I think it is a terrible idea to revert the policy
I disagree, strongly, with the assertion that questions are so often “reasonably clear.” I feel that I have often assumptions made that are incorrect. I also flatly deny—as objective fact—the claim that there is “no harm” in answering such questions. Mistaken assumptions create an unholy mess for no good reason.
Reverting the policy is an optimization for sand
One of the core foundations of Stack Exchange is “optimize for pearls, not sand.” That means that effort is focused on ensuring that the best questions get the most attention, that the best answers get the most attention, and the people capable of providing the best answers spend as much time as possible on writing those best answers to the best questions.
Meanwhile, questions from users who don’t come back, who don’t read comments, have almost no value. To undo this policy is, to my mind, optimizing for help vampires. We don’t have many of them, but this policy is part of why. And we do have some.
The closed status is super-important here
In contrast to one popular comment here, in reality, it is closing questions that has no real harm. Yes, we want to help, we want to answer questions—but there’s not really a shortage of questions to answer.
But we need the system to be clear before a question can be answered.1 No if’s, and’s, or but’s. That’s just a fact. Every user on this website needs to know that. Every question that needs a system but doesn’t have one is a problem, a danger. In that state, it is ripe for wasted time, hurt feelings, and huge arguments. It is an excellent way to cause our best contributors to spend more time on our worst contributions.
So we close them—perfect! Now there’s no risk of those things happening. And with an engaged user, the closing doesn’t have to be long—in my experience, re-opening after a system tag is edited in happens very quickly on this site.
Closing questions protects the site and protects its users—both those who asked the question and those who would answer it.
The educational opportunity that the policy provides is real
As discussed, we really absolutely do need a clear system before any answers can be written, and every user needs to understand that. And they do, very quickly, learn that. Because of this policy.
Off the top of my head, I can only really think of one serious help vampire that we have on this site. I wouldn’t trust that user to so much as give their own question a quick once-over to make sure they haven’t left anything out or broken their links—to say nothing of actually tagging their question correctly—but that user does actually manage to use system tags correctly. Because that user knows that a missing system tag will actually get their question closed. The other myriad errors they can’t be bothered to fix, on the other hand, will usually be fixed for them by the community, and in any event will rarely inhibit their getting an answer, so they don’t bother.
Now, I am not suggesting that the community stop editing questions generally to get them into shape. This is literally one user, you don’t shape policy around one problem user. But it illustrates the power of this policy for educational purposes, and also goes on to show why it’s so important for system tags, and not other tags or other kinds of edits—because system tags are that critical, and cause that much of a problem in their absence.
Every edit could instead be a question closure followed by making the user edit things themselves; that would teach them something. The lessons so learned would not, in most cases, be worth doing that, though, so no one (least of all me) is suggesting that. But system tags are in fact unique in their importance, and therefore a case in which that educational value is uniquely high.
The Stack Exchange system isn’t helping
The precise examples given all over the place, on an ad hoc basis, I would not particularly mind admitting that they can be safely edited to include a system tag even without the querent’s interaction. But there is simply no mechanism to restrict things to only those kinds of questions. You suggest eliminating the policy—having no policy. That isn’t only going to apply to the kinds of questions you have listed—it simply isn’t.
Ideally, we would want every edit introducing a system tag (by someone other than the OP) to get attention—we would want oversight on whether or not that edit was valid. It would be great if we could make that always a suggested edit, say, or something people could vote on like closing the question is—but we cannot. That’s a pretty serious problem—because it’s very, very likely that edits are going to be missed.
Hell, for that matter, it’s not even hard to imagine the OP missing it—because there’s no way to enforce that users have to comment on the question about editing in the tag, and a new user might not notice the tag or recognize its signifcance. If the game they are playing is related to—but distinct from—the game the question was incorrectly tagged with, and they’re a new player of the game as well as a new user of our site, they may not recognize that answers are off, either—after all, if they knew the answer to the question, they wouldn’t have asked it.2
And then we are misleading users. Pretty clearly not in anyone’s best interests, least of all ours. It’s a fairly lengthy sequence of steps—I wouldn’t expect this kind of thing to happen immediately or often—but each step along the way is pretty plausible to me, which means the whole chain itself doesn’t strike me as outside the realm of possibility. And I’d very much prefer it were.
I would likely feel very different about the scenario if we could enforce oversight, enforce notification to the OP about what’s happened and what its significance is, or even better, actually limit things to certain kinds of questions, but we can’t. Stack Exchange provides no mechanism for us to do so; unless the right people happen to catch that the edit happened, it’s all too likely that we won’t even know of a problem.
The only solution to this that I see is the current policy—at least that means that everyone on the site knows system-tag edits aren’t allowed. Everyone on the site knows that if they see a system tag edited in by someone who isn’t the OP, that’s not OK. Without the policy, they won’t know that anymore. They won’t know if that was an acceptable edit or not. They’ll have to find the same evidence that whoever made the edit used, or look for a comment explaining it, or otherwise dig deeper. And they may not be qualified to do so. That’s a serious problem.
You are seriously downplaying the risks and dangers here
I see a whole lot of “preparing for the best” going on, and a whole lot of what really seems like head-in-the-sand-ism about what the worst is going to be like. This is not going to be as neat or clean as it seems that some people expect. As just discussed, we can’t restrict things to such cases, nor can we guarantee oversight to do so manually. Which means there are eventually going to be serious problems, and worse, serious problems that are allowed to stand long enough to become big fights. But the argumentation in favor of removing the policy doesn’t seem to recognize that.
For example,
Reverting the “don't guess the system” policy should not mean “guess the system at will”
except that’s exactly what it will mean—to some users. We absolutely have users who are simply vastly over-confident in their own judgment. We see users label things as “obvious” beyond any rational justification all the time. In fact, here, and really more broadly in most human discourse, “obvious” is often used as a crutch to obfuscate that the speaker doesn’t really have evidence to support their position (or can’t articulate it, but I think it’s reasonable to assume we want people to be sure enough to be able to articulate it). “Obvious” isn’t.
An example I mentioned in comments was something like
The weapons table in the Player’s Handbook includes the unarmed strike...
Someone familiar with D&D 5e might well recognize this—the issue of unarmed strikes and whether or not they are weapons and what the significance of that is are all pretty sources of confusion in that system. It might be obvious to such a user that this is a D&D 5e question. Except D&D 3.5e also has a product named Player’s Handbook, and it also includes a weapon table, and that weapon table also includes the unarmed strike. And the two systems handle unarmed strikes very differently (thanks to some errata in D&D 5e), so any answers resulting from this question are going to be very different in each system.
Certainly, not every product is titled “Player’s Handbook,” which gets re-used constantly. I am not aware of any system other than D&D 5e that has a product titled Xanathar’s Guide to Everything. But I’m not certain of that—for all I know, this Xanathar character has his roots in 2e or something, and this title reflects a 2e book I’ve never heard of. I don’t know. And without that kind of encyclopedic knowledge, it is difficult to leverage such evidence—and easy to misuse it, if you are wrong about how unique something is. More importantly, not only does whoever performs the edit have to know it, so does everyone who provides oversight—a tall order!
The long and short of it is that we are going to have blind guesses, followed by arguments that “no it really is obvious, I can just tell.” That is going to happen, sooner or later. And when it does, it’s all-too-likely that someone’s going to see the question with the tag, not question it or check the edit history—why would they?—, and write an answer. And then if it turns out to be wrong—let’s be honest with ourselves, when it turns out to be wrong, because sooner or later there will be such a question where whatever was “obviously” so wasn’t—then there is going to be a lot of wasted time, a lot of hurt feelings. That’s not good for anybody. It doesn’t help the querent, it doesn’t certainly doesn’t help the answerer, and it doesn’t help the site.
We’ll have optimized for sand, and gotten grime instead.
The arguments are going to be worse, not better
One of the stated reasons to revert the policy is to reduce arguing about the policy. That completely ignores the fact that arguments about system-less questions are going to get worse without the policy.
The simple fact of the matter is that the strict, hard-and-fast policy does nothing so well as it kills arguments. There is no room for debate—you don’t guess, period. If someone wants to argue about it, they have to argue about the policy as a whole, not the particular case. It eliminates wasting time getting into the weeds on a particular question.
Without the policy, we are going to instead of arguing about what is or isn’t obvious. Edit wars where people insert tags they consider “obvious” and others revert them, so on and so forth.
Serious danger of becoming even more “D&D 5e Stack Exchange”
As Rubiksmoose points out, 65% of our questions were about D&D 5e. That isn’t a good thing. There probably isn’t anything we can do about it, but it’s definitely worrisome—the number of D&D 5e questions is great, but the too-low number of non-5e questions is something of a concern.
Reverting this policy in light of that is only more worrisome. People incorrectly assume D&D 5e all the time—even when questions are tagged with something else. How unwelcoming would it be to have your question about another system turned into a D&D 5e question, and then for there to be a big fight when it gets changed back? How much is that going to cement the idea that we only handle D&D 5e?
In the extreme, we can even imagine us ending up with D&D 5e questions that were never meant to be D&D 5e questions. That’s only going to make matters worse.
A potential alternative—that I doubt anyone wants
The Stack Exchange system does have a solution to the problems we have here, where we could ensure oversight and consensus before allowing answers to a question. Namely, we could allow system tags to be edited in without the OP’s input after a Meta discussion agrees that the situation is “safe” to edit.
But does anyone want that? Going and creating a Meta discussion, waiting for consensus, that takes time. To be fair, there’s strong reason to wait 24 hours, to give people in all timezones a chance to contribute. In that time, hopefully, the OP would return and edit their question anyway, rendering the whole Meta discussion moot. And if they don’t, that really makes the effort put into salvaging the question pretty dubious. And it still won’t help anyone “champing at the bit,” since they’re still gonna have to wait.
So I’m not sure it’s worthwhile to even bother mentioning such an exception to the policy—would anyone even bother using it if it was an option? Is it worth anyone’s time to do so, even if they’re willing? Is it worth anyone else’s time to contribute to the consensus? I don’t think so.
Unless it’s a question that doesn’t have anything to do with a particular system but those are not the subject of this discussion.
Assuming it’s not a self-answered Q&A, which is probably a fair assumption for a new user—and even if it were, in that case there’d be an answer, and the user would presumably have enough familiarity with the system to say other answers are off.
