be defining what "good" means – d'alar'cop Nov 17 '14 at 9:18
• While both approaches seem equivalent, I think one of them will lead to a constructive solution while the other might not. We should start with the first principles of Q&A: what makes a good question? Then I think we can figure out how those apply to riddle questions, and see what works and what doesn't. – user20 Nov 17 '14 at 9:21
• OK, but what's a good question in that case is far bigger than us. And it's probably all over the main meta site, no? – d'alar'cop Nov 17 '14 at 9:22
• This is true; the definition of a 'good question' has mostly been fleshed out over the past five years. We need to figure out how it applies to our questions, though, in order to produce content that works well with Stack Exchange's core philosophies. – user20 Nov 17 '14 at 9:32
• @Emrakul I see, so really the question you want to see answered is: "How does the general SE 'good question' translate to us?". I didn't really feel quite that. However, I will stand behind the hypothesis that defining what we want to be (as a site) will answer that question for us in the most efficient and clear way. I mean, how can we define how 'good question' translates to us without knowing what we are!!? – d'alar'cop Nov 17 '14 at 9:35
• I agree with d'alar'cop. What is the point of defining a "good question" if it may end up being declared off-topic. We should define what's on-topic first. – pacoverflow Nov 17 '14 at 15:53
• I also agree with d'alar'cop and disagree with Emrakul. Plenty of things are good. Marshmallows are good. Videos of speedcubers are good. An SE needs to have a specific niche and purpose, and then judge goodness in that light. Otherwise, it will become Reddit. Apparently, "puzzles" means lots of things to different people. Goodness is also too subjective and abstract to be the first criterion, and we've seen plenty of cases where some users love a thing and others hate it. – xnor Nov 17 '14 at 18:17
• I will throw my hat into the $Site2$ ring. I came here because I wanted to make it into a space like PCG.SE but for regular puzzles. But some of the puzzles that we've been getting would never make it 5 minutes before getting closed and deleted if an analogous version were to be posted on PCG. – Joe Z. Nov 17 '14 at 19:32
• @JoeZ. I think this is where a key argument is happening in the community... poem riddles: yay or nay. Obviously I think they do have a clear definition (if they are sufficiently carefully crafted) – d'alar'cop Nov 17 '14 at 19:44
• I feel like one line we can draw is that challenge-riddles that depend on language (not just the way words are written, but semantics; such as "I can run" having two meanings in that riddle whose answer is "a clock") should probably be off-topic. The nuances of a language are too concrete that ambiguity is near unavoidable, and that different people won't understand the same riddle due to that. – Joe Z. Nov 17 '14 at 20:05
• "People who don't have a single question or answer" is not beside the point -- it's the active members who will (and should) decide what happens. On the other hand, the "sudden...influx of traffic" that Emrakul mentions may not be users who are long-term active, and thus their behavior may not be indicative of the "right" direction for the site. – jscs Nov 17 '14 at 20:09
• Let me stress that I'm speaking here as someone who is so far really just a reader of the site, but who nonetheless finds it hilariously dumb that 50% of what I read here is completely marked up with spoiler tags. As a reader, I'd lean more towards visiting to read "Site 1"-type questions, where there's something to learn. – jscs Nov 17 '14 at 20:09
• @d'alar'cop: Yeah, PPCG is surely a good precedent for "Site 2", and I've participated there from time to time. It's always been a slightly awkward fit in the SE format, though, and I guess the same thing is playing out again here. – jscs Nov 17 '14 at 20:17
• @JoshCaswell indeed lol. I was thinking that we should study PPCG and learn from its journey and see why it works. especially since it's what many people want and regardless is what is happening. – d'alar'cop Nov 17 '14 at 20:20
• @JoeZ. i feel that you are right... but here is a fresh one where OP states they were careful to make it fit well... puzzling.stackexchange.com/questions/4907/… also I think we'd lose too much by excluding such puzzles... – d'alar'cop Nov 17 '14 at 20:24

As one step in this process, we should be aggressively deleting answers that give absolutely no explanation. While normally I would say "downvote them," this has become such a rampant problem that I'd be totally in favor of deleting them.

Expert answers to expert questions. Not guesses to expert questions.

Here's a short list of answers which have absolutely no explanation associated with them (a lot of these are on one or two questions, but that's just because I'm lazy): One, two, three, four, five, six, seven, eight, nine, ten, eleven, (deleted) twelve, (deleted) thirteen, (deleted) fourteen.

Okay, so they are answers technically, but they are also very, very low quality. Each of these is a one-line answer that gives absolutely no explanation as to why it's right, and should (at least) be downvoted. That makes them simply guesses.

But because this problem is so rampant, deletion may be the preferred route. We'd want to leave a comment, letting users know why their answer was deleted, so they don't just suddenly come back to confusion. Something like, possibly:

Hello! Unfortunately, as this answer does not provide an explanation as justification, it has been removed. If you can edit it to include a complete explanation as to why your answer is correct, feel free to flag it and it may be undeleted. Thank you!

We have other categories of quality issues, too, and this isn't meant to address those. However, I (at least) find these answers terrible, and am inclined to say they should be outright removed. If this idea gets actionable support, and you see such an answer, flag it.

• How about leave a comment asking them to improve the answer, and if they don't do it within a day then delete it? After the first comment, they may learn to provide better answers and won't provide poor answers again. – pacoverflow Nov 17 '14 at 17:41
• @pacoverflow Yeah, I just had that thought and edited the answer. We'd want to leave a comment asking them to edit/improve their answer, and if they did they could flag it for undeletion. (In my experience, such a thing rarely happens, though it's always polite to extend the offer and a short explanation.) – user20 Nov 17 '14 at 17:43
• I'm skeptical that bad answers will become good if explained. Going from "The bird took a helicopter" to "By piloting a helicopter, which can typically reach speeds of 100 MPH [citation], the bird can cross the distance in 3.6 minutes without flying" doesn't change the fact that it's a cheap cop-out on a puzzle with no right answer, and one of many posted. But perhaps increasing the barrier to posting will help to deter such answers in the first place and keep such questions off HNQ. – xnor Nov 17 '14 at 18:24
• @xnor Two different issues. This particular solution isn't designed to solve that problem. – user20 Nov 17 '14 at 18:25
• @Emrakul Then I'm not understanding the reason for the proposal. Can you please give an example of something it's solving that isn't this? – xnor Nov 17 '14 at 18:26
• @xnor One, two, three, four, five (from one question); these answers have no explanation. Six, seven, eight, nine... none of these answers explain why they're right. – user20 Nov 17 '14 at 18:29
• @Emrakul Those are all on vague guess-the-thing riddles, and are still examples I believe would not become good by being explained. What would convince me of this being a separate problem would be if high-quality questions with a precise answer were receiving answers that are clearly correct except that they lack explanation. – xnor Nov 17 '14 at 18:35
• @xnor That's... fair, actually. What d'you think we should do to address this? – user20 Nov 17 '14 at 18:37
• @xnor Actually, as long as we keep seeing these answers, I think we should have a site policy on them. If we later identify and remove the source of the problem, that's great - and optimal. But in the interim, I think this would improve site quality. – user20 Nov 17 '14 at 19:22
• @Emrakul Let's continue in chat. – xnor Nov 17 '14 at 19:24
• As a user who mostly just reads the site, I was unaware of the policy against answers without explanations and have been upvoting them. I'll stop doing that immediately, of course, but I think it might be worth your while making the expectation that answers include good explanations more clear. Commenting on offending answers will help with that. It there's a meta question explaining the policy and demonstrating consensus, I think it would also be helpful to link to that discussion in comments. – Kevin Nov 19 '14 at 22:40
• I agree, but not the title of the first example, for instance. It is "Guess The Object"... it's hard to avoid guess-answers in such a case:) – klm123 Nov 23 '14 at 11:42

@Gilles @Emrakul @d'alar'cop

Well, there is a big problem behind all these discussions. Obviously, most of the time we are talking about question that have been answered long time ago or were never in need of an answer.
So just to make clear:

• Are there too many riddles? - Yes. Many riddles are crap. Frequency is too high - people's demand for attention. Sad.

• Do riddles need a definitive answer? - Subjective. Doesn't matter anyway.

• Do riddles fit the intended Q&A idea of SE? No, they don't. No matter what kind or answer. This is black or white here full stop

• Are riddles and their difficulty/clearness a matter of taste? If so, they will again absolutely not fit the Q&A scheme. Plus, riddle is riddle no matter how hard. Wrong place to discuss this.

• How can we make riddles suitable for the intended Q&A-scheme ? We can't, in my opinion. Riddles are no questions of seeking knowledge but of seeking entertainment. We need to distinguish between those two.

Now, what we have to talk about is whether to allow riddles/puzzles anyway or not? It's as simple as it is. It hurts but we have to admit that.

So I'm asking the real questions:

• How can SE provide quality?
• Is SE willing to break with the traditional style?

From my point of view, very often I found help on SE (programming an maths) and it is a great and well-arranged source of wisdom - there are a lot of intelligent people around here and I really appreciate that. So in other words: the Q&A-system is somehow self-sustaining. Questions cause answers cause competent people.
In addition to that, I love the style of SE which is working well (also with riddles). Up-voting and marking the correct answer kind of kills most of the trolls around. Thanks for that.

Riddles and breaking with the Q&A-system allows for almost anybody to talk nonsense without being wrong, technically. Which in the end does not guarantee quality. So on the one hand I don't see a reason for a pure, Q&A based riddle website. Well, there may be some people who need help with writing riddles or somebody who has to do some "de-riddle" homework. But in general I think SE will not be in need of puzzling if there is no more original content from the users around here.

On the the other hand, however, I want to ask you to think about how we could define goals together.

• So how can users, in addition to SE itself, provide for more quality?

• How can we at the same time protect 'the freedom of writing'?

I kindly want to ask for cooperation on rules for quality (if possible) and about how we can break Q&A and at the same time keep it somehow (this is not a riddle).

@d'alar'cop I hate to say it but we have almost nothing to convince them. So I'd advice to work together on a solution as mentioned before.

I thought about some solutions towards quality. On of the most reasonable ones is called "anti-bounty". People who ask questions can flag answers and mark them as "quality issue" resulting in decreasing the reputation many users are interested in sigh. Sometimes I get the impression that the reputation-rush is a big issue regarding quality.
Another possibility would be to remove some reputation for ever answer rating below zero. Hope you get my point.

More solutions appreciated.
(feel free to correct typos)

• I sort of missed the verdict. Are we saying now that maybe riddles don't really fit here :p – d'alar'cop Nov 20 '14 at 11:00
• In my opinion we have to admit that they don't fit the original SE, yes (technically). But that's no reason to close the website. We can adept. – Avigrail Nov 20 '14 at 11:02
• well that's the issue. they are not going to do that... so if it is true (and we accept) that it doesn't fit, then we are toast. I mean, if that's the case then that's that... off to cyanide.net – d'alar'cop Nov 20 '14 at 11:03
• I got the impression we can't convince them by stretching the rules. – Avigrail Nov 20 '14 at 11:05
• But I don't feel that we are stretching the rules so much as we are repairing their misconceptions... they don't seem to respect it, or think it's easy. NOT that they have answers on them... or know that it's as respectable as a cryptic crossword – d'alar'cop Nov 20 '14 at 11:09
• Maybe you're right... maybe we are just stretching the rules.. although what about codegolf.se? – d'alar'cop Nov 20 '14 at 19:05
• PCG.SE stretches the rules as well. Perhaps stretching the rules with Puzzling.SE may not have been as good an idea. – Joe Z. Nov 21 '14 at 0:44
• @JoeZ. Or maybe it's fine? I don't know honestly. I realised that a riddle is more like an event - rather than a standard question - and does remind me of a PPCG game. – d'alar'cop Nov 21 '14 at 1:19

Q1: What are the first steps to raising the quality of questions and answers on our site?

A1: Insist that a question is on-topic only if the OP does not know the answer to it.

From my experience with other stack exchange sites (PPCG is new to me), my impression was that this was the purpose and definition of a SE site. A person asks a question because they want to know something.

I also like the riddles. Since d'alar'cop has mentioned that we are looking at 2 different sites...

Q2. What specific actions do we need to take to push us in the right direction?

A2: Open a Site 2 (PuzzleChallenge.stackexchange), for the sole purpose of posing challenges.

For example, there are two SE sites for maths: Overflow and Math.SE.

The Challenge site can insist that the OP must know the answer. Quality for such a site needn't be enforced. You can just look at challenges from users with a reputation for good questions.

Alternatively, this site can have the riddles and the new one can be Site 1.