Revisiting the Identify-this-game ban

Question

I think the time is ripe to revisit our identify this game rules, following this week's discussion on chat and meta.

The status quo

The current policy is that any question asking to identify a game or modification is off-topic for this site.

The reasons brought forward behind the policy are that:

• History shows the criteria in these questions are necessarily:
  • warped by time and memory
  • possibly wrong altogether
  • insufficient to identify just one game
• From this last point it follows that the only person who can identify the one right question is the asker himself.
  • The asker might have stopped checking for questions
  • The asker might be summarily dismissing what's actually the "right" answer over a discrepancy between what he recalls and what the game actually is
  • Mind reading and 20 question guessing games are not what we're experts in
• These questions are not helpful to the internet at large
  • It is difficult to close these questions as duplicate. You might theoretically have two ITG questions with the same body and two different answers - both "correct". You also can have two different questions with the same answer.
  • These questions have lower view counts and vote counts than average
  • These questions typically do poorly in converting new users into frequent users

The proposed change

Hoping I can distill LessPop's post into a paragraph correctly, we should change our policy to:

Questions asking to identify a game or parts thereof are not too localized if they reference actual screenshots, music or other artifacts thereof. Others are too localized and must be closed accordingly.

This is okay because:

• We're not requiring vague recollections of memories, but actual evidence.
• These are not mind guessing exercises - the correct answer is not what the asker had in mind. It's about identifying elements from a game.

The problem of usefulness to the internet at large mostly remains, however. I could counter with "google reverse image search", but that's kind of stretching it.

To make everybody's life easier, while these questions still ask to identify a game, we would use the game-identification tag for these questions rather than the old and smelly identify-this-game tag. (Credit)

Topicality remains an orthogonal matter; in other words ITG can be on- or off-topic, but that must be decided on a per-question basis; this is a policy on whether or not the question is constructive. I don't think "identify-this-font" questions are on-topic for us, there's a Graphical Design Stack Exchange if you really want people to tell Helvetica apart from Arial to you. Still, this:

• removes ITG from the list of questions that automatically are off-topic.
• still leaves us with a easy to enforce, hard and fast rule on what's okay and what's not okay
• leaves the room to close a question as not constructive for other regular reasons (see the previous version of this post); this post makes a good example of an ITG that wouldn't be closed Too Localized but would be closed Not Constructive.

I for one thing this makes a ton of sense and would like to make it our policy onwards. There doesn't seem to be significant opposition on the previous question, but I expect this question title to raise more than a few eyebrows.

So, make yourselves heard: do you like this or not?

Answer 1

I'm posting this answer because I think Mark Trapps answer while excellent, gets a bit away from the core concern raised by this question in favor of talking about a tag than anything else.

So, to speak to the core question: Should we allow requests for identification of games based on the reference to actual visual, textual, or audio artifacts of that game?

I'm of the opinion that we should. I fail to see how these questions share any of the faults that led to the initial ban on ITG. They are not guessing games, and they are not one-per-answer popularity fests waiting to happen. To address the concerns raised by Matthew Read:

Q: Won't this lead to a flood of bad questions by people that don't understand the rules or the standards? Isn't this what we tried with that template for ITG questions that failed horribly?
A: First off, no, this is nothing like what we tried with that template, which merely attempted to encourage people to post their vague memories in detail. We were still playing a guessing game against bad memories. As for the possibility of bad copycat questions, I just don't really believe it. And even if that is the case, I don't really have a problem with needing to close a few more questions. Our moderation team and large number of 3k, 10k, and 20k members are more than adequate to handle the burden. Especially when there is a very clear standard of what is and is not acceptable.

Q: Isn't this just as vague and unreasonable a standard of evidence as "Only the really detailed ITG's are okay?
A: NO. I'm not sure how much clearer of a standard you can have than what is effectively Screenshots or GTFO. (To clarify, pointing to a timecode in a widely availiable video, or the song on Level X of Game is more than sufficient for what I'm talking about. The key is that an answerer should be able to access the same content and look at the same thing.)
As to the concern about how large of a sample, of a song, or how detailed a screenshot - we get vague questions every day about all kinds of subjects. We close them or keep them open based on a concensus of high rep users as to whether the question is constructive or answerable. I see no reason why our existing, and already high-standards shouldn't be applied to these questions. We actually had a question about a sound effect, (I want to say Police Quest?) that was answered almost instantly because it was sufficiently distinctive. I also think the concern is overstated, simply because you need to consider where people are going to be sourcing these artifacts from - I mean, you raise the concern of a screenshot of the sky. I counter you with in what world is someone going to have a screenshot of a blank sky, with no distinguishing features, and wonder what game it came from?

Q: My ITG was closed and it was really good!
A: I'm sorry, but this is a really petty argument, and it just holds no water with me whatsoever. You claim that this is a vanishingly small subset of ITG questions. I'd argue that you're misunderstanding the motivation behind the sort of questions with which this change is concerned. ITG's of the sort that you asked, and which prompted the blanket ban, were essentially memory exercises. It was about attempting to remember something the asker had forgotten. The questions that this change seeks to allow, are instead, truly requests for identification, in the strictest, most unambiguous sense possible. It's not about trying to track down a memory. It's about seeing a game being played in a Youtube clip and wondering what game it is. It's about hearing a variant on a games theme song and wondering which game in the series it appeared in. It's about "I have a thing here that is clearly and unambiguously a video game or from a video game, but I don't know what game it is." To be brief, it's about the difference between Identify this game and Help-me-remember this game.

You're right, the primary difference is in the quality of the evidence. Which is to say, it differs because there is any evidence at all. Which is all of the difference in the world when it comes to solving the core problems of ITG's answerability, poor compatibility with voting, and yes, even usefulness to the internet at large, as being able to point to the source of the artifact in question greatly enhances the searchability and broad usefulness of the question.

To address Mark Trapps concerns:

I think you're falling into the same trap as opponents of this change, TBQH, by getting hung up on terminology. I'm not sure why the tag matters so much. The tag as it existed was dead, and bringing it back to serve this new class of questions doesn't strike me as inherently problematic. There's nothing in Badp's proposal that in any way diverges from exactly the sort of questions that we've been discussing for several days at length now as being perfectly good and answerable. If you'd like to propose an alternative tag to avoid confusion, than by all means do so, but post-facto tagging of these questions based on their answers, as was done for the Breaking Bad question, while probably helpful for searchability, is not a viable broad solution. We need to tag these questions with something, and identify-this-game has a bunch of baggage. In the interest of clarity, game-identification is as good as anything else. A clear tag wiki and an educated base of close voters should ensure that the nightmare scenario you've described doesn't happen.

So, for those that want a tl;dr:

• These questions are good, we should allow them.
• ITG questions based on the memory of the asker are still bad, we should still not allow them.
• I like Pizza.
• What we tag them as is up for discussion, but we have to tag them with something and game-identification is as good as anything else. identify-this-game has a bunch of Meta baggage that is not worth dredging up and causing confusing meta searches to happen.

Answer 2

Opening the floodgates? I don't buy it

I feel that the argument of "If we ever allow anything even remotely approaching a question that asks for identification of a game, we're going to open the floodgates and will have to start closing thousands of questions and argue with people in meta all day long over their specific questions!" to be kind of overblown. Of course, to be fair, I just worded it in overblown way. My point is, this argument alone doesn't seem sufficient to me to think we should ban otherwise-good questions.

We want to require a good source of inquiry, not simply "any old screenshot"

A literal reading of "Screenshot or GTFO" thing misses the point too. I upvoted LessPop's answer and agree with it, but based on what I've seen in chat I think people are getting hung up on the "screenshot" idea too literally, and "Screenshot or GTFO" is not a very quotable rule. Or rather, it's very quotable, but in a way that misses the point. The Breaking Bad question is perfectly fine when you just list the show and episode number that the game was in. A screenshot or video clip of the episode in question can certainly enhance the question, but even without it you've already done your job by giving a very solid source that people can refer to when they look at your question. No, you can't just throw a screenshot into any old question and we're going to say that it's suddenly a perfectly valid game identification question. It would be really, really easy to follow the letter of whatever rules we set up for that while still providing a screenshot so bad as to be useless. For example, what game is this from?

Okay, with Google Image search, you can probably find out where I got that and reverse engineer from there, but the point is, this question is not really answerable. It's vague enough to have become a guessing game. The entire problem with ITGs was that no matter how many recollections they contained, they were guessing games, and the only one who could say if they were right or wrong was the person who asked the question.

The real problem is with questions that have answers that can only be validated by the person who asks the question

So a question that posted the above screenshot, or some other less ridiculous but still really generic screenshot, has the same problems. It's so vague as to be unanswerable. And it's not really helping the Internet be better. A screenshot might help you get the answer you're looking for (as in the case of the Breaking Bad question), but the point isn't "Hey, guys, I have this screenshot and I'm wondering what it's of". That's often not going to be an answerable question, nor will it be terribly useful to other people. However, "Hey, I was watching Breaking Bad, Season 3, episode 1, and Jessie was playing a game in it, and I'm really wondering what that game was?" is an extremely answerable question, and is likely to be a question other people have.

It seems ridiculous to me that we'd close the Breaking Bad question simply because it vaguely resembles the identify-this-game questions, and Oh no don't do that or you'll open the floodgates! Users' tiny little brains will get confused and they'll revolt!. Yes, it's asking for game identification, but beyond that, it's got a source people can go to and it's got a single answer that can be verified. I know I've made those sorts of searches all the time after watching a TV show. Sometimes, no one can figure out the answer, but that doesn't mean that it doesn't have a single, correct answer that does not require questioner's agreement to become correct.

Now, if we want to say questions such as the Breaking Bad one are off-topic because there's a Movies & TV site, fine, but again, that's just looking at one specific case. I'm sure there will be other cases where we've got a very specific source for the question that people can reference, which means we can give a very definitive answer.

The point is, let's not close good questions just because they vaguely resemble bad ones

Here's the deal. Answerable questions related to gaming are good. Unanswerable/guessing-game questions, where only the asker can verify an answer as correct (and maybe not even he can do it!) are bad. I think we're getting way too bogged down in weird special-case rules for this. Just like when Congress tries to legislate Internet policy, trying to come up with hard and fast rules and handling of corner-cases is going to be an exercise in futility. 99% of the time, it's obvious which of those categories a question falls into. Why are we pretending like no one is going to understand the distinction? Why are we so afraid of having to disagree with someone and close a question anyway in the rare case that they don't understand the distinction and want to argue about it? Isn't the entire point of having a community that we can handle things like this in a manner that isn't ridiculously bureaucratic?

If this were a SE site dedicated to music lovers, I'd love to be able to ask questions like "What was the song that played in the last 3 minutes of episode 1 season 1 of Some TV Show X?" I Google for stuff like that all the time. They have none of the problems that ITGs have. Similarly, when I saw that episode of Breaking Bad, I had the same thought about the game and after the episode Googled for the answer. These are questions people have. They often require expert knowledge. Why are we trying so hard to say no to them?

Answer 3

As I read it, you're not talking about Identify-this-game questions, as we've known and loved(loathed?) them. You're talking about a different brand of question that happens to involve "identifying" and "games" (or parts of games).

If all you're really arguing for is a change to the terminology, I'm not sure I find that request reasonable, for the simple reason that I don't think it'd result in a net increase in clarity (and I think it might do the opposite).

Answer 4

Due to a discussion in chat today, I don't believe that this exception will do much of anything beyond turn us into a glorified Google-bot for these questions.

We're going to get two types of game-identification questions:

1. A game screenshot a user has pulled off of the web somewhere.
2. A reference, link, or picture to a game being played in a TV show somewhere.

The only scenario I can realistically see where our expert knowledge would be required is in the latter.

Scenario 1 - A game screenshot a user found online somewhere
With the former, the chances of a screenshot being a standalone, without any context available whatsoever, is going to be vanishingly small. You may find a picture somewhere in a gallery of someone going, "Hey, this looks pretty cool!". At that point, without any context, all a user has is a picture. We're not going to be the first resource someone asks to play the "Identify the Screenshot!" game. It's going to be Google.

And Google allows you to drag a picture from your computer onto the image search, and it will give you not only a best guess, but also provide pictures that are visually similar. In most cases, this is going to be sufficient for someone to figure out what game that is from, all without our intervention.

More savvy Internet users are going to realize that there are also resources such as Tineye, which allow you to give it a picture, and it will bring back results of where that picture's been found online. Chances are if Google doesn't know where it's from, you can use Tineye to figure out where else it's been used, and figure out what game they're talking about from one of the original results.

As an example, I used a picture @Pixel uploaded to disprove the fact that most screenshots are going to include context from where they were pulled from. The picture can be found in chat here.

I first ran it through Tineye. It, surprisingly, doesn't return any results. Which means it was probably taken specifically to disprove my point. I could very be wrong about this, in which case I apologize.

However, taking that same picture, and uploading it to Google's Image search, gives us the name of the game as Neocron 2. Honestly, I never would have known that without someone doing the legwork for me in chat; my image searching is just giving confirmation that that's what it is.

But in the event that Google has no idea where it's from, Tineye isn't the only reverse image resource available. Googling reverse image search brings up Google's own service, Tineye, and a blog post that lists a Top 10 reverse image search engines from 2010.

Many of those links probably won't work anymore. The point I'm making is that there are resources out there, beyond Google, that will take the picture and give you results as to what it might be.

In cases like this, our expertise is hardly required, or will barely be thought of. Who wants to wait for a human answer when an automated search engine can do it for you?

In the (probably extremely rare) event that Google AND Tineye don't return anything that can be used to identify the game, chances are going to be extremely rare that a human can identify it. Google, at the very least, should not only point to the area where the user grabbed it in the first place, but also where it originally came from, and everywhere else it's been used. One of those results will tell the user what it is. The chances are vanishingly small that we can identify a picture that Google can't.

Scenario 2 - A reference, link, or picture to a game being played in a TV show or similar
I'm going to use the Breaking Bad question as an example, as it's somewhat of a unique case. If you Google breaking bad video game jesse (Thanks to @RonanForman), the first hit is to a blog, which explicitly states what game it is, and the fact that it had not been released at the time of airing.

If this question had been asked when the episode originally aired, I don't think we would have had enough information to conclusively prove (or disprove) whether or not the blog had it right. Short of someone from id dropping in and going, "Hey, yeah, that's from our new game", it probably would have been extremely difficult if not impossible to identify that.

And it would also run us straight into the talking about unreleased games problem that we can't seem to go yay or nay about. But that's a whole other can of worms, one I won't get into.

Leaving the Breaking Bad example, we will probably get lots of questions where a user can provide a clip or a point in time in a TV show, point to it and go, "Which game is that?". And for all we know, it could be a game that's not real. It could be created specifically just for that point in time. In which case the correct answer is, "It's not a real game, it's something made up for the TV show."

---

So, after this long winded explanation, I'm NOT in favor of "Screenshot or GTFO". That turns us into a Google-bot. The only exception I would be in favor of, for lack of a better term, "TV clip (link/pic/reference/whatever) or GTFO". Screenshots, Google does better than we ever could. The only time where our expertise would ever be needed is in TV clips, because Google hasn't invented a reverse movie search engine yet.

Answer 5

I don't much have an opinion on this, but it may interest people to know that SciFi.SE allows "identity this" questions, and back in March made an interesting blog post on it. Might provide new perspective, might not.