Wikipedia talk:Synthesis on video games

opening the discussion

I do think there's a problem here. Maybe it's important to understand what the problem is before we get into a solution. It's reasonable to worry about the impact of any proposed solution. I don't think it's reasonable to dismiss this problem. It exists.

I happen to think that a VERY narrow exception to WP:SYN is the only logical solution. The alternative to that is strict enforcement of WP:SYN, which I believe would impose a top-down rule on the WP video game community that harms the quality of articles more than it helps. A narrow exception would be more compatible with what video game editors already do.

Arguably, this is already policy if we combine WP:IGNORE with WP:SYN. But I think a clarification of how these rules interact would be extremely helpful. It would avoid a number of edit wars. Policywonker (talk) 21:48, 19 March 2008 (UTC)

I think a key part of WP:SYN is that it is improper to synthesize data to advance one's position. For example, in video games, it would be violating SYN to state that because a game is currently the top at one of the meta-ranking sites, that it is the best game of all time.

The question is, is stating, taking the FF case, that "FF2 marks the first time that Cid and chocobos appeared" a violation of SYN. I feel that it is not for these reasons:

• Is it factual? Yes.
• Can at least some part of that statement be sourced? Yes.
• Is it a contentious fact? No.
• Is the statement trying to push a POV? Definitely not.

True, it's not the cleanest statement ever in terms of sources; you can't turn to a page in a book and says that it says exactly that. However, it is a far cry from what I've seen in NOR and NPOV in terms of whats been added. I believe the other two examples given are the same.

In other words, in lack of actual cases where a VG article has been significant affected by someone calling these type of statement as WP:SYN, I think we're fine, so I am curious if this point was brought up due to some edit that related to this. Yes, video games do have a stigma on WP in that we just don't have academic sources, period, not so much in how we approach sourcing or avoiding SYN. I think this is more a matter of making sure articles focus less on the plot and game and making sure that the development, reception, and other aspects are improved so that the articles feel less like fancruft than not, a typical problem with any contemporary work on WP. --MASEM 22:27, 19 March 2008 (UTC)

Quick comment- slightly irrelevant, but the statement "FF2 marks the first time that Cid and chocobos appeared" is actually properly sourced. The source provided explicitly states this on two of the pages in the source, just not the first one. (Guyinblack25 ^talk 22:39, 19 March 2008 (UTC))

Interesting. My bad. I still think the broader point stands, though. If need be, I can dig up more examples. Policywonker (talk) 22:49, 19 March 2008 (UTC)

No worries. Just figured you'd want accurate examples for your discussion. (Guyinblack25 ^talk 15:22, 20 March 2008 (UTC))

I would agree with your interpretation of WP:SYN, and I think this would be a great way to clarify WP:SYN if other people would agree.

Other than the examples I've offered, there's also the Artillery game article. There is no source that classifies artillery games as a subgenre of strategy game. But there are many, multiple sources that classify individual artillery games as strategy games. [1] <-- that was determined to be a violation of WP:SYN. Which is a fair interpretation of the WP:SYN policy. But removing all those sources may have been counter-productive, since the information that remains now is totally unsourced.

I think your interpretation of WP:SYN would be to put the references back in (perhaps as a bundle). But WP:SYN is not clear at this point, thus causing edit wars and potentially destructive edits. I think a clarification of WP:SYN should address situations just like this. Policywonker (talk) 22:49, 19 March 2008 (UTC)

(Slightly off-topic response) This is definitely a tricky subject. You're right that video games are lacking academic analysis and study. However, there are several examples of VG articles that have made it to FA and been able to comply with WP:SYN.
What is I think is lacking is the knowledge of how to find the necessary sources (if they are out there) and write the content in a NPOV way so that WP:SYN doesn't even enter the equation. I hate to say this, but of the 800+ members in Category:WikiProject Video games members, there is a very small amount that know how to write encyclopedic articles. And it's not like these editors are stupid, they just aren't aware of the numerous policies and guidelines (some of which are rather complicated) that need to be taken into account when writing on Wikipedia.
I know I wasn't aware of most of them until I started trying to find them. I'm sure there's still a fair amount out there that I don't know about. Hopefully some of our efforts like the newsletter will help educate members about all this. (Guyinblack25 ^talk 15:22, 20 March 2008 (UTC))

I definitely think that's an issue. But it's not necessarily the case that the wikipedians need to adapt. Perhaps wikipedia policy needs to adapt for an area with the unique qualities of video games -- which are relatively new, constantly changing, and with very little scholarly research. Wikipedia is supposed to be built bottom up, not top down. Perhaps the WP:video games community reflects the way things should be.

I also think WP:SYN is less of an issue for articles about individual games -- which can be sourced with a game review, or an instruction manual. This is much more of an issue for articles that cut across games (like genres, series, gameplay, or technologies). These articles might state well known facts about a pattern in the game industry that have not been chronicled in a reliable source.

The solution might involve a narrow exception to WP:SYN for video games (or other new phenomena). It may also revolve around why or how articles are synthesized together, which have been discussed here. That is, it is acceptable to take "A said this, B said this, C said this" and say "several scholars said this". Or to take "A has X, B has X, C has X, but D doesn't" and say "many, but not all, have X".

(BTW, I changed the bad example to a good one. I can still provide more examples, if necessary.) Policywonker (talk) 20:10, 20 March 2008 (UTC)

You do speak the truth about writing series, genre, and broader VG topics. They are more difficult to write and source; though it has been done before. I still feel that part of it is just doing the proper research. Every now and then when reading a review or feature, I'll spot a generic statement about a series or genre. You find enough small pieces and you can make a bigger picture that is properly sourced.
Personally, I think the Wikipedia community has given our video game articles a fair amount of leeway already. I'd be a bit wary of pushing it further. Besides, with video game degrees popping up more and more, I think the academic side of our sources should pick up in the future.
Having said all that, I unfortunately don't have any other solution. I do admit that writing VG articles here is very difficult and things shouldn't continue in this fashion. But I worry that quality will be sacrificed if we try to alter the guidelines to make it easier to write. (Guyinblack25 ^talk 20:31, 20 March 2008 (UTC))

Just to be devil's advocate though... policy is written from the bottom up (as it should be). And WP:SYN is ignored in the way I'm talking about already. Especially in video game articles. That is, people abide by WP:SYN when they use one source to color the interpretation of another source. But people ignore WP:SYN when they simply summarize multiple sources. The former is original research. The latter is a kind of synthesis, but can't really be called original -- there's no "inventive step". It's just a straight compilation.
See below. I think Father Goose is onto something. Policywonker (talk) 03:48, 23 March 2008 (UTC)

SYN is fine, certain interpretations of it are wrong

I've seen people invoke SYN any time a general claim is made about multiple sources, especially primary sources. That is conceivably OR (if the claim is contended) but not SYN. SYN only applies to when a claim in one source is combined with a claim in another to distort what the first one said. That is clearly what the example provided under WP:SYN is referring to.

If someone wants to make a claim, for instance, that "Kenny dies" is a recurring gag in South Park, even if one couldn't cite a secondary source supporting that claim, it's still not "SYN". It's a collective description of multiple primary sources, and trying to invoke SYN on it is just ignoring what SYN is targeted at in the first place. If people claim that it's an incorrect description, then it's just plain old OR, and still not SYN.--Father Goose (talk) 02:48, 21 March 2008 (UTC)

Thanks for checking in. This is very helpful. Just to confirm, let me try a hypothetical:

• One source says A is an X, and a Y.
• Another source says B is an X, and a Y.
• Another source says C is an X, and a Y.
• So, you state in the article "many Xs are Ys" (let alone "Xs are generally Ys"). But it's unsourced.

Would the last statement be a violation of WP:SYN? Or would that just be a healthy summary of the research out there?
If it would help, let me use a real example:

• One source says Episode 1 of Southpark involves Kenny dying.
• Another source says Episode 2 of Southpark involves Kenny dying.
• Another source says Episode 3 of Southpark involves Kenny dying.
• So, you state in the article "Kenny generally dies in every episode of Southpark" "Kenny dying is a recurring gag in southpark". But it's unsourced.

If you can confirm that... my last question is if we can clarify or amend WP:SYN to reflect this kind of example? To avoid this kind of dispute, and encourage helpful, neutral, factual research? Policywonker (talk) 03:58, 23 March 2008 (UTC)

Again, I think it depends on what X/Y are. WP:SYN makes the issue if you are trying to use such synthesis to advance a position, so if X/Y are potentially arguable statements, then this would be in voilation. But in, say, the South Park example, that fact is very far from pushing a point, and thus there's nothing wrong with applying SYN there. --MASEM 04:41, 23 March 2008 (UTC)

Wait, are you saying the South Park example should be contested as WP:SYN, or that it is a type of "synthesis" that is not a problem? (The specific way Policywonker put it, "Kenny generally dies in every episode" has problems since they dropped the gag at a certain point, but that's still not SYN, just an inaccurate description of the primary source(s). More on that later.)--Father Goose (talk) 05:20, 23 March 2008 (UTC)

Okay, I used a bad example. I guess no matter how many examples you show where "Xs are Ys"... if you can find one example of a source that says "X is not Y", then it's a counter-example. But these are fundamental challenges of human knowledge and science -- anything is vulnerable to a counter-example. Even what seems to be true today can turn out to be false tomorrow. The standard isn't truth, but verifiability.

At any rate, back to the main point... is there a way we can explain what kinds of synthesis are acceptable, and what kinds are violations of WP:OR? I think we're close to describing it: it's probably a compilation of some kind, and meets the WP:V standard in the sense that there's a good enough amount of research to support it and no known counter-examples (e.g.: the sun rises every day... so far). If we can describe what kinds of synthesis are acceptable, then I think this would be a huge help on the WP:SYN section of the WP:OR policy page. Policywonker (talk) 07:43, 23 March 2008 (UTC)

I'm saying that the South Park example is perfectly reasonable and is not an evocation of WP:SYN, since the final point "Kenny generally dies" doesn't seem like a POV-pushing statement. As a counter-example, I could list several episodes where South Park jokes at Christianity, and then come up with the conclusion "thus, the South Park producers are anti-Christian", which, of course, is very POV-ish and WP:SYN comes into play; all that I could say if there's many jokes at Christianity is that "South Park episodes frequently make jokes about Christianity". Again, the point here is that the statement that you are trying to demonstrate needs to be considered: the most opinionated it sounds, the more likely it is a violation of SYN. In many of the examples given, the statements trying to be stated are certainly far from POV pushing, and thus all should be ok. Mind you, finding sources is still the better approach, and the synthesis should only be done if you absolutely feel it is needed to provide the full context of the work, and you should always be prepared that someone may challenge that statement anyway. --MASEM 12:18, 23 March 2008 (UTC)

The Southpark example shows why we should not have an exception for video games. One would indeed violate SYN by drawing the "recurring gag" conclusion, too. The analysis of what is a recurring gag and what is simply something that occurs a few times (say, Chef wearing a hat) cannot be done by Wikipedia editors, as that would be original research. We leave that kind of analysis to reliable, secondary sources. It gets worse in video games, as the only times you'd make use of such an exception would be in contestable cases. Fighting over the details of video games belongs on forums, not on Wikipedia. Simply apply WP:IAR on really obvious cases, and prepare to be shot down for it if it turns out not to be so obvious. User:Krator (t c) 21:41, 23 March 2008 (UTC)

Isn't this covered by WP:V? Something that occurs a few times but not more would be easy to verify.

"The sun rises every day" is a good example. We can never know that it's true. But we can verify that it has been true many times. And if someone could find an example where it hasn't been true and cite it, then we'd make an edit accordingly. That's not a wikipedia problem, that's a problem as old as humanity.

Truth is not the standard for wikipedia. The standard is verifiability. "Chef always wears a hat" would be verifiable, even if we had to do some synthesis to get there. Cite a few examples where he's wearing a hat. And if someone could find a counter-example, the information would be falsified.

Of course, even something that is

Policy proposal: "Compiling several pieces of research to make a generalization -- without any inventiveness, interpretation, or insight -- should not constitute WP:OR or violate WP:SYN." -- I fully concede that this might sometimes result in a few inaccurate over-generalizations. BUT if the statement was sourced in accordance with standards of WP:V, then someone else will be able to find research that shows the generalization is false. The "Kenny dies every episode [1][2][3]" statement would then be edited to be "Kenny dying is a recurring gag in many episodes [1][2][3], although he does live through several episodes[4][5]."

Just my interpretation of WP:SYN. What do other people think? Policywonker (talk) 23:53, 23 March 2008 (UTC)

The policy I follow, and it's never failed me, is "don't write anything that's incorrect". Speculation and overgeneralizations are incorrect. Statements that are uncontroversially accurate (such as "Kenny dies is a recurring gag") are correct. Statements that are less well-known but can be backed up by a source are correct. Insisting on a source when making a correct and uncontroversial description of a primary source (which is, by implication, the source) is incorrect. If someone makes an incorrect description, delete it (or rewrite it) because it's incorrect. It's not OR, it's not SYNTH, it's just inaccurate. And if you do describe a primary source in a way that is uncontroversial, calling that OR or SYNTH is incorrect.--Father Goose (talk) 02:58, 24 March 2008 (UTC)

Policy proposal (first strawman)

It seems like we have at LEAST a strawman proposal out there. Let me try to articulate it. Feel free to criticize it constructively, so we can clarify WP:SYN as much as possible:

• Policy: Compiling several pieces of research to make a generalization is not a violation of WP:OR or WP:SYN, so long as this generalization involves no inventiveness, interpretation, or insight. This kind of generalization must still meet the standards of WP:V and WP:NPOV. If this generalization is verifiable, then it should also be possible (in theory) that someone can offer a counter-example. Speculation would violate WP:CRYSTAL.
• Acceptable Example: Three different sources about three different episodes of South Park say "Kenny dies in this episode" Three different episodes of South Park have Kenny dying. Thus, you cite the three episodes as primary sources and say "A recurring gag in South Park is Kenny dying."

Would you make any changes? What is your interpretation of WP:SYN? Policywonker (talk) 23:53, 23 March 2008 (UTC)

For the Kenny example, I feel it's not even necessary to have "sources that say it". When describing a primary source, you can work straight from the primary source (see WP:PSTS). It can be nice if you have a secondary source which backs up what you write, if for no other reason than it helps to quiet some people who are under the mistaken impression that primary sources are impermissible.

Separately, if it was a gag that only appeared a few times (such as Mr. Hankey), I'd recommend listing which episodes it appeared in, but "Kenny dies" was in almost every one of the early episodes, so I'd say it's just basic information about the entire series. It's certainly a central trait of the character, so not surprisingly, it's mentioned several times in the article Kenny McCormick, and no particular sources are provided to back up the claim -- anybody who watches the show would immediately say, "yep, that's true".--Father Goose (talk) 02:04, 24 March 2008 (UTC)

I changed the example to reflect what you just mentioned. Makes sense. Would you change anything about the policy statement made before the example? Policywonker (talk) 02:53, 24 March 2008 (UTC)

Aside: rename of this proposal is coming ...

I might move/rename this page to reflect its new scope. This isn't so much about video games anymore. It's about Generalizations. Are generalizations just good summaries of verifiable and neutral materials? Or are generalizations inventive observations, constituting a violation of WP:SYN and thus original research? I might call this WP:Generalization or WP:GEN. I'll likely be pestering you all for more feedback as I continue to make changes. So don't be shocked if there's a name change. Policywonker (talk) 00:24, 24 March 2008 (UTC)

I've been busy lately, and will likely be busy over the next month. But in that time, I expect to think about this a while. Seeing as this proposal reflects a problem (and practice) that is more widespread than video games, I'm going to broaden the rationale. I'm also going to tighten up the proposal itself, and see if I can get more feedback from a new class of users. This has been a good start. See you all later in April. Policywonker (talk) 16:48, 29 March 2008 (UTC)

Okay, see you then. Policy changes are often a long-term affair anyway; there can be a lot of inertia to overcome.--Father Goose (talk) 21:24, 29 March 2008 (UTC)