WMF Quizzes/2025/Solutions

Overall notes

This quiz had some minor technical issues (eg the music didn't work for some reason, the leaderboard was hard to see), but overall the execution of the quiz was significantly better than 2024. It was also tightly contested, with one point being the difference between the first and second-placed contestant. The quiz was pretty challenging at the high end, but there were also many opportunities for people to pick up points.

The answers

Option D (Q7838)
See below
Option A (10² < N < 10³)
Option B (Wikimedia Commons)
Option B only (Edit all user rights (userrights))
Option B (A person or group trying to make themselves look better via their article on Wikipedia)
Could be disclosed: option B (A conlang (constructed or artificially created language)), option D (An extinct language) and option E (A language with a standard Unicode ending). Cannot be disclosed: option A (A simplified variant of another language) and option C (A dialect of French).
Option E (Sockpuppetry)
Required to disclose: option B (Board of Trustees) and option C (AffCom (Affliate Committee)). Not required to disclose: option A (Steward), option D (U4C (Universal Code of Conduct Coordinating Committee)) and option E (CRC (Case Review Committee)).
Option A (Increase the protection to full-protect)
Option C (A former system used to track and report bugs and requests in MediaWiki and other WMF projects)
Can be uploaded: option A (Japan Railways), option B (Nederlandse Spoorwegen (Dutch Railways)) and option C (Amtrak). Cannot be uploaded: option D (Indian Railways) and option E (SNCF)
Option C (Hilfe:Spielwiese (= Help:Playground))
Egypt = 2008, South Africa = 2018 and Kenya = 2025
a, e, f, английский and baba respectively
Option A (Supreme Court of India)

Answer for Q2:
Column A	Column B
Wikipedia
Wikisource
Wikinews
Wikibooks
Wikivoyage

Explanation by question

Question 1

This question was done reasonably well with over half the answers correct, given that not everyone actually "uses" Wikidata.

This question is just bookwork. Options B and C can be discarded since it would not make sense for Wikidata to not use a unique identifier (which Swahili isn’t, and even then, shouldn’t it be Kiswahili as that’s how it’s referred in that language?). Option D is invalid, and option E redirects to the Wikidata namespace, which is used for pages related to Wikidata itself.

Note that the options were written in such a way that knowing that Q7838 referred to Swahilli wasn’t needed.

Question 2

This question was done well - the vast majority got this one. A few ran out of time, however.

Nothing much to do here – basic identification. Two of the answers are very obvious since the answer is directly in the image; this is by design.

Question 3

This one was somewhat harder than expected, with option B being a popular distractor and a significant minority picking the other three numbers.

If we look at the number of wikis that can be attached by SUL (Single User Login), which would all be publicly-editable wikis, we get N ~ 851. That is between 10² and 10³, which is option A.

Note that this excludes fishbowl (publicly readable but not editable) and private wikis (which require a login to even read, beyond the main page or so).

Question 4

I expected this one to be pretty easy, and while most did get this right, a surprising minority chose options A and D for some reason.

A pretty straightforward problem. The idea is that we have a video under a free license (Creative Commons) which we want to upload to a WMF project, and the natural answer would be Commons since all the projects can make use of the video.

Options A and D are technically possible, but not a good choice since other projects cannot make use of the video without additional work. The other two options are not projects that are designed to store media.

Question 5

As expected, most had no idea about what the question was talking about. Over 12% of people did get it right, which is reasonable for what I expected as the hardest question in this quiz, even if it appeared to be lucky guesses or using faulty reasoning. All of the options (except E perhaps) were popular.

A very challenging problem targeted at the top end of participants, as this one requires significant depth in an area most people are likely to have little to no idea about.

First break down the problem. The important parts are that user B has no permissions at all beyond a new user, and that user B has one global role (with our task to add rights to that global role). The next deduction is that CheckUser Wiki, while a WMF wiki, is a private wiki and hence not connected through SUL, so you’ll need to make that change without logging in to that wiki.

This itself is very non-obvious and I do not expect more than a handful to get this far, as it’s plausible to believe that private wikis are “disconnected” from the rest of WMF for security reasons – but the userrights-interwiki technology is pretty dated (it was used right from when stewards were created around the end of 2004 or so).

Then, look at each option on its own:

Option A is tempting, as can be seen from the name, and is something that is actually required. The problem is that this role is useless if you cannot do anything else.
Option B is not enough on its own, but is actually the correct answer. We’ll explain further below.
Option C has two problems:
- The account has already been created on the target wiki
- The permission is linked to CentralAuth, and hence will only work on SUL wikis (which this is wrong).
The permission in option D does not exist.
Option E is irrelevant to the question.

Now, why does option B work (and not option A)? The answer is that with option B, the following sequence is possible:

Use Special:UserRights on Meta-Wiki to give themselves steward. This gives access to userrights-interwiki (option A)
With this, go again to Special:UserRights and enter B@checkuserwiki and give themselves bureaucrat on that wiki that way.

Keep in mind that any combination of options involving B could work, but the question wants us to pick the fewest number of valid options.

P.S: I suspect a few people may not realise that private wikis cannot be accessed by having global permissions, and come up with option B because userrights allows to change any right on any global wiki connected via SUL. Such people are just lucky here.

Question 6

This was a bit harder than expected, with just over half the attempts being correct. Options D and E were chosen by a significant minority of candidates.

Consider each option individually:

Something adjusted from the computer of a royal palace is not likely a “fanboy” (who would be independent users protective and ardent users/lovers of something).
This is the correct option. In this case, it is clear that because the modification was done in a royal palace, that a group (or Mabel/Friso itself, as was actually the case) wanted to make themselves look better by removing information that Mabel lied.
Nothing was blocked, so option C is illogical.
The police had nothing to do with this, and no one in the palace made a complaint, so option D is also illogical.
Mabel/Friso (or someone in the royal palace) is not a “random user”, not to mention that nothing in the fragment suggested that they were correcting an “error”.

Question 7

As expected, very few got all the way to the end, though almost everyone who attempted this question got some of the points so partial marking helped.

These kinds of questions tend to be challenging to get full points on, despite there being little in the way of advanced reasoning required. This is because the topic is not in an area I expect most to be familiar.

“Simple” versions of a language are not allowed, and more so for Wikisource where the goal is to archive historical documents. -> not allowed.
Conlangs, such as Esperanto, have always been permitted as long as they meet specific criteria (such as having an ISO language code). -> allowed.
Wikisource is the one project where extinct languages are allowed, because there’s a use-case for archiving extinct languages. Now, they are not recommended (the preference is to use a “major” language or Old Wikisource if possible), but they are within the rules. -> allowed
A dialect of French is too similar to an existing language to be allowed. This can be tricky to reason through due to the presence of languages such as Egyptian Masry Arabic, which do have their own languages. -> not allowed
That’s a positive and something virtually all languages have, so yes that’s fine. -> allowed.

Question 8

Done reasonably well; most of the participants got this one. Those who got this wrong picked options A, B and C with roughly equal frequency.

Despite the unfamiliarity of many of the terms in this question, the question was carefully written with multiple “escape paths” which should mean that most participants can get this right. Now look at the terms presented:

"Mayavi-Lutapi" memberships – I expect very few to figure this one, even though for those who do, it is a huge clue. Mayavi (hero) and Lutapi (villain) are Malayalam-language cartoon characters and the presence of this phrase means that it is not allowed for users to have a “good” account and a “bad” account at the same time (i.e, the bad account vandalises, while the “good” account reverts them).
“Pseudo-idols” – this requires deducing that “idols” is a copy (of usually a religious deity). Again, for those who get it, big clue, but I don’t expect many to.
“Wikipedia tigers” – the phrase is too vague to be useful.
“using multiple memberships” – I expect this to be the key for many participants, as that is a clear clue to sockpuppetry.
“official policy” – that’s going to be true for most of the options, and hence isn’t going to be terribly useful.
“escape from other editors” – Probably a clue, but it could also apply to some of the other options.
“falsely accused of being someone else's idol” – I expect this to be a major clue for many participants. While many may not recognise the “idol” term (like the second phrase), the rest of the phrase is a clear hint of having multiple accounts.
“Aparamoorthy Administrators” – probably a religious reference, and a pretty useless hint.

The page is wikipedia:ml:വിക്കിപീഡിയ:അപരമൂർത്തിത്വം.

Question 9

As expected, not many got all the points, but everyone who attempted this on time got at least some of the points.

Like Q7, I expect few to get all the way to the end, but partial marking should get most people quite far as some of the options are not difficult for average participants.

Affiliation Committee requires public disclosure. This is somewhat tricky to deduce from the name, but the application process for this role is public and requires disclosing the real name (this differs from Ombuds and the CRC whose applications follow a similar cycle – those applications are private).
CRC does require private disclosure, but not public. In fact, CRC members are not allowed to disclose that they are on the committee at all as long as they are part of the committee.
Steward and U4C applications are managed by the volunteers and do not require disclosure.
As a “public-facing” role, Board of Trustees applicants must disclose their real name publicly, and are usually referred to as such throughout the application process (and as a member).

Question 10

This question was considerably harder than expected, and somewhat controversial as a few complained that the question was ambiguous. All of the options (except C perhaps) were chosen by a significant number of people, probably missing the "simplest and most effective solution" part of the question.

This reasoning question primarily targets participants working at the lower-to-mid end of this quiz, though I suspect that the options may cause some difficulties.

This is the expected answer. All the LTA does is vandalise a single page, so fully protecting the page will solve the problem. Also, the main page is the one page that should be fully protected anyway – or the system administrators may do it for you anyway if there’s a potential threat (as they did to a few major wikis ~1.5 years back).
This is not only unnecessarily heavy-handed, but “detecting” vandalism is never going to be fully accurate and one can expect both false positives and false negatives to occur from this process.
This won’t solve the problem, may end up inconveniencing new contributors, and the LTA can simply wait longer.
This is a nonsensical option, especially since a lot of users visit the wiki but do not exist – such an approach will block all of them.
This could work in theory, but
- New users can’t edit the main page anyway as the main page is already semi-protected
- Wikispecies doesn’t even have FlaggedRevs

Question 11

Given that Bugzilla was retired in 2014, this was done pretty well with over three-quarters getting this right.

This recall question is likely to be a case of “you either get it or you don’t”. The only thing to do is to notice that this is an image of the Bugzilla mascot, and then it becomes clear what the image is referring to – Bugzilla.wikimedia.org. Note that Bugzilla is not a VCS as it isn’t where code was stored – there were other places for that at that time (and Phabricator isn’t the place either today).

Question 12

Like with the other questions of this nature, everyone who attempted got some of the points but only a minority managed to get all the way to the end. Surprisingly, a lot of people thought that Amtrack's logo cannot be uploaded into Commons, while most could answer the SNCF part correctly.

I'm aware of commons:File:Indian_Railways_Logo_Red_Variant.svg and am trying to figure out how that was uploaded. I believe that is an image that does not belong at Commons.

This is a “fat-curve” question, i.e, one that spans a wide level of participants. There are options I expect most to be able to figure out, with a couple being more discriminating. Hence, I expect a wide range of scores for this one. The theme of this question is determining whether the images meet the PD (public-threshold) definition of originality for Wikimedia Commons.

The logo for Indian Railways is definitively not “simple” and would not qualify to be uploadable to Commons.
Japan Railways and NS (Dutch Railways)’ logo is pretty geometric and just a couple of shapes -> qualifies. Same for Amtrak.
SNCF is a tricker case, since it isn’t obvious whether it qualifies for PD or not, and the logo is kind of textual. The answer to this is no, and it’s actually because SNCF is French, and the threshold for originality is lower in France compared to the US. Since Commons requires the image to qualify in both the US and the country of origin, this won’t work.

Question 13

This was surprisingly poorly done, with less than half the answers correct. Option A turned out to be a very big distractor as over one-third of all attempts chose that option. The fact that the answer was in the "Help" namespace also appeared to cause some confusion.

This reasoning question was not meant to be difficult, as the correct option is fairly obvious.

Option A has two problems:
- IP addresses don’t have user pages
- Even if we’re dealing with a regular user, random gibberish is usually deleted with the user sent to the sandbox.
Options B, D and E are mainstream pages.
Option C is the correct one, as the name suggests. It’s just what most other wikis would call the sandbox.

Question 14

This was very well done amongst those who attempted with nearly everyone getting all the points. However, a significant number of people struggled with the 30-second time limit and ran out of time.

This is meant to be an easy gimme, though some of the weaker candidates may get confused between Egypt and South Africa. Not much to say here otherwise.

Question 15

This was very difficult, as expected, and the only question where none of the participants got all the way (though a few got 4). However, given that even the "easier" options were targeted at the more able candidates, it was good to see nearly 77% of all attempts getting some of the points.

This is a very challenging recall question, and only slightly easier than Q5 (with more opportunities for partial marking). In particular, it requires recalling how four magic words work, and while some of them are common and accessible to the more able participants, some are obscure.

The first two blanks test knowledge of the GENDER magic word, with the first deduction being that no user is given at all, so it will return a “default” value. Recall that the ordering is male|female|neutral.
- When all three parameters are provided, it defaults to what’s provided in the wiki, which is gender-neutral. Hence the second blank is e.
- When only two parameters are provided, it defaults to male. Hence the first blank is a.
The third blank tests knowledge of the PLURAL magic word. If the absolute value of the parameter is 1, it returns the first value, otherwise the second. Since |-1| = 1, it returns f as the third blank.
The fourth blank tests knowledge of the LANGUAGE magic word, and in this case, it is asking for the word “English” in Russian. The lack of distractors make this blank relatively doable.
The last blank is perhaps the most difficult option in this question. Notice that “baba” is clearly not a valid date. In this case, it will result in simply repeating the invalid text, and not (blank) or an error message.

Question 16

This was unexpectedly very difficult, with only a small minority getting this right. Option C in particular was very common, which makes sense given the sentence quoted, but a significant number of participants chose options B and D, which were clearly wrong.

Not much to say for this one.