Talk:Bayes' theorem/Archive 8

Archive 5

Archive 6

Archive 7

Archive 8

Are there copyright violations in the image being discussed now?

If there are copyright violations, please let me know which images and I will pursue that. Please keep the discussion in this section to only about copyvio. Constant314 (talk) 19:45, 28 September 2022 (UTC)

There was talk of this image, the first in the "Interpretations" section, being copyright infringement. However, the developers of the video games in question have IP guidelines which delineate what they do and do not allow in terms of the usage of their IP. They seem to allow usage of their IP for non-commercial and/or educational purposes. Furthermore, because of the nature of the image (Original work imitating a small amount of the IP primarily for nonprofit educational purposes), fair use likely applies to the image. Baldemoto (talk) 20:37, 28 September 2022 (UTC)

I've addressed this in Talk:Bayes'_theorem#Response_from_the_illustrator. Cheers, cmɢʟee⎆τaʟκ 00:39, 29 September 2022 (UTC)

Yes, you have, but I think it would be useful to add the rational to the image page on Wikimedia. Constant314 (talk) 16:36, 4 October 2022 (UTC)

Protected edit request on 4 October 2022

I request restoral of the previous version of this page for the following reasons:

1. The image in the previous version has an involved editor who has been making modifications in response to suggestions and requests for improvements made on this talk page. He is addressing the complaints.
2. The consensus appears to be going that way.
3. The last change which is now protected was made by editor that is now blocked.
4. The previous version was the stable version before the present edit squabble (I don't think it reached the war level).

By no means do I wish to imply that the issue is resolved, but rather that the previous version is a better version to be presented to the public while the issue is being resolved. Constant314 (talk) 16:29, 4 October 2022 (UTC)

Consensus is not yet reached, so we should wait for that instead of hastily going that way. 0xDeadbeef 00:02, 5 October 2022 (UTC)

 Not done for now: We should wait for the closer of the RfC to make this edit. 0xDeadbeef→∞ 11:35, 9 October 2022 (UTC)

Please add the 1966 Palomares B-52 crash as an example

The article doesn’t mention the 1966 bomb search, which is described by the BBC: “The lost nuclear bombs that no one can find https://www.bbc.com/future/article/20220804-the-lost-nuclear-bombs-that-no-one-can-find“. See also https://medium.com/war-is-boring/how-to-find-a-missing-h-bomb-aef5660ed51a.

Having looked over the angst created by an illustration, I have no desire to get involved in calculating the number of angels who might dance on the head of a pin. (I don’t know if this meme will be understood by all Wikipedia editors.)

For the record, I have no idea of the significance of “among us” in any form. If, in a hundred years, Wikipedia is viewed as we now look upon the oft-quoted 1911 Encyclopedia Britannica, will it be intelligible? Humphrey Tribble (talk) 02:10, 2 January 2023 (UTC)

I don't think that there are enough details there to make a useful example. Constant314 (talk) 03:10, 2 January 2023 (UTC)

@Humphrey Tribble: I don't think there is much sense of mentioning the use of Bayesian concept in that specific case, even if we have enough technical and mathematical data. If we do, may be we should also mention in Archimedes' principle it's been used in construction of USS Constitution, or describe in Airbag the role an airbag played in Stefan Eriksson's crash investigation. --CiaPan (talk) 13:09, 2 January 2023 (UTC)

Regarding your latter point: as was pointed out in the RfC above, if Among Us loses all relevance in the future, to the point that the image is unintelligible, then we can replace the image in the future. (Also, is the 1911 Encyclopedia Britannica really "oft-quoted"? It's clearly an outdated source; I'd imagine that most people would quote the newer editions.) Edderiofer (talk) 22:05, 17 February 2023 (UTC)

Removing a meme

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

---

Location of the meme: Interpretations fist image. Someone please replace the "diagram" 2405:201:23:9AD4:24D5:467:EBFB:15DE (talk) 05:00, 28 September 2022 (UTC)

I have to agree that the diagram isn't a very good visualization. But I prefer the Probability of assassin given suspicious behavior as those ideas are easy to grasp. Constant314 (talk) 05:08, 28 September 2022 (UTC)

I mean the meme is already 2 years old... I hate memes (talk) 05:35, 28 September 2022 (UTC)

Oppose. It is a completely valid diagram and I do not see how it would warrant the grounds of removal. There is no policy that says we cannot use a meme in an image. Captain Chicky (talk) 00:29, 23 October 2022 (UTC)

Oppose. IMO "it's a meme" should not on its own be grounds for removal, unless there's some Wikipedia policy or guideline concerning memes that I'm unaware of. See Wikipedia:Humor#Humor_in_articles; the image, though a meme, satisfies all the usual article requirements. Edderiofer (talk) 05:54, 28 September 2022 (UTC)

Pretty sure the assassin version is an edit of the original visualisation. I see no reason to inject humour into an existing informative visualisation in a non-productive way. Could be considered vandalism Fuhsini (talk) 17:25, 2 October 2022 (UTC)

See my rebuttal to you below, but once again, there is no Wikipedia policy or guideline I'm aware of that implies that "it's a meme" alone is grounds for removal. If you would like to argue this point, I'd welcome you to link such a policy or guideline. Edderiofer (talk) 11:17, 3 October 2022 (UTC)

It is vandalism in my opinion, yes, "it is a meme" is not a reason, but also, there is no reason to keep this image NightJasian (talk) 00:40, 28 May 2023 (UTC)

Info The original contribution of this image appears to be in good faith.

Support I'd favor the original over the Assassins version, as the former seems more generally familiar. However, I think a new image altogether that uses a real-world example would be ideal. Determining true positive rate when testing a population for a disease is a classic example. Perhaps User:Cmglee would be interested? Cheers all and thank you for flagging the potential vandalism! Scientific29 (talk) 06:05, 28 September 2022 (UTC)

Strong Support Isn't this a copyrighted character design? Pretty sure that's grounds for removal even without a replacement. --JiFish(^Talk/_Contrib) 08:16, 28 September 2022 (UTC)

Info Just for reference, the developers of the game have a page dedicated to their IP guidelines, and it appears that, generally speaking, non-commercial and/or educational use of the IP is allowed. Baldemoto (talk) 15:45, 28 September 2022 (UTC)

So the image has an invalid license. I would still be infringing the copyright if I was to use this for monetary purposes given that I saw the CC-BY-SA license. 0xDeadbeef $\to \infty$ 15:47, 3 October 2022 (UTC)

• Neutral: anything with a consensus. My main complaint is that there is no explanation within the text. You would have to already understand Bayes to get anything out of the illustration. Constant314 (talk) 11:04, 28 September 2022 (UTC)

Weak Support. I'm fine with the original over the current version, but I think the best solution is an illustration of the classic, low base-rate, disease testing example.  cjquines  (talk) 21:16, 28 September 2022 (UTC)

Strong Support Seriously, this is not how Wikipedia should work, this is like seeking attention. It is a valid diagram, sure, but uneccessary NightJasian (talk) 00:28, 28 May 2023 (UTC)

I request to replace the meme with the following updated Image: https://commons.wikimedia.org/wiki/File:A_geometric_visualisation_of_Bayes%27_theorem.png I hate memes (talk) 05:48, 28 September 2022 (UTC)

Oppose. That image looks like a badly-edited version of this one. It doesn't obey the licensing terms of the original, and is hence a copyright violation. And what's this about astigmatic eggs? Edderiofer (talk) 06:03, 28 September 2022 (UTC)

 Not done for now: Given the discussion above, there does not appear to be a consensus for this at this time. That being said, the suggested replacement is presumably inferior, as it is in PNG format rather than SVG. The inconsistency in the typefaces (the table column headings versus the rest of the text) and variations in what should be identical images (coupled with what appear to be random whiteouts, which are especially noticeable in the white eggs) are also somewhat problematic. -- Kinu ^t/_c 06:07, 28 September 2022 (UTC)

Both images are by the same Wikipedia contributor. If there are any concerns about the new version's quality perhaps they should be approached? --Escape Orbit ^(Talk) 12:18, 28 September 2022 (UTC)

The suggested replacement isn't by the same contributor as the original image (astigmatic/bearded) and the current image (assassin/suspicious); it's by @I hate memes. Edderiofer (talk) 13:26, 28 September 2022 (UTC)

Ah, sorry. Didn't realise there was a third image in consideration. --Escape Orbit ^(Talk) 17:10, 28 September 2022 (UTC)

Oppose. Thank you for putting this together. However, I agree with Edderiofer, the quality on this edit is too low and the example doesn't make sense.Scientific29 (talk)

Neither. Both images are bad, because the presented set of data maintains constant ratios across columns (consequently, across rows, too). As a result, it explains nothing about counter-intuitive results for highly dependent events, like those described in the Examples section. --CiaPan (talk) 07:21, 28 September 2022 (UTC)

By "both images" I assume you mean "the image proposed by I hate memes" and "the image that that image was derived from", but not "the image currently in the article", which doesn't have this problem? Edderiofer (talk) 12:42, 29 September 2022 (UTC)

Good catch on the constant ratios across the rows and columns in the suggested images. This would make (File:Bayes_theorem_assassin.svg) more useful, in my opinion. --Kinu ^t/_c 00:10, 2 October 2022 (UTC)

Strong Oppose. Not only the image uses examples that can't be used across times (no one will guess why it was an "assassin" or even that figure too as well in the future), this is just silly. The art was exceptionally good, but more real life examples are needed for this. Replace the Among Us figures with objects such as stickmen or bandits. And yes, let's not beat around the bush: this is just a meme. Silliness for a non-silly subject matter feels out of place (this article was written in a frozen register after all) 112.202.252.91 (talk) 16:06, 2 October 2022 (UTC)

I think instead of assassins, it out to be criminals. It should be clearer: calculating the probability that a person is a criminal if he acts suspiciously. Pr{ criminal | suspicious behavior }. That would make the proportions better. There may be many criminals in a population, but there are relatively few assassins. Constant314 (talk) 20:03, 2 October 2022 (UTC)

I would like to propose a rather radical reduction of the interpretation section. Remove both figures. The first is essentially a duplicate of an earlier figure. The second isn't really about interpretation. Remove the example as there are already plenty of examples. Remove all that Gerd Gigerenzer stuff which seems vaguely promotional, or at least advocates for a particular approach. Constant314 (talk) 23:28, 28 September 2022 (UTC)

Oppose removal of meme image. I'm not understanding what's so sacrilegious about the image. If anything, it could even make the concept its trying to illustrate easier to understand. I vote keeping it. — That Coptic Guy (talk) 16:23, 2 October 2022 (UTC)

Strongly support removal [Attacks and other off-topic content removed ~ ToBeFree (talk) 18:16, 2 October 2022 (UTC)]

1. this is a *meme*. a thing that was not expected to be known by most people. ask a 3rd grader and they'd know what astigmatism is, but never played among us. i'd bet a pretty penny for this one. better yet, ask your 60 something year old statistic professor if they know among us.

2. the only redeeming quality of this image is that it was scaled properly. that's it

3. i cannot believe people think that a "less dry" example is far better example for a style guide that requires its editors to be dry.

49.228.246.58 (talk) 17:50, 2 October 2022 (UTC)

‘’Oppose removal’’. In all honesty I think your average third grader is going to be more likely to know what Amongus is than an astigmatism. More importantly though there is a clearly connection between being an assassin and being suspicious than having a beard and astigmatism so I think it’s a more intuitive example, regardless of someone’s familiarity with the game. Ganondox (talk) 09:18, 2 November 2022 (UTC)

As I said to somebody else, the image was a humourous edit of another perfectly okay visualisation, it's not a new contribution. It's an unproductive edit, doesn't help general understanding of the concept, and could even be considered vandalism. Funny for a while but quickly gets old and isn't encyclopedic content Fuhsini (talk) 19:27, 2 October 2022 (UTC)

Definitely not vandalism. It was a good faith attempt to make the subject more accessible. Constant314 (talk) 19:59, 2 October 2022 (UTC)

Once again, skewing a visualisation to pair an internet meme is not making anything more "accessible". If anything, it's the reverse Fuhsini (talk) 20:07, 2 October 2022 (UTC)

Response from the illustrator

I'm surprised how much controversy my image has caused. At the time, I thought it was a good, albeit humourous illustration of Bayes' theorem as I felt people could relate to judging someone's innocence based on suspicion. I chose Among Us as it was topical and much less dry than disease testing. (The original scenario of astigmatism/beard was also more relatable than medical testing.)

I intentionally designed the icons to remind one (who is familiar with the game) of a crewmate but made it different enough to not violate copyright: zoom in and compare with an Among Us screenshot. My version is more curved, such as the elliptical visor and round backpack. The knife is also a different model.

Personally, unlike in other diagrams in the article, illustrating the theorem using fractions of respective areas makes it much clearer why Bayes' theorem works. I didn't mean to imply constant ratios, and the numbers I chose do not imply that. I'd be happy to modify the diagram if someone can suggest a way to make the diagram not imply that and yet be readable.

May I request that either diagram is evaluated on its merits without letting the edit war (and some rather less constructive edits in its history) affect your opinion?

Thanks,
cmɢʟee⎆τaʟκ 00:34, 29 September 2022 (UTC)

I agree with suspicion and assassins (maybe criminals would be better). That is the way people think. He is suspicious, maybe he is a criminal. My objection is that it is essentially a duplication of figure 1. As far as I am concerned, this illustration could replace figure 1. I think perhaps you could address the copyright issues on the image page on Wikimedia. Constant314 (talk) 01:00, 29 September 2022 (UTC)

Strongly support removal The image makes the whole thing less than serious, and, to be frank, makes Wikipedia look like something of a joke. GreenBeret1439 (talk) 13:29, 2 October 2022 (UTC)

There is nothing wrong with a little whimsy in Wikipedia. Constant314 (talk) 19:56, 2 October 2022 (UTC)

It was an unneeded change to an already okay visualisation, and it's less generic, thus inherently harder for people who aren't familiar with among us to understand, and wikipedia is not a site for those who are familiar with internet memes, so it's not more accessible at all. If wikipedia allowed unproductive but nonetheless "whimsical" edits all the time, then the site would be rich in vandalism, because that's exactly what vandalism is about, edits with humourous intent that are however unproductive. Fuhsini (talk) 20:05, 2 October 2022 (UTC)

Vandalism is an intentional attempt to disrupt or damage Wikipedia. The editor that made the edit explained his intentions. The edits are good faith edits. You may not think that it is not an improvement, but you are accusing a good faith editor of vandalism when it clearly is not vandalism. Personally, I think suspicion and assassin is a whole lot more meaningful than astigmatism and beard. Constant314 (talk) 20:13, 2 October 2022 (UTC)

Let me rephrase what I said. Vandalism normally revolves around humour paired with unproductivity. I may have misunderstood the editor's intentions, and I am sorry to both you and cmglee for that misunderstanding, but regardless of the editor's intentions, whether it was in good faith or not, it was an unproductive edit, and I wholly do not believe that humour should be randomly injected into an encyclopedic data bank especially when it is not paired with ultimately meaningful contribution. If it was an original depiction, then my perspective would be different. But it isn't an original depiction, it is an edit of another one, the single edit being the use of among us characters to reference an internet meme. Neither suspicion and assassins nor astigmatism and beards are more meaningful or relevant than each other on an article about a theorem about predicting probability. Fuhsini (talk) 20:17, 2 October 2022 (UTC)

That's not at all how Wikipedia defines "vandalism":

On Wikipedia, vandalism has a very specific meaning: editing (or other behavior) deliberately intended to obstruct or defeat the project's purpose, which is to create a free encyclopedia, in a variety of languages, presenting the sum of all human knowledge.

Also, although you argue that the only edit of the image was to use Among Us crewmates, there is another edit in the image which I believe is productive. In the original image, the two events "Astigmatic" and "Beard" are independent, while in the new image, the two events "An assassin" and "Being suspicious" are dependent. Bayes' Theorem is better illustrated with dependent events than with independent events, and thus the new image is better in this respect. Edderiofer (talk) 11:02, 3 October 2022 (UTC)

RfC on illustration

The following discussion is an archived record of a request for comment. Please do not modify it. No further edits should be made to this discussion. A summary of the conclusions reached follows.

Option 2. After 4.5 days of no new comments, it appears that no further input is likely. 0xDeadbeef→∞ 12:41, 14 October 2022 (UTC)

---

What should be the image in figure 2 of the current article?

• Option 1: File:Bayes theorem visualisation.svg. This compares being astigmatic vs. having a beard or not.
• Option 2: File:Bayes theorem assassin.svg. This compares being suspicious vs. being an assassin or not.
• Option 3: None. Figure 2 should not exist.
• Option 4: Something else.

0xDeadbeef 06:32, 4 October 2022 (UTC)

Discussion for RfC on illustration

• Option 4 if possible, or Option 2 in the meantime. I make the following arguments for Option 2 over Option 1:

• In Option 2, the ratios of people in the two rows "An assassin" and "Not an assassin" are different (3:1 vs 2:6); i.e. that "Being suspicious" and "An assassin" are dependent events. Contrast this with Option 1, where the ratios of people in each row are identical; i.e. "Beard" and "Astigmatic" are independent events. The former is a more relevant illustration of Bayes' Theorem.
• Option 2 is accessible even to people who are unaware of the internet meme it is based on or the game Among Us, as even without knowledge of the meme or the game, most people know what an assassin is, what suspicious behaviour is, and that assassins may exhibit suspicious behaviour more than non-assassins. The same probably cannot be said of astigmatism, and the lack of a clear link between someone being bearded and someone being astigmatic makes Option 1 less accessible than Option 2.
• Option 2's subject matter (assassins with suspicious behaviour) give a better motivation for the topic, as it is understandable to want to catch assassins before they assassinate people. The same cannot be said of bearded people or people with astigmatism.

and the following counterarguments against points people have raised in favour of outright removal under Option 3:

• Figure 1 in the article does not show how Bayes' Theorem applies; it only explains how P(User|Positive) may be found via a frequency box and makes no reference to Bayes' Theorem. Thus, having Figure 2 (in the form of Option 1 or Option 2) in the article aids the text. (If anything, I would even argue that Figure 1 should be removed as it's redundant by Figure 2.)
• While I agree that Option 2 is humorous in nature due to it being based on an internet meme, this alone is not grounds for removal. Wikipedia does not have a guideline on the use of humor in articles, although there are some Wikipedia essays on it (e.g. WP:HUMOR, which suggests four article requirements that Option 2 very much abides by; and this op-ed about the use of humor in articles).
• Option 2 is not vandalism under Wikipedia's definition, as it is a good-faith edit, so WP:VANDALISM doesn't apply.
• While the character design in Option 2 is based around the characters from Among Us, the creator of the image Cmglee has explicitly stated that they designed the icons to not violate copyright. Whether it still violates copyright may still be an issue which I will discuss further below.

I however agree that, although Option 2 is better than Option 1 and better than nothing, an ideal solution would involve a different image:

• Ideally the image should involve a subject more relevant to Bayes' Theorem, such as medical/disease testing, rather than the somewhat-more-outlandish assassin-uncovering. It may be dry, but it's evidently a classic example because it's understandable and relevant to people.
• Failing that, the characters may need to be redesigned so that there is no question about copyright. (And while we're at it, maybe make those knives look a bit more like knives and less like the character's mouths?)

but this will be up to whichever editor decides to undertake making a new image. Edderiofer (talk) 08:08, 4 October 2022 (UTC)

• Option 2 (assassins). I agree with almost everything that @Edderiofer: says. If option 2 were selected, I would later request slight modifications. However, I have a problem with Astigmatic. In that case P(A|B)=P(A|B*)=P(A). That is the most boring case and it fails entirely to illustrate the usefulness of Bayes. It is the null case, or the case where Bayes is not needed. Constant314 (talk) 14:36, 4 October 2022 (UTC)
• Option 2 - The difference between Options 2 and 1, which was the existing image, is negligible in my opinion. Option 2 offers a humorous and enlightening alternative to the original image, and despite having no foreknowledge of Among Us, could be understood just as well. Furthermore, Option 2 is of a higher resolution and is a newer image with a representation that stands out. No copyright issues as the illustrator rightfully re-created the crewmates and uploaded an svg/generic file. I also agree with much of @Edderiofer said. As such, I see no issues with it and think it should be the presiding image in the article. — That Coptic Guy (talk) 15:31, 4 October 2022 (UTC)
• Option 2. Yes, it is a pop culture reference (and as alluded elsewhere there may be copyright issues that need to be addressed), but an illustrative example that is accessible and slightly less dry has pedagogical value. Barring that, rework this option to use different context. Option 1 is not very useful due to the conditional and marginal probabilities being equivalent, e.g. $P(B\mid A)=P(B)=0.4$ provides no insight into Bayes' theorem. --Kinu ^t/_c 22:28, 4 October 2022 (UTC)
• Option 2. Bayes theorem is better understood with non-independent events. There are no issues with the reference to the popular game, since it can be understood by people who do not know the game. This makes it easier to understand and more approachable than the beard/astigmatic example. 0xDeadbeef 01:12, 5 October 2022 (UTC)
• Option 2 (weakly). Agree with what has been said so far. The example is instructive and accessible, especially to younger readers who may not know what “astigmatism” means. Whether Among Us will be relevant in 10 years is another story, but we can deal with that later.Wham Bam Rock II (talk) 10:50, 5 October 2022 (UTC)
• Option 4 if possible, or Option 2 in the meantime. Option 2 is a better alternative to option 1 as people interpret astigmatism and beard as independent events, something not useful for learning Bayes theorem. "Astigmatic" is also a hard word, I interpret its meaning to be "Strabismus" from the illustration, as I do not aware of the actual meaning, even that does not contribute to learning the actual theorem though. Still, I am a fan of a more 'conservative' and 'gold standard' example like medical screening tests. -- Sirakorn (talk) 12:33, 5 October 2022 (UTC)
• Option 4 or Option 2 until a suitable replacement can be found. GreenBeret1439 (talk) 01:15, 6 October 2022 (UTC)
• Option 2 Glad we now have a more reasoned discussion. On reading the feedback, I concur that the astigmatism/beard one is inferior (independent events and same probabilities). I still think the suspicion/guilt scenario is clearer than disease testing and would like to keep the labels "Assassin" and "Being suspicious" as their initials are "A" and "B" (saves readers having to remember, "What's A (or B) again?") Frankly, I see nothing wrong with referring to a meme as long as it doesn't prevent someone unfamiliar with the meme from understanding it. However, if it is felt that the icon is too similar to Among Us, I'll redraw it with another icon: let me know. Cheers, cmɢʟee⎆τaʟκ 22:06, 6 October 2022 (UTC)
• Option 4 if Option 2 causes yet another edit war when the full protection is lifted. DizzyTheMan (talk) 17:59, 8 October 2022 (UTC)

  If they go against consensus, then they are editing disruptively. There is no point to !vote for Option 4 if the only intention is to make people who dislike Among Us happy. 0xDeadbeef→∞ 03:11, 9 October 2022 (UTC)

  In addition, if people are removing the image without understanding that the image was added as a result of consensus, then that might indicate that they are probably not frequent Wikipedia editors, and so an easier solution to prevent edit-warring might just be to keep the article semi-protected. Edderiofer (talk) 06:58, 9 October 2022 (UTC)

• Option 4 as I didn't find either image particularly helpful. John (talk) 13:27, 9 October 2022 (UTC)
• Option 2 as humour is a very useful pedagogical tool. I don't like Among Us (I !voted keep in a deletion discussion for Template:User hate Among Us which I have on my user page) yet this is a good example regardless. Chess (talk) (please use {{reply to|Chess}} on reply) 22:23, 9 October 2022 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.