Wikidata talk:WikiProject Video games

From Wikidata
Jump to navigation Jump to search
Home

The WikiProject dedicated to the world of video games!

Video game genres. Again[edit]

The fact that we have ever-growing amount of genreless games made me think practically.

I've done a run connecting Wikidata with StopGame (Q72406450) database. StopGame only uses 19 genres ("action", "add-on", "adventure", "arcade", "cards", "casino", "casual", "educational", "fighting", "for adults", "for kids", "logic", "massively multiplayer", "online", "racing", "rpg", "simulator", "sport", "strategy") which makes it easy to import, and we currently have about 2000 genreless elements that have a link to StopGame.

Problem is, this genre classification is way too generic (say, Half-Life (Q279744) would be an action game (Q270948) game, not first-person shooter (Q185029) nor shooter game (Q4282636)), and somewhat controversial (puzzle video game (Q54767) games got classified as either adventure game (Q23916) or logic video game (Q60617925), and we don't use the latter at all in Wikidata currently, we don't even have a Wikipedia article for that). More than half StopGame games are some combinations of Action, Adventure and RPG tags. I could start importing genres, but I doubt if it would be helpful.

I can express my doubts in two questions:

  1. Is it true that unaccurate / too generic genre is always better than no genre / indie game (Q2762504) only?
  2. Is it okay for element to potentially have dozen of redundant genres so we'd have to manually deprecate them?

If the answer to both questions is "yes", I can think of an ecosystem with some bots importing data from whichever source they found and the other bots deprecate redundant genres (say, deprecate massively multiplayer online game (Q862490) if massively multiplayer online role-playing game (Q175173) is set). If "no", I guess I can import some non-controversial genres like fighting game (Q846224), but I still don't see a way to solve the problem in the big picture. Facenapalm (talk) 21:25, 20 March 2023 (UTC)[reply]

Thanks for raising this.
I agree with what you say, and unfortunately I would answer "it depends". For example, I would not bat an eye at mass-adding shoot 'em up (Q1044478) even though rail shooter (Q2127647) or tube shooter (Q66230650) might be more appropriate (because shmup is not that broad and the others are quite specialized) ; but Half-Life as action game feels way too broad indeed.
I think a lot of databases or stores just do not have good enough genre vocabulary for us: most boil down to a handful of very basic genres (and sometimes supplement with free-form user tags, like Steam does). Exceptions I can think of are Glitchwave, PCGamingWiki and UVL. Perhaps IGDB and VideoGameGeek. Jean-Fred (talk) 10:40, 21 March 2023 (UTC)[reply]
There're currently 1.3k genreless games connected with VGG, 1.2k with UVL, and 0 with Glitchwave (including items with only non-gameplay genres liike indie game (Q2762504) set). Without digging into details, that alone won't let us to solve the problem mid-term.
PCGamingWiki might be fine (5.5k items), but relying on single platform-specific database clearly won't suffice.
IGDB looks good on paper (14.5k items), but it's actually hard to parse, same with MobyGames. To get first-person shooter (Q185029) for half-life, you need to extract "shooter" from Genres field and "first person" from Player Perspectives field. Sometimes you need to look at Themes as well. I'm not saying that's impossible, but that's too much work for me personally. Besides, not all of the IGDB games have genre set.

I'm quite convinced that "let's only import data that's clearly non-controvercial" approach won't help us with increasing amount of genreless games. Facenapalm (talk) 12:04, 21 March 2023 (UTC)[reply]
But each website/database may tag a game with genres that might be conflicting with each other. How would we circumvent that? --Poslovitch (talk) 09:22, 23 April 2023 (UTC)[reply]
I would also like to see a matching table first. Maybe that could also be used to clean up some of the messy genres up here. Matthias M. (talk) 18:35, 2 March 2024 (UTC)[reply]
One way might be to use depicts (P180) , for example, you might say that a game "depicts" racing, or fighting, or ninja, or puzzle or outer space. We might be less inclined to say a game "depicts" educational, or rpg, or "for adults". Incidentally, usage of depicts (P180) is used often in other Wikidata projects which we might learn from where they map media contents to tag-like entity objects. For example, you can imagine that all the ninja-themed or ninja-focused games all depicts (P180) ninja (Q9402) but all those ninja-themed games are likely in different genres, some as rpg's, some as strategy, etc. In other words, usage of depicts (P180) allows a much easier path for applying narrow or broad category tags. Use of both genre (P136) and depicts (P180) can provide richer metadata about video games and their content as well as content type/style in addition to gameplay elements, for example, "apocalypse" versus "apocalyptic". I'd avoid usage of main subject (P921) within video game instances, for now. depicts (P180) would seem to fit well in the data model under "Characters and setting". Thadguidry (talk) 01:53, 4 October 2023 (UTC)[reply]

To me indie game (Q2762504) is not a genre, at least not something that we would put into Template:Infobox video game (Q5623861) in German Wikipedia, so due to this junk value I can't really use Wikidata as a source. Can we do the distinction between AAA vs. Indie separately, please?  – The preceding unsigned comment was added by Matthias M. (talk • contribs) at 21:06, 1 January 2024 (UTC).[reply]

I saw that independent film (Q459290) is used both as a genre and instance of (Q21503252) so assume there is no real consensus yet in the realm of media to fix this. Matthias M. (talk) 10:22, 28 February 2024 (UTC)[reply]
I agree that "indie game" is indeed not a genre, nor "independent film" is (and it's actually a mistake to use latter in P31 - see de jure and de facto uses. I would have preferred them to be listed in has characteristic (P1552) as it's the best variant comparing to instance of (P31), genre (P136), form of creative work (P7937). But a detailed hierarchy has yet to be built for video games qualities that should be placed in P1552. Solidest (talk) 21:13, 28 February 2024 (UTC)[reply]
I moved dōjin game (Q906556), fan game (Q1395577) and indie game (Q2762504) to has characteristic (P1552) now. Matthias M. (talk) 20:14, 4 March 2024 (UTC)[reply]

Imported 20k new video games[edit]

Hello! Together with User:Facenapalm, I wrote a script to collect Steam IDs from IGDB that weren't already in Wikidata, and then filtered them down based on a number of criteria. Namely, excluding early access and unreleased games, excluding games with no English support, and excluding games with non-Latin characters in their titles. I think all of those - other than maybe early access/unreleased games - should be imported in the future, but for now I wanted to focus on the ones I could verify and work on, and I only speak English. The idea isn't to import every game on Steam, but to import all of those that are on both Steam and IGDB, but not in Wikidata.

Then I used Facenapalm's scripts - including one he wrote for me to make the import process a lot faster via QuickStatements - to actually create the games in Wikidata based on their Steam IDs. The result is 20,800 new video game items on Wikidata, pre-populated with release date, platforms, game mode information, supported languages, and Steam ID. We're in the process of now running further scripts to auto-import IGDB IDs, PCGamingWiki IDs, Lutris IDs, MobyGames IDs, etc. based on these Steam IDs.

If anyone notices any problems with the new items, please notify me so I can investigate and fix them! Unfortunately, there's a decent chance this created some duplicates (although we were checking for matching Steam IDs and IGDB IDs before we created any of the new items, so hopefully there are not many), and I'll be on the look-out for new uniqueness constraint violations as we import other IDs for these items.

Best, Nicereddy (talk) 19:04, 10 September 2023 (UTC)[reply]

Nice job. Given how many new games are released daily on Steam, an automated solution is a good idea. Could you link the QuickStatements batch? Matthias M. (talk) 19:51, 10 September 2023 (UTC)[reply]
@Matthias M.: Unfortunately I had to split them up into batches of ~250-300 games at a time, because otherwise QS would fail to process the batches. The batches I can see right now on my batches list were 212711, 212710, 212709, 212707, 212706, 212705, 212704, 212703, 212702, 212700, 212699, 212698, 212697, 212677, 212672, 212671, 212668, 212679, 212667, and 212664, but there were definitely more that don't show up on my list anymore. Probably more useful, I have a list of all the QIDs of the items I created here. Nicereddy (talk) 20:17, 10 September 2023 (UTC)[reply]
Nice work :) Just an idea, but if you feel like writing up the process, a quick blog post would be awesome ;) Jean-Fred (talk) 08:19, 11 September 2023 (UTC)[reply]
Quick question: as I was skimming through the (mainly bad) preliminary matches for Steam in Mix-n-match, I noticed F-16 Multirole Fighter (Q5423590) − it has a Steam ID and a IGDB ID (even two), it’s in English… How come an item was not created for it? (not that I’m complaining that it did not do a duplicate in that case ;-). Jean-Fred (talk) 08:44, 11 September 2023 (UTC)[reply]
@Jean-Frédéric: Sure, I'll consider writing a blog post about the process! :) I think the only reason it didn't create a duplicate in that case is because Wikidata prevents you from creating an item with the same name and description as another item. That's probably the only thing that protected us here :) Nicereddy (talk) 12:58, 11 September 2023 (UTC)[reply]
  • Thank you for this import, however along with the new games you have also added duplicates (just some examples [1], [2], [3], [4], [5], [6], [7], [8], [9], [10] and many others...). Can you please check where else there are such and if possible merge them with already created items? I'm afraid there could be thousands of them. @Nicereddy, Facenapalm: Please check if there are errors in the script. Before using the script further, for the time being, I would suggest suspending imports through it as it is necessary to investigate why duplicates were imported. Regards Kirilloparma (talk) 05:56, 12 November 2023 (UTC)[reply]
Here are the batches to check for duplicates: [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47], [48], [49], [50], [51], [52], [53], [54], [55], [56], [57], [58], [59], [60], [61], [62], [63], [64], [65], [66], [67], [68], [69], [70], [71], [72], [73], [74], [75], [76], [77], [78], [79], [80], [81], [82], [83], [84], [85], [86], [87], [88], [89], [90], [91], [92], [93], [94], [95], [96], [97], [98], [99], [100], [101], [102], [103], [104], [105], [106], [107], [108], [109], [110], [111], [112], [113], [114], [115], [116], [117], [118], [119], [120], [121], [122], [123], [124], [125], [126]. Regards Kirilloparma (talk) 05:56, 12 November 2023 (UTC)[reply]
@Kirilloparma 😬 that's not good, thanks for bringing this up. I'm surprised there are so many dupes, my script for collecting these games should be checking for any games with either Steam or IGDB IDs already present on Wikidata and excluding any that matched. Unless all of these items were missing both of those when the import occurred, they should've been skipped by the import process.
FWIW, the simpler list of QIDs for items imported by the batches are these three Gists: [127] [128] [129], we should probably use those as a reference to look for dupes. It's near midnight where I'm at, but I'll try to figure out what went wrong with our scripts tomorrow morning. Nicereddy (talk) 06:06, 12 November 2023 (UTC)[reply]
Hm, although having said that, I guess I only added the Steam ID-based filtering earlier today. The first two batches may not have been checking against that, and I don't know if Facenapalm's script had been checking for that. That may be the ultimate source of the problem.
Although, the unique value constraint on the Steam ID should've reported any violations that came from that mistake, and I don't see a huge increase in uniqueness violations on the database report for Steam IDs.
If the problem is that all of these items were lacking both a Steam ID and IGDB ID, and that caused us to create dupes as a result, that's a much harder problem to solve ahead of time as part of the script without pulling in a bunch of other external IDs as well :/ Nicereddy (talk) 06:18, 12 November 2023 (UTC)[reply]
Oh, and I can confirm Facenapalm's script does check for Steam IDs already in Wikidata before generating the QuickStatements creation commands. So I think this safeguard should have been fully working even without my script checking for that part of things. So we need to determine whether the problem was that the items _had_ a Steam ID and neither of our scripts managed to catch it, or that the existing items lacked a Steam/IGDB ID, and so weren't caught by our safeguards. Nicereddy (talk) 06:21, 12 November 2023 (UTC)[reply]
Checked a couple of items − looks like the original item did not have the SteamID or IGDB.
Title comparison could be another check to do. That would be a bit annoying with homonyms ; but it could/should have caught things like “DORAEMON STORY OF SEASONS: Friends of the Great Kingdom” vs “Doraemon Story of Seasons: Friends of the Great Kingdom” (exact same spelling, just different case). Things like “Demon Slayer -Kimetsu no Yaiba- The Hinokami Chronicles” vs “Demon Slayer: Kimetsu no Yaiba – The Hinokami Chronicles” are tougher, but exact match or lowercase match might bring such duplicates down. Jean-Fred (talk) 07:01, 12 November 2023 (UTC)[reply]
We can probably look through the unique value constraint violations on database reports for all the extra IDs we imported (PCGW, MobyGames, etc), and then compare the QIDs from those to the QIDs from my Gists. That should help us find any other dupes that were created in this process. I can update my script to check for homonyms as well, to try to prevent this in the future. Nicereddy (talk) 14:13, 12 November 2023 (UTC)[reply]
Ok, I investigated further and found a number of games where the QuickStatements creation failed due to Wikidata's duplicate checking. So comparing against the titles of existing games on Wikidata is definitely worthwhile, since that was hundreds (maybe over one thousand) of games that already exist and are just missing Steam/IGDB IDs.
A few examples from just going through this batch are: Alan's Automaton Workshop (Q117850916), Overlanders (Q106452970), Unheard - Voices of Crime (Q120752847), The Gallery (Q116265794), Madness Beverage (Q110400846), Redout 2 (Q113652710), Speedway Racing (Q119792408), Illusion: A Tale of the Mind (Q97344438), Dadish 3 (Q113645484), Tunnel of Doom (Q110247852), Arcade Spirits: The New Challengers (Q112946721), Fishing Adventure (Q106879791)
None of these resulted in duplicates, but also none of them had a Steam or IGDB ID, and so weren't caught by the script. Wikidata/QuickStatements prevented them from being created via their own duplicate checking based on name matches. That said, it proves that we should probably protect against this given how many of these there were just from one of the batches.
I'm going to write a script to pull the names of every game on Wikidata and check for dupes amongst the names of games created as part of these imports. Based on the safeguards we already implemented (Wikidata's dupe checking + Steam/IGDB checking), I think there probably aren't that many (at least, that haven't been fixed already). Nicereddy (talk) 16:25, 12 November 2023 (UTC)[reply]
While adding missing Steam IDs, I noticed that there are around 60 games with a Steam Deck compatibility status and no top-level Steam ID. @Kirilloparma would you mind adding the Steam IDs to those items via OpenRefine (or a script, whatever you like)? SPARQL Query here Nicereddy (talk) 17:14, 12 November 2023 (UTC)[reply]
Ok, initial run of my script gets me about 825 potential duplicates. A chunk of these are just games with generic names ("basketball", "pinball", etc), but a lot are actual duplicates.
I'm going to improve the script formatting, pull in some extra metadata, and start de-duping. Nicereddy (talk) 19:49, 12 November 2023 (UTC)[reply]
I think that before checking an item for Steam ID, we should first check the existence of the item itself, as Special:NewItem suggests, and then apply the criterion for checking Steam ID. Also, as it was mentioned here, it is desirable to take into account both upper and lower case titles, which may help. Regards Kirilloparma (talk) 04:22, 13 November 2023 (UTC)[reply]
Ah, I see that you've already started checking and improving the script. Nice! :) Regards Kirilloparma (talk) 04:24, 13 November 2023 (UTC)[reply]
> While adding missing Steam IDs, I noticed that there are around 60 games with a Steam Deck compatibility status and no top-level Steam ID
@Nicereddy: ✓ Done, see toollabs:editgroups/b/OR/6e996953768. Regards Kirilloparma (talk) 04:29, 13 November 2023 (UTC)[reply]
Assuming this is one of them. [130] When you import re-releases, don't set the Steam release as only publication data. A 1987 game is now marked as 2023 in all supported languages thanks to mindless bots. Matthias M. (talk) 17:49, 5 January 2024 (UTC)[reply]
Weirdly, I think this was part of an initial test batch I did before Facenapalm created the QuickStatements converter script. So, essentially part of this mass-import but created in a slightly different way.
Unfortunately there's not a very good way to detect re-releases like that (short of filtering out any games with a 4 digit number in the description, I guess? But that'd obviously have a ton of false-positives, and miss some re-releases as well). Interestingly, since this "contains both the Amstrad CPC and ZX Spectrum versions of the game", there's probably an argument to be made that this is its own work (assuming they're not separate executables, maybe they are?). Nicereddy (talk) 18:33, 5 January 2024 (UTC)[reply]
Alternatively, one might argue that this Steam entry is not a game but a compilation. Some databases, like OGDB, consider that compilations do not compile games, but game releases − so in this case two releases of the same game. (We do not consider releases (yet) so I’m not really advocating for this! ;) Jean-Fred (talk) 21:33, 5 January 2024 (UTC)[reply]
The Killing Cloud (Q122373507) is another one. Marking a 1991 as 2021 makes matching difficult. Matthias M. (talk) 18:36, 7 January 2024 (UTC)[reply]

Metacritic (IMPORTANT UPDATE)[edit]

applsdev Arlo Barnes BugWarp Coloradohusky CptViraj Cupkake4Yoshi Cwf97 Cynde Moya Danrok Datumizer Dexxor Diggr Dispenser Dollarsign8 DoublePendulumAttractor EdoAug Edolusill Eniehack Facenapalm Floyd-out FullyAwesome Harshrathod50 Jean-Frédéric Jotamide Keplersj Kirilloparma Lewis Hulbert LotsofTheories Macocobovi Macrike Master Of Ninja Matthias M. Metafire18 Nicereddy Nw520 Oduci Poslovitch Rampagingcarrot RampantSpirit Sanqui Santer Sight Contamination Sir Lothar thgiex Tomodachi94 VGPaleontologist Wd-Ryan WikiSyn YotaMoteuchi

Notified participants of WikiProject Video games

While browsing the site for the sake of interest, I went into its source code, although not immediately and rather late, but still noticed that Metacritic not only updated the database and stopped using the platform format, but also, as it turns out, moved to numeric identifiers. See the following examples:

URL slug → hogwarts-legacy; numeric ID → 1300523817
URL slug → resident-evil-4; numeric ID → 1300571786
URL slug → baldurs-gate-3; numeric ID → 1300501848
URL slug → starfield; numeric ID → 1300486989
URL slug → cyberpunk-2077-phantom-liberty; numeric ID → 1300577900

Note: These numeric IDs can be found by the productId: or title:{id: string using Ctrl+U key combination on the page you are interested in. It is noteworthy that the first numeric ID starts (if I'm not mistaken) with 1300000001, and the current last one at the moment of writing this thread ends with 1300610109. I think that such a significant number gap is to avoid confusing human-readable identifiers with numeric ones (example [131], [132]), and probably also to avoid reusing them.

This literally changes everything. I would of course suggest that as soon as a new property is created, we migrate to numeric identifiers as we did with MobyGames, but it is important to note the main drawback of numeric identifiers so far. On the game page with numeric ID unlike the human-readable one, for some reason not all things are displayed. For instance, in the top right corner you can see only user reviews and when you go to the critic reviews in the View All Platforms section it's completely empty, although they are available by non-numeric identifier. I don't know what this has to do with, but I think that the database is still being updated, so it's not completely implemented yet. I suggest that while the Metacritic staff is still working on the database, we should update the identifiers (i.e. switch to the new format) of the current broad property and use it as the main for now. Once the numeric identifiers are okay, we'll move to the new property. In other words: create a property for numeric identifiers (qualifier) and use the property with URL slugs (main) for now as in the case of IGDB (example).

There is another option, which in my opinion is the most optimal. Since we decided to get rid of the general Metacritic ID (P1712) property and not to use it for video games, we can create two separate properties: one for human-readable identifiers (already proposed) and one for numeric identifiers. What do you think? Regards Kirilloparma (talk) 02:55, 25 September 2023 (UTC)[reply]

We could handle it like Internet Game Database game ID (P5794) which is matched and Internet Game Database numeric game ID (P9043) which is added by bots/scripts. Matthias M. (talk) 07:57, 25 September 2023 (UTC)[reply]
Also all the Metacritic identifiers that links to video game companies have gone dead. Trade (talk) 13:59, 25 September 2023 (UTC)[reply]
Joy. Thanks for flagging this. At this point, I would borderline consider a retrospective and posthumous “Metacritic company ID”, move all IDs to it, and change the formatter URL to Wayback Machine. Jean-Fred (talk) 15:26, 25 September 2023 (UTC)[reply]
(We have 472 IDs with `company/`, it’s not nothing). Jean-Fred (talk) 15:28, 25 September 2023 (UTC)[reply]
> We could handle it like Internet Game Database game ID (P5794) which is matched and Internet Game Database numeric game ID (P9043) which is added by bots/scripts.
@Matthias M. Yeah, that's exactly what I was talking about. I'm going to propose a separate property for numeric IDs, so here it is. In the end I think it would be much easier to maintain both properties for video games rather than messing around with a generic Metacritic ID (P1712) property.
> Also all the Metacritic identifiers that links to video game companies have gone dead.
@Trade Yes, unfortunately Metacritic decided to remove the companies for some reason, however the good news is that we still have over 10,000 archived URLs left, so nothing to worry about.
> At this point, I would borderline consider a retrospective and posthumous “Metacritic company ID”, move all IDs to it, and change the formatter URL to Wayback Machine.
@Jean-Frédéric Agreed. In this case it really makes sense to create such a property since company IDs are well archived. Regards Kirilloparma (talk) 23:32, 25 September 2023 (UTC)[reply]
Will you propose the property? Trade (talk) 02:21, 30 September 2023 (UTC)[reply]
Sure, ✓ Done: Wikidata:Property proposal/Metacritic company ID. In the meantime, let's also create a proposal for gaming publications while we're at it. Regards Kirilloparma (talk) 18:26, 30 September 2023 (UTC)[reply]
Before switching to the new property, I started correcting the current game IDs to the new scheme (example). Once the new property has been created, we will move the corrected IDs into it, but this time without the game/ prefix. Regards Kirilloparma (talk) 15:26, 1 October 2023 (UTC)[reply]
I have completed update to the new format (see toollabs:editgroups/b/OR/bf12b3e0e36), the only thing left is for one of the property creators to create a property (I can't since I proposed it). @Jean-Frédéric, Trade: If it's not too much trouble, can you please create a property? It won't even take 5 minutes as I have already prepared a ready to run QS file, see Wikidata:Property proposal/Metacritic game ID#For property creators. Regards Kirilloparma (talk) 17:53, 2 October 2023 (UTC)[reply]
Trade has created Metacritic game ID (P12054), now all that remains is to create a property for the numeric ID (qualifier for P12054) that was marked as ready. Here is the QS batch ready to run for it. Regards Kirilloparma (talk) 00:52, 3 October 2023 (UTC)[reply]
You’re moving from the old value? I thought you were going to copy, so that downstream reusers have time to adapt. I assume some Wikipedias are currently using Metacritic ID (P1712) Jean-Fred (talk) 08:31, 3 October 2023 (UTC)[reply]
As I mentioned earlier in this discussion, I think that potential reusers will follow our approach and start to gradually, as it is happening with Mobygames, move from a broad to a specific property. To encourage folks to use the new property, I've gone through various sections of Wikipedia and replaced the old property with the new one in the {{Metacritic video game}} and other templates/modules (example), thereby letting everyone know that there's now a separate property for games. I tweaked and updated ([133], [134], [135], [136], [137]) these templates to make them work properly (see en:God of War Ragnarök#External links, sr:Hogwarts Legacy#Спољашње везе, no:Mario Bros.#Eksterne lenker, ko:호그와트 레거시#외부 링크). I have also submitted a request on ruwiki in order to start using the new property in the authority control template and it should be added to the module in a couple of days (if there are no objections). Participants in the various Wikimedia projects can do the same, which I strongly recommend, since the generic property should no longer be used for games, although it is still allowed to be used in references for review score (P444) property, but no more than that. By the way, I've already added a complex constraint to track values of P1712 which aren't used in references. Regards Kirilloparma (talk) 20:40, 6 October 2023 (UTC)[reply]
You have been extremely thorough (way more than I would have managed to be if I had tried to tackle that!) :) That sounds great. Jean-Fred (talk) 06:46, 7 October 2023 (UTC)[reply]
Thank you! :). Regards Kirilloparma (talk) 04:31, 10 October 2023 (UTC)[reply]
@Kirilloparma Please don't delete Metacritic ID (P1712) just yet [138] because it is still in active use at Template:Metacritic (Q114776504) so there is a chance you will destroy Wikipedia articles at least in German Wikipedia. Matthias M. (talk) 17:44, 12 January 2024 (UTC)[reply]
Well, the Metacritic identifier has not been deleted in this case, but moved to a new property for video games.
I can certainly pause for now and not move the replaced property identifiers, however please note that Metacritic ID (P1712) property is no longer used and should not be used for video games (see this note) as of September 2023. I strongly discourage you to resume using it as you did here, instead please use the separate one for video games (together with complementary Metacritic numeric game ID (P12078) property, it's use is optional, but helpful). As for the German Wikipedia, you should create a new separate Metacritic template for video games and use exactly P12054 instead of P1712. Regards Kirilloparma (talk) 23:30, 12 January 2024 (UTC)[reply]
I can probably migrate German Wikipedia, but there is already a strong sentiment against using Wikidata because automated destructive actions are possible, so you are not making my life easier. Dutch Wikipedia is also affected. Matthias M. (talk) 10:28, 13 January 2024 (UTC)[reply]
@Ivan A. Krestinin: who operates User:KrBot which also deletes Metacritic IDs which causes disruption in downstream Wikipedia. Matthias M. (talk) 11:26, 14 January 2024 (UTC)[reply]
@Matthias M. I think KrBot is mostly automated, we can probably remove something from the constraints/talk page of the Metacritic ID property and it will stop trying to fix this. Nicereddy (talk) 17:09, 15 January 2024 (UTC)[reply]
Stopping it would be as easy as removing the first two {{AutoFix}} templates on the talk page, but I don't really think this is a good idea. The IDs have already been migrated to Metacritic game ID (P12054) for a few months now, we don't really want to migrate nearly 10,000 IDs backwards and just end up with a ton of redundancy, so I'd think the best course of action is to migrate any wikis still using Metacritic ID (P1712) to the new game subproperty. Lewis Hulbert (talk) 17:35, 15 January 2024 (UTC)[reply]

Counter-Strike Xbox Edition[edit]

Would anyone mind if i split the Xbox Edition of CS into it's own seperate article? The Counter-Strike Wiki consider them two different games Trade (talk) 16:50, 10 October 2023 (UTC)[reply]

Sounds good to me - assuming that wiki is correct it's actually derived from Counter-Strike: Condition Zero (Q847820) rather than just being an Xbox port of Counter-Strike (Q163628). Problem is most external IDs just lump the Xbox edition together with CS because they share a name despite effectively being different games. Lewis Hulbert (talk) 03:49, 11 October 2023 (UTC)[reply]

How do we model fictional characters that have not only voice actors, but also people who have provided their face, body or something else for them?[edit]

I have been wondering about this for a long time and now I will give some concrete examples. Let's say we have a specific game, and there is a corresponding voice actor for it, but the character's face itself is based on a completely different person. So, what is the best property to use here and do we need some new separate one? For instance, in this case I tried to use inspired by (P941) although I am not completely sure that it fits here and that is why I decided to open a thread so that we can use the exact modelling scheme in the future.

Here's another good example. We have a character Abby (Q63190645) that was built on 3 different people:

  1. The motion capture and the voice acting was done by Laura Bailey (Q2364867).
  2. The face of the character is based on Jocelyn Mettler.
  3. The body of this character was provided by athlete Colleen Fotsch.

Looking at this example I started to think that it might be better to create a new property like "modeled by", "modeled after", "game model", "game character portrayed by", face or body model", "game character model" or something like that with a possible additional applies to part (P518) qualifier, but on the other hand I think we already have one that would be suitable for these cases. Looking at the enwiki article about this character, we see that the phrase based on is used everywhere, which makes sense and we can model it as follows:

Example 1: Abby (Q63190645)

voice actor
Normal rank Laura Bailey


add value

Example 2: Jill Valentine (Q840368)

voice actor
Normal rank Nicole Tompkins
applies to work Resident Evil 3
0 references
add reference


add value

Example 3: Max Payne (Q898615)

voice actor
Normal rank James McCaffrey
applies to work Max Payne
0 references
add reference


add value
based on
Normal rank Sam Lake
applies to part face
applies to work Max Payne
0 references
add reference


add value

Example 4: Vaas Montenegro (Q35525972)

voice actor
Normal rank Michael Mando
applies to work Far Cry 3
0 references
add reference


add value
based on
Normal rank Michael Mando
applies to part face
applies to work Far Cry 3
0 references
add reference


add value

Example 5: Alan Wake (Q21010056)

What do you think? Regards Kirilloparma (talk) 02:05, 19 October 2023 (UTC)[reply]

Instead of based on (P144) I'd prefer model (P2634) (this is also the property used for animated characters). - Valentina.Anitnelav (talk) 13:57, 20 October 2023 (UTC)[reply]
Though I haven't looked too thoroughly, I agree with Valentina. I might expect based on (P144) to be used for subjects whose character as a whole is the basis of the fictional character. If it's just likeness, model (P2634) seems most suitable. EdoAug (talk) 16:19, 20 October 2023 (UTC)[reply]
The examples above with model (P2634) as suggested make sense to me. Jean-Fred (talk) 20:19, 3 November 2023 (UTC)[reply]

Video game creation systems[edit]

How do we model that a video game was created on the Roblox platform? Is "distributed" correct?--Trade (talk) 15:45, 30 October 2023 (UTC)[reply]

I suppose distributed by (P750) along with software engine (P408). Similar to RPG Maker games. In particular, the Nintendo 3DS version of RPG Maker, RPG Maker Fes (Q60636482), only offered a way to download games made in the engine over the internet from its official servers (those are close to being lost media since the servers were shut down on 2023-03-31). --Sanqui (talk)
It looks like some of the existing items use distributed by (P750) -> Roblox (Q692989) and software engine (P408) -> Roblox Studio (Q97167512), which makes perfect sense to me. The usage of platform (P400) is a lot more inconsistent however - how should that be handled? It currently ranges from just duplicating every platform that Roblox is available on, to including just a select few, or even having Roblox (Q692989) itself (which is currently explicitly disallowed by a constraint on P400). -- Lewis Hulbert (talk) 13:09, 31 October 2023 (UTC)[reply]
Are these similar to video game mods, such as Defense of the Ancients (Q215936), Team Fortress (Q2164966), or Tekkit (Q19838382)? EdoAug (talk) 16:09, 31 October 2023 (UTC)[reply]
No, Roblox is explicitly a game creation system. Trade (talk) 04:06, 3 November 2023 (UTC)[reply]
Would instance of (P31)=custom game mode (Q86442388) be suitable for such? I'm assuming these are all played within the Roblox game client, though I might be wrong! depends on software (P1547)=Roblox (Q692989) as well, if that's the case? EdoAug (talk) 20:36, 3 November 2023 (UTC)[reply]

Video game achievements[edit]

Could someone give me feedback on the model item i cooked up? Key To The City (Q123348995)

If you have an opinion whether or not individual achievements should have items this will be the place to voice it

As far as i can tell there are no major video game databases that lists individual achievements across all three major platforms (PC, PS and Xbox)

Trade (talk) 13:17, 6 November 2023 (UTC)[reply]

Some feedback:
  • I think inception (P571) should have a reference, even if it's "inferred from date of publication" or something. Achievements can be tied to a publisher's platform so they can be added at a later date.
  • I don't think present in work (P1441) is appropriate for stating the game. The description claims "this (fictional or fictionalized) entity, place, or person appears in that work as part of the narration". Achievements are, at least typically and by definition, not a part of the narration, with the exception of a few breaking-the-fourth wall meta games (I'm thinking The Stanley Parable (Q7766240)). Some achievements are not a "part of" the game either, as games like Sonic the Hedgehog 2 (Q35627) did not have any achievements on release but have gained quite a few with later rereleases.
  • numeric value (P1181) appears to be used for constants, additionally, it has a stability of never changes (Q23611288). I don't think that's true for points awarded for achievements.
I am not fundamentally opposed to achievement items but I wish you good luck with modeling them correctly! --Sanqui (talk) Sanqui (talk) 16:36, 6 November 2023 (UTC)[reply]
We do need some way of linking achievement items to their respective works and to state the Gamerscore value of Xbox Achievements Trade (talk) 17:31, 6 November 2023 (UTC)[reply]
If we do want to track these, I think you could probably mass-import these from Steam relatively easily. Do you know if there's any way to permalink to a specific Steam achievement in Steam? Like is there a numeric ID for them in Steam?
I don't have a strong opinion on whether these are worth tracking, I'm pretty ambivalent tbh. The modeling here looks good to me, though. Nicereddy (talk) 16:02, 10 November 2023 (UTC)[reply]
No but third party websites are always an option Trade (talk) 23:30, 10 November 2023 (UTC)[reply]

@Jean-Frédéric, Kirilloparma, Nicereddy:--Trade (talk) 10:10, 8 November 2023 (UTC)[reply]

The modeling looks correct, but there is one additional bit of metadata for Steam achievements which is the API NAME of the actual achievement, such as that returned when using the Steamworks Web API and as shown on SteamDB website, such as "GAMECOMPLETION100". It's used as an identifier of the achievement and set by the publisher/partner and is not returned on HTML web views, but only with the Steamworks Web API. Achievements properties are detailed here: https://partner.steamgames.com/doc/features/achievements One other thing of note is that both Display Name, and Description are actually localized string values sometimes, where the default is English https://partner.steamgames.com/doc/api/ISteamUserStats#GetAchievementDisplayAttribute It might be a "good idea" to request to add a new Property for that API NAME identifier such as "Steam achievement API NAME". That would help further Linked Data sharing among websites and applications of Steam Achievements (I cannot comment on Xbox ecosystem however) --Thadguidry (talk) 04:05, 28 November 2023 (UTC)[reply]
Could you propose the property please? Trade (talk) 10:33, 14 December 2023 (UTC)[reply]

Place of publication and digital download[edit]

Should Steam games have their place of publication set to worldwide? Seems like that is the intended purpose of the worldwide item Trade (talk) 22:57, 16 November 2023 (UTC)[reply]

A brief research tells me it appears to be possible for publishers to set up region locking for Steam games, so I suspect it's not guaranteed that a game published on Steam is available worldwide. --Sanqui (talk) 08:58, 17 November 2023 (UTC)[reply]
How do we model which country a game is locked from? Trade (talk) 12:56, 17 November 2023 (UTC)[reply]
No idea on that modelling, but I think there is also US Sanctions on Iran and North Korea that would play a role, and it would be tedious to list that on every items, despites it being likely the correct way. Misc (talk) 11:57, 18 November 2023 (UTC)[reply]
Or we could just not be overly literal when applying the item Trade (talk) 22:13, 23 November 2023 (UTC)[reply]
No. Specifically, Steam's App Ownership Flags are actually internal usage only since an Owner can change Region, AND RegionRestricted and RegionLocked (both of which are NOT SETTINGS BUT ERROR CODES RETURNED https://partner.steamgames.com/doc/api/steam_api) can change over time based on a number of factors like laws, age-gates, and Steam approval processing, etc. Steam internally controls this and modifies the overall criteria. Content Restrictions are also a moving target and should not be modeled since it depends on the user and not necessarily the game or app. See BlsLowViolence on https://partner.steamgames.com/doc/api/ISteamApps Finally, sales or general "doing business with" for Steam is controlled by The Office of Foreign Assets Control of the U.S. Department of the Treasury and where Steam gives additional information on the Parter FAQ here https://partner.steamgames.com/doc/gettingstarted/faq --Thadguidry (talk) 05:10, 28 November 2023 (UTC)[reply]

Bad use of genre help[edit]

Hi! I'm new to wikidata. I was looking at the task "clean up bad use of genre." It seems like the Wikidata:WikiProject Video games/Statistics/Genre list is bot-created. If it is created from all genre tags that exist on video game concept schemes, is the reason that it is bad is that not all the genres that have been added to video games are defined as video-game specific? Please let me know how far off the mark I am! Dkreisst (talk) 20:51, 21 November 2023 (UTC)[reply]

The list is created from all items that are used on video games as genres regardless if the items themself are actually genres
"retrogaming", "ray tracing", "hypothetical Axis victory in World War II", etc Trade (talk) 22:20, 23 November 2023 (UTC)[reply]
I removed some of the obviously wrong ones up that were input devices or something else. Most of them were marked as imported from English Wikipedia. Whoever did that unsupervised and with no additional clean up step should stop that immediately. Matthias M. (talk) 18:20, 2 March 2024 (UTC)[reply]

PlayStation Store IDs[edit]

I've noticed a lot of products on the PlayStation Store now use this URL format: https://store.playstation.com/concept/10003735/ which isn't compatible with the existing PlayStation Store properties. Is there anybody more knowledgeable on this? Does the product also have an ID that works with the existing properties, or is a new property required for this "concept" URL? Lewis Hulbert (talk) 03:58, 4 December 2023 (UTC)[reply]

I have reached out to Trang Ho at SIE to ask them about this. My hunch is that this is part of their larger data modeling changes happening for AI and machine learning across their data strategy (where I've noticed many job postings that align with that need). Stay tuned, and let's wait till I hear back from Trang Ho. Thadguidry (talk) 03:22, 9 December 2023 (UTC)[reply]
@Thadguidry did you end up getting any response here? Nicereddy (talk) 17:04, 31 December 2023 (UTC)[reply]
I did not get a response, unfortunately. But my hunch is very likely correct, since most of the gaming industry is onboard with Linked Data efforts and pushing along with them, thankfully. In fact, many of the publishers have on staff semantic experts which a few I know and are friends with. One thing to keep an eye on is the usage of the concept URLs. To see how that might be used to link data across Sony Playstation sites and partners. I'll be keeping an eye on that myself. I'll keep trying other ways to get more information like from partners who might know more specifically.  Thadguidry (talk) 23:46, 31 December 2023 (UTC)[reply]
See also: Wikidata:Property proposal/Playstation Store Concept ID. Lewis Hulbert (talk) 11:58, 12 January 2024 (UTC)[reply]

Lux and Lux Deluxe[edit]

@Mess I think you made a mess here. Lux Delux (Q122416203) which was Lux Deluxe and Lux (Q3840938) should remain separated as they are different games released at different times with different IDs in external databases. Please undo your merge. Matthias M. (talk) 16:29, 9 December 2023 (UTC)[reply]

@Matthias M.: I apologize for the misunderstanding: I've reverted my edits as suggested by you. Anyway, both the English Wikipedia and the Italian Wikipedia articles talk about the 2002 videogame (so that's why I did the merge), even if the English article is more focused about the whole Lux videogame series. Probably the current associations among Wikidata items and Wikipedia articles should be rearranged in a better way. Best regards. --Mess (talk) 16:44, 9 December 2023 (UTC)[reply]
I see. That is a common problem on older and less maintained Wikipedia articles where initial game, successor and series are mixed. I created Lux (Q123735371) (the series) now as a common denominator, so you can hopefully connect the Wikipedia articles after changing them to match. Matthias M. (talk) 17:55, 9 December 2023 (UTC)[reply]

Numerical IDs have vanished[edit]

Hall of Light ID (P4671) changed the subdomain and redirects to a slug:

http://hol.abime.net/2632https://amiga.abime.net/games/view/indiana-jones-and-the-last-crusade-the-graphic-adventure

The numerical IDs cannot be extracted easily from the HTML source code. I only found it as part of the screenshot filenames. This does not impact Wikidata yet as old links still work, but how should new entries be added manually?

Same inconvenience for Lemon 64 ID (P4816) where we host the numerical ID but the website uses slugs

https://www.lemon64.com/?game_id=1513https://www.lemon64.com/game/lemmings

However, here you can find the game ID in the metadata, which is still less accessible than harvesting from URLs. Redirects work. Matthias M. (talk) 12:05, 28 December 2023 (UTC)[reply]

Ooh, this is unfortunate. @Jean-Frédéric any chance you have a contact at Hall of Light you could ask about this? Nicereddy (talk) 17:07, 31 December 2023 (UTC)[reply]
Ah, yeah, noticed that too, but thought there were in the HTML somehow... (I did find the HoL IDs as <div id="version_list_2632".) I think the easiest I found is to use the LemonAmiga entry to find the HoL ID… ^_^" Jean-Fred (talk) 18:30, 31 December 2023 (UTC)[reply]

WikiProject Video Games Wikidata Telegram group chat[edit]

Hello! For folks who are interested, we have a Telegram group chat for the WikiProject you can join if you're interested to discuss editing activities and such. We've had it for a few months, but I was a bit wary of sharing it more widely in the talk page for fear of spam. I've set the link to require approval by a group admin, so it may take a few hours to actually get into the channel.

applsdev Arlo Barnes BugWarp Coloradohusky CptViraj Cupkake4Yoshi Cwf97 Cynde Moya Danrok Datumizer Dexxor Diggr Dispenser Dollarsign8 DoublePendulumAttractor EdoAug Edolusill Eniehack Facenapalm Floyd-out FullyAwesome Harshrathod50 Jean-Frédéric Jotamide Keplersj Kirilloparma Lewis Hulbert LotsofTheories Macocobovi Macrike Master Of Ninja Matthias M. Metafire18 Nicereddy Nw520 Oduci Poslovitch Rampagingcarrot RampantSpirit Sanqui Santer Sight Contamination Sir Lothar thgiex Tomodachi94 VGPaleontologist Wd-Ryan WikiSyn YotaMoteuchi

Notified participants of WikiProject Video games

Best, Nicereddy (talk) 17:12, 31 December 2023 (UTC)[reply]

I requested to join, my display name is 香酥鴨 - I promise I'm not a spam bot! --Lewis Hulbert (talk) 19:13, 31 December 2023 (UTC)[reply]

Performed another mass-import, ~9600 new video game items[edit]

Using the same scripts/tools as the mass import we did in September, I did a small mass import on December 30th to add 590 video game items (thanks as always to Facenapalm for many of the scripts I used!). Then I noticed that my script for gathering the potential Steam IDs had a bug in it that was causing it to miss a ton of games that *could* be imported, so I fixed that and ran it again. Today I imported a further 9,080 new video game items from Steam/IGDB.

One notable enhancement versus the previous batch a few months ago is that this one checked for potential dupes and filtered them out (essentially, any game that happened to have the same name as an existing video game item on Wikidata got filtered out). So hopefully there won't be any problems with duplicates this time around.

Please tell me if anyone sees any problems with the newly-created items. The QIDs for the Dec 30 batch are here and the QIDs for the Jan 1 batch are here. I'll be enriching these over the next couple days with additional external identifiers and other automated data from my and Facenapalm's scripts.

Happy New Year :) Nicereddy (talk) 04:17, 2 January 2024 (UTC)[reply]

Nice job. If you keep pulling in the latest stuff from Steam, then if we concentrate on getting older games by other means, we might get somewhere in terms of completeness. Matthias M. (talk) 11:05, 3 January 2024 (UTC)[reply]
But "same name", should not be necessarily filtered out, right? A few games indeed have the same name. It should look at other properties/metadata to know if it's truly the "same game"? Instead, import dupes, then those additional checks could be done after an import using a ShEx entity checker and rules within it to remove dupes? Thadguidry (talk) 02:40, 12 January 2024 (UTC)[reply]
@Thadguidry technically yes, and that's how we did it for the first major batch in September, however that resulted in hundreds of potential dupes that I had to go through, so it wasn't really worth the trouble. Nicereddy (talk) 17:10, 15 January 2024 (UTC)[reply]

MC wiki dab[edit]

can someone help distinguish English Minecraft Wiki (Q123456803) and Minecraft Wiki (Q105533483)? Arlo Barnes (talk) 18:07, 15 January 2024 (UTC)[reply]

I'm not sure what is going on here, both items just seem like duplicates of the same thing as they currently are. I'm guessing that Q123456803 specifying "English" may be saying that only the English part of the Fandom wiki migrated away, while other languages (e.g. https://minecraft.fandom.com/pt/wiki/Minecraft_Wiki) remained using Fandom. The Fandom wiki does still exist in English though, so I think it would be better to have distinct items for the Fandom wiki and minecraft.wiki instead of trying to do... whatever this is.Lewis Hulbert (talk) 18:39, 15 January 2024 (UTC)[reply]

Hi folks,

Looking at Wikidata:WikiProject Video games/Statistics/instance of, we currently have 175 -instance of (P31)video game remaster (Q65963104), 157 of video game remake (Q4393107) and 78 of video game reboot (Q111223304) − sometimes with a of (P642) qualifier, resulting in for example The Last of Us Remastered (Q19057897)instance of (P31)video game remaster (Q65963104)of (P642)The Last of Us (Q1986744).

I would suggest deprecating this modelling in favour of -instance of (P31)video game (Q7889) + -based on (P144)The Last of Us (Q1986744)subject has role (P2868)video game remaster (Q65963104)

When it comes to the why, I would echo many of the points made by Spinster (about movies, but I think they apply here too) in Wikidata_talk:WikiProject_Movies/Properties#Precision_of_P31_and_subclasses_of_types_of_film_(again...).

In particular:

  • to me, P31 is about the very core nature of something. Being a remake or a remaster is not, to me, more defining of The Last of Us Remastered than being, say, a PS4 game, an action-adventure game, a 2014 game…
    • worse, whether something is a remake/remaster/port can be an endless debate with sources disagreeing − we can deal with that, but it feels wrong to have such shaky grounds for the P31.
  • information that can be stored in other properties (like indeed the platform, the genre, the date, etc.) should be in these statements, not stuffed into P31.
  • drilling-down the class tree (whether directly in WDQS or with eg my ExLudo user-script) is slow and occasionally returns more than what you want (for example, you would get compilations and expansion packs since these are currently subclasses of Q7889 [which might be debatable in itself, but I think it’s fine enough])
  • the data modelling relies on the use of of (P642), which is deprecated.
  • As far as I can tell, other databases do not consider remakes/reboots/remasters as a top-level entity either (the exception is IGDB, which I really find odd).

(There are more subclasses in Wikidata:WikiProject Video games/Statistics/instance of that I’m not too happy with, but let's keep this discussion about these 3 for now)

Looking forward to your thoughts!

Jean-Fred (talk) 14:41, 17 January 2024 (UTC)[reply]

 Support, I fully agree with the raised points. Sjoerd de Bruin (talk) 14:47, 17 January 2024 (UTC)[reply]
 Support, there may be some cases with remake and remaster where video game edition (Q60997816) may be better as the P31, but I can agree with all of these three being deprecated at minimum. --Lewis Hulbert (talk) 15:02, 17 January 2024 (UTC)[reply]
 Support. Solidest (talk) 15:32, 17 January 2024 (UTC)[reply]
 Comment, but I don't think it's quite right to use the pair based on (P144)original gamesubject has role (P2868)remaster – such a statement would imply that it's based on a remaster, isn't it? In this case, I would suggest adopting the music project practice and use remaster/reboot/remake and the other qualities that characterise the video game release in has characteristic (P1552), directly (not as qualifier). And then we can create a separate class, like music release attribute (Q108040195), to group all these things together and define a set for this property. So this could be a real replacement of "of"-like qualifiers with direct properties instead of just swapping labels. Solidest (talk) 15:32, 17 January 2024 (UTC)[reply]
My understanding is that what you describe would be the case with object has role (P3831), not with subject has role (P2868). I do agree with you that we should make more use of has characteristic (P1552) ; but in that case I think that remasters and remakes are fundamentally a relationship with another work, which is better expressed with based on (P144). Short of creating properties “remake of” and “reboot of” etc., which I did consider − it’s also what’s floated by the Relationships among video games: Existing standards and new definitions (Q50180192) paper − but decided (so far) that the frontiers between these concepts are too blurry to make dedicated properties workable, and that we are better served by based on + qualifier (as I’m writing this, though, I do accept that based on would not really work super well for reboots ; perhaps rather as a qualifier on part of the series (P179)?) Jean-Fred (talk) 20:28, 17 January 2024 (UTC)[reply]
Yeah, you are right about subject/object, my bad. I agree that "based on" with qualifier sounds smoother and more readable with "remaster" and "remake". But perhaps this could be used in parallel with has characteristic (P1552) = remaster + remake? Since "reboot" as a qualifier doesn't seem quite fitting in either "based on" or "part of series". I just see more opportunity here in uniform standardisation than relying somewhere on qualifiers, somewhere on a second property, and somewhere on a third property. All these things look to me like part of the same qualities set. Speaking of uniform standardisation, next we could also move video game port (Q64641660), sequel (Q261636) from p31 to p1552. And there's much more potential of clearing up and standartizing p31s, like use group of video games (Q116741534) in p31 while move video game compilation (Q16070115), video game bundle (Q96240590) to p1552 as well, and so on. In the end we will have a fixed set of 8-10 recommended P31s that we can lock in the schema, and the similar loosier set for p1552. And I'm not sure that it will be possible if some things remain in the scheme as qualifiers - it just gives extra unsystematic space, which eventually result again with into what we have in p31 especially since qualifiers are rarely monitored and corrected. Solidest (talk) 22:39, 17 January 2024 (UTC)[reply]
 Support Like Sjoerd, I fully agree with the raised points. — Envlh (talk) 17:18, 17 January 2024 (UTC)[reply]
 Support This makes sense to me. Nicereddy (talk) 18:02, 20 January 2024 (UTC)[reply]
 Support Subclassing is not wrong in a sense. It can be useful. "It's a video game. Oh, but what KIND OF video game? It's a remaster. Oh, a video game remaster? Yeah." But I think your first comment is the most compelling, about having a single way to show some parentage of a video game, is the right move forward. And based on (P144) is exactly the right property to use to show some kind of parentage, and then just qualify the children. -- Thadguidry (talk) 01:31, 23 January 2024 (UTC)[reply]
 Support as this will increase compatibility with Mix'n'match (Q28054658) which does not support subclasses of video games so it will ignore everything marked solely as video game reboot/remaster (and expansion) Matthias M. (talk) 10:14, 24 January 2024 (UTC)[reply]
 Support. Done for Constructor HD (Q123456086). YotaMoteuchi (talk) 23:57, 16 April 2024 (UTC)[reply]

Thanks all for chiming in. Given the support, let's do this. I’m not sure I can reliably automate this, so I’ll work my way trough this query:

SELECT DISTINCT ?item ?itemLabel ?type ?typeLabel ?of ?ofLabel ?based ?basedLabel ?role ?roleLabel WHERE {
  VALUES ?type { wd:Q4393107 wd:Q65963104}
  ?item wdt:P31 ?type ;
    p:P31 ?p31statement.
  OPTIONAL { ?p31statement pq:P642 ?of. }
  OPTIONAL {
    ?item p:P144 ?basedStatement.
    ?basedStatement ps:P144 ?based.
    OPTIONAL { ?basedStatement pq:P2868 ?role. }
  }
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
Try it!

Jean-Fred (talk) 19:36, 10 April 2024 (UTC)[reply]

(Solidest raised this on our Telegram channel, let’s have the discussion here too for transparency and traceability)

To model the person who wrote the script/scenario of a game, we currently recommend to use author (P50). Solidest argues (and I agree) that this property is more designed for literary works, and that we should rather be using its subproperty screenwriter (P58), designed for audiovisual media. P58 already has a require-constraint on video game writer (Q3476620) (since 2019)

Despite not being documented, this seems to have become the de-facto standard, with P58 used with 428 video games and P50 with 218 video games.

I think Solidest’s proposal makes absolute sense. The only hitch (as pointed out by folks in the English Wikipedia Video game Wikiproject) is that the English label screenwriter is not typically used for video games. That said, the French label scénariste (= scenario writer) and the English aliases (written by, writing credits, scenarist, story by) leave no doubt to me about the intent of the property. P50 feels also in imprecise: do we mean "author" as in the scenario writer, "the sole creator" (like say, Chahi for Another World, where we would rather use P178) or as in Auteur (in the artistic sense, like has been said of Fumito Ueda, Hideo Kojima or Goichi Suda − something somewhat fuzzy that we don’t want to model I think)

So unless there are any objections, I suggest we do the following:

  • [TODO] Update Wikidata:WikiProject_Video_games/Properties#Staff to use P58 instead of P50
  • [TODO] Check with any Wikipedia uses of the property (like infobox field) and help them through the changes.
  • [TODO] Go through the 218 uses of P50 and migrate them, where relevant, to P58 (I don’t think we can automate that, per the unclarity I pointed to above)
  • not much to do on P58, as it already has the relevant constraint
  • [MAYBE?] Add a constraint to P50 to prevent its use on video game (Q7889) items.
  • [OPTIONAL] if we are really bothered by the English screenwriter label, start on the property talk page a discussion to change it to "scenarist" or "scenario writer" or whatever

Looking forward to your thoughts!

Jean-Fred (talk) 10:29, 18 January 2024 (UTC)[reply]

Thanks for starting this discussion. In the enwiki discussion yesterday I gave some additional arguments in favour of P58 over P50:
– a writer in video games is closer to a film screenwriter than to writing literary prose or poems - there is a narrative that is played intended to be played out or performed visually (essentially is written in dramatic mode)
– on wikidata author (Q482980) is described in the context of literature and books, screenwriter (Q28389) is described in the context of cinema, and for video games there is a separate video game writer (Q3476620), the latter is not connected to any of the properties, which should be corrected.
– articles on enwiki actually describe video game writers in the context of en:screenwriter and the corresponding subsection in en:writer#screenwriter, not in en:author.
– additionally, as already said, "author" can be understood as the sole creator of a game, and I was looking through the use of author properties yesterday and found some of these cases, e.g. Portal (Q274897) now lists both game director and writers in author, or Seastalker (Q7441958) has game designers lists as authors. But these are rather rare cases, as this is mostly being exported fromthe writer parameter from in enwiki.
Other suggestions that have come up on enwiki is to create a separate property for a video game writer. This doesn't seem to me to be the most optimal, as it should actually be not much different from the position described by screenwriter (P58) - it's all about writing scripts for further performance in different environments and mediums, such written works most often conform to the rules of dramaturgy (dialogues - acts - narrator/author's comments). Video games aren't much different from cinema in this regard.
So I  Support all of the above points, including the title change for P58. Also, this should probably go to the P58 discussion, but en:Template:Infobox television uses 4 different parameters for the following: (1) "writer" - a generic post combing screenwriting + story by position. (2+3) "screenplay" and "teleplay" for scripts specifically, with different titles displaying when using different parameters. (4) "story" - a separate property for an idea or original story. I'm not sure how such a variety fits into 1 property and perhaps it is to be determined, with the potential of creating a new property. Since some other infoboxes also have the practice of separating "screenplay" and "story by".
And the last but not least thing to do is to add the alias "video game writer" to screenwriter (P58) with already mentioned listing video game writer (Q3476620) in Wikidata item of this property (P1629). Solidest (talk) 17:03, 18 January 2024 (UTC)[reply]
There has not been much input here ; but there has not been any opposition either :) Can I assume that no-one has any concerns about this? Jean-Fred (talk) 17:32, 4 March 2024 (UTC)[reply]
There is also a discussion on the screenwriter (P58) talk page, started a month ago, about changing the default English label regarding comics, which is relevant to what we are discussing here: see Property talk:P58#Default English label. Solidest (talk) 14:20, 29 March 2024 (UTC)[reply]

Which of the three titles should have preferred status? --Trade (talk) 16:48, 28 January 2024 (UTC)[reply]

Current title is the right one. [139] — official site has "3" title instead of "three". And subheadings are normally marked first with a colon and only secondly with a hyphen. Solidest (talk) 08:31, 1 February 2024 (UTC)[reply]

Evolve (Q16267814) & Evolve Stage 2[edit]

Any complaints if i split these two into seperate items? It's hard to argue they are the same game given the massive differences Trade (talk) 23:35, 22 February 2024 (UTC)[reply]

I think that's probably fine? It feels a bit weird since the game is still marketed under the same exact name and uses the same identifiers on storefronts, though. Nicereddy (talk) 16:21, 25 February 2024 (UTC)[reply]
Most databases (MobyGames, GameFAQs, GiantBomb, PCGamingWiki, etc.) cover both as one though. Jean-Fred (talk) 18:54, 25 February 2024 (UTC)[reply]
That's also a good point, I guess it should be kept as one. Nicereddy (talk) 20:45, 25 February 2024 (UTC)[reply]
Both CSGO and CS2 uses the same Steam ID, that doesn't change the fact that they are fundamentally two different games Trade (talk) 22:14, 25 February 2024 (UTC)[reply]

Business models[edit]

  • free-to-play
  • microtransaction
  • battle pass
  • subscription

Is there any other business models for video games that we should accept?--Trade (talk) 22:15, 25 February 2024 (UTC)[reply]

Perhaps buy-to-play (Q28456863) or pay to play (Q594235) for games that require purchasing the game, but doesn't have any additional payments once the game is bought? DoublePendulumAttractor (talk) 05:57, 26 February 2024 (UTC)[reply]
shareware (Q185534) for the era before free-to-play and possibly also adware (Q193345) for mobile games. Matthias M. (talk) 09:59, 26 February 2024 (UTC)[reply]
Relevant: the currently used values at Wikidata:WikiProject Video games/Statistics/Business model. (Of these, I’ve always found crowd funding (Q348303) a bad fit for business model (P7936) (as you can crowdfund just as well free-to-play or one-time-transaction games) and should rather be moved to funding scheme (P6195)?)
Also relevant: PCGamingWiki’s Monetization ontology Jean-Fred (talk) 10:30, 26 February 2024 (UTC)[reply]
Yes, crowd founding is not a business model. It always struck me as odd. Bots tend to tag that and need a change. Matthias M. (talk) 14:40, 26 February 2024 (UTC)[reply]
PS: Steam Greenlight (Q61905393) and pay to win (Q108523183) are not business models to me. name your own price (Q17144905) and pay what you want (Q846616) seem redundant. Matthias M. (talk) 14:46, 26 February 2024 (UTC)[reply]
What do you suggest we do with Steam Greenlight (Q61905393) instead?--Trade (talk) 15:31, 26 February 2024 (UTC)[reply]
Steam Greenlight (Q61905393) might be better as a qualifier to distributed by (P750)Steam (Q337535), maybe using has characteristic (P1552). DoublePendulumAttractor (talk) 16:27, 26 February 2024 (UTC)[reply]
Could you make a batch to move the item? Trade (talk) 21:56, 26 February 2024 (UTC)[reply]
I moved them to approved by (P790) now. Matthias M. (talk) 13:17, 3 March 2024 (UTC)[reply]
There must be a reason why name your own price (Q17144905) and pay what you want (Q846616) have two seperate Wiki articles--Trade (talk) 15:26, 26 February 2024 (UTC)[reply]
From the wiki pages, name your own price (Q17144905) looks like it means making a price offer and having a sale occur only if the seller agrees. It looks like a type of auction, where multiple buyers can make offers for the same item, and not something that can be used for business models in the context of video games. pay what you want (Q846616) (and simillarly pay what you can (Q17121671)) gives buyers a choice of what to pay, as long as it's over a floor amount. This looks like it'd be more relevant for video games. DoublePendulumAttractor (talk) 16:23, 26 February 2024 (UTC)[reply]
"pay what you want (Q846616) (and simillarly pay what you can (Q17121671)) gives buyers a choice of what to pay, as long as it's over a floor amount" What if there is no floor amount tho?--Trade (talk) 21:55, 26 February 2024 (UTC)[reply]
Having no floor would be the same as it being free, but with the option of including a donation. I guess in that case, it'd also qualify as donationware (Q10267) but that as a business model. Examples would be games like SimSig (Q1288814). DoublePendulumAttractor (talk) 23:22, 26 February 2024 (UTC)[reply]
careware (Q1035935) also comes to mind (which I used for Caper in the Castro (Q86753491)) Jean-Fred (talk) 17:22, 4 March 2024 (UTC)[reply]

How would you tag a game that is fully or partially hidden behind Patreon subscribtion or other similar platforms?--Trade (talk) 15:31, 26 February 2024 (UTC)[reply]

You already mentioned subscription business model (Q322948). Matthias M. (talk) 09:37, 28 February 2024 (UTC)[reply]

The Finals (Q113586393)[edit]

  • The Tournament consists of 16 teams (48 players) in a bracket set up.
  • The first bracket features four matches that are played simultaneously.
  • Each match consists of four teams with three players each (4x3=12)

Should i set the maximum number of players (P1873) as being 48 or 12? Trade (talk) 21:31, 15 March 2024 (UTC)[reply]

Ooh, that's... complicated. I would probably say 12? Nicereddy (talk) 00:11, 16 March 2024 (UTC)[reply]

Moving identifiers[edit]

Any complaints if i move the identifiers and interwiki links (excluding ENWP) from Final Fantasy (Q12391356) to Final Fantasy (Q99416119)? Trade (talk) 22:37, 22 March 2024 (UTC)[reply]

Same applies from Angry Birds (Q165094) to Angry Birds (Q111372948)--Trade (talk) 22:39, 22 March 2024 (UTC)[reply]

I'd like to move this from instance of (P31) to has characteristic (P1552) as well and replace it with just video game (Q7889) to comply with our model and to make it visible to Mix'n'match (Q28054658). Same for the subclasses open-source video game (Q1015325) and freeware video game (Q116278114). What do you think? Matthias M. (talk) 08:11, 29 March 2024 (UTC)[reply]

Sounds good to me, this information isn't really top-level instance of (P31) importance, better suited for copyright license (P275) and has characteristic (P1552). Lewis Hulbert (talk) 08:34, 29 March 2024 (UTC)[reply]
 Support Yes, absolutely agree. Jean-Fred (talk) 13:58, 29 March 2024 (UTC)[reply]
Okay done. Matthias M. (talk) 14:09, 29 March 2024 (UTC)[reply]

Which values should be listed as input devices?[edit]

I figure we might as well discuss the issue so we can make it somewhat consistent. I noticed there is a huge variety in how different items utilize input device (P479) for specific platforms Trade (talk) 02:08, 8 April 2024 (UTC)[reply]