Wikipedia:Articles for deletion/Method of Equal Shares


 * The following discussion is an archived debate of the proposed deletion of the article below. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review).  No further edits should be made to this page.

The result was no consensus. I don't see a 3rd relist bringing additional input. A merger could be discussed outside this discussion Star   Mississippi  15:41, 19 March 2022 (UTC)

Method of Equal Shares

 * – ( View AfD View log | edits since nomination)

This concept seems to fail WP:GNG. Gscholar query for "Method of Equal Shares" gives only two minor hits - a preprint and one paper with a single citation. This has been recently nominated for deletion discussion on pl wiki (pl:Wikipedia:Poczekalnia/artykuły/2022:02:23:Metoda równych udziałów), with a note that the creator of this article likely has a COI (the only other page they created is that for one of the co-authors of the paper cited in the Method., see ongoing Articles for deletion/Piotr Skowron). Piotr Konieczny aka Prokonsul Piotrus&#124; reply here 13:09, 24 February 2022 (UTC) Please add new comments below this notice. Thanks, Qwaiiplayer (talk) 13:59, 4 March 2022 (UTC)
 * Note: This discussion has been included in the list of Politics-related deletion discussions. Piotr Konieczny aka Prokonsul Piotrus&#124; reply here  13:09, 24 February 2022 (UTC)
 * The reason why "Method of Equal Shares" gives only 2 hits is because the rule has been previously called Rule X in the early papers. The original paper has 33 citations. This is the list of papers studying the rule:
 * Dominik Peters, Grzegorz Pierczynski, Piotr Skowron: Proportional Participatory Budgeting with Cardinal Utilities. NeurIPS'2021.
 * Dominik Peters, Piotr Skowron: Proportionality and the Limits of Welfarism. EC 2020: 793-794.
 * Rupert Freeman, Anson Kahng, David M. Pennock: Proportionality in Approval-Based Elections With a Variable Number of Winners. IJCAI 2020: 132-138.
 * Virginie Do, Matthieu Hervouin, Jérôme Lang, Piotr Skowron: Online Approval Committee Elections. CoRR abs/2202.06830 (2022).
 * Piotr Skowron, Adrian Górecki, Proportional Public Decisions, AAAI-2022.
 * Roy Fairstein, Reshef Meir, Dan Vilenchik, Kobi Gal: Welfare vs. Representation in Participatory Budgeting. CoRR abs/2201.07546 (2022).
 * Martin Lackner, Piotr Skowron: Approval-Based Committee Voting: Axioms, Algorithms, and Applications. CoRR abs/2007.01795 (2020).
 * User:Motzarcik&#124; reply here 14:00, 24 February 2022 (UTC)
 * As I pointed out in pl discussion, even if the above papers discuss this theory in WP:SIGCOV (which is not apparent from their titles), they are almost all co-authored by one person (Piotr Skowron). There is no evidence of wider recognition for this theory. A short summary could be merged to Skowron's biography. --Piotr Konieczny aka Prokonsul Piotrus&#124; reply here 13:21, 24 February 2022 (UTC)
 * Delete At best this could be a sentence in Participatory budgeting algorithm. Lamona (talk) 19:01, 26 February 2022 (UTC)
 * A few arguments for keeping:
 * The method is taught at courses at major universities and is described in a book. For example, it is taught at Harvard https://sites.google.com/view/optdemocracy22/schedule (the lecture about participatory budgeting, the name of the rule used there is Rule X), and at the University of Warsaw. It is described in the book being published by Springer (the open-access version is here: https://arxiv.org/abs/2007.01795). There, the method is called Rule X, and is mentioned 50+ times. Springer is a serious scientific publisher; the topics covered in books are those which are considered established in the literature.
 * Analogous pages exist for analogous methods.. Examples include Proportional approval voting, Sequential proportional approval voting, Satisfaction approval voting, and many others. The knowledge of how to implement proportionality in committee elections is covered at Wikipedia. It seems reasonable that the knowledge of how to implement proportionality in the context of participatory budgeting (PB) should be available for common readers, especially given that PB elections are becoming increasingly common.
 * The topic is relevant for many people. Many people participate in elections, and many cities are now considering introducing participatory budgeting or changing their election methods. It is important that the knowledge of how different methods work is available and easily accessible for general public audience. When talking to a few cities, I noticed people where surprised, when they've heard the example with red and blue projects (the one that is mentioned on the page). They did not expect that the used method returns such results and they were not aware that proportionality can be achieved. In a way, why should they, since this knowledge is not easily accessible.
 * Citations. Regarding recognition, pointed out by User:Piotrus, the articles mentioned above are those that define the method, and study their properties. In order to check how often the method is referenced in the literature, one would need to look at papers that cite these works. For example, the query <<"Rule X" Peters Skowron>> in Google Scholar returns 46 hits. Yet, some papers refer to this method not using its name explicitly, so the number of references can be larger. Is couple of dozens of references in the scientific publications a lot (given that the method has been proposed three years ago)? I think that yes, but this is a subjective criterion. For comparison the query "Sequential proportional approval voting" returns 35 hits. Note that WP:GNG does not specify how many times the name should be mentioned.
 *  Satisfaction of WP:GNG. Notice that the papers that I have mentioned went through the review process, where their soundness and significance was judged by independent researchers. The fact that there were several papers published on the specific rule in top venues in computational social choice (NeurIPS, AAAI, IJCAI, EC; I recommend looking for "ranking of computer science conferences") means that this research and knowledge is accepted and important for the relevant community. In particular there exist reliable and independent sources for the knowledge. — Preceding unsigned comment added by Motzarcik (talk • contribs) 05:49, 27 February 2022 (UTC)
 *  Relisted to generate a more thorough discussion and clearer consensus.

Please add new comments below this notice. Thanks,  Sandstein   17:03, 11 March 2022 (UTC)
 * Motzarcik we would still need significant sources to support the article. You seem to know of sources so please list them. Numbers of hits on searches is not sufficient given all of the shortcomings of search engines. Also note that Google scholar does list some articles by the authors you mention (Peters and Skowron) but the number of citations given for them (10, 9, 7, 5, 1, 1) is not very high. Lamona (talk) 19:50, 8 March 2022 (UTC)
 * All things considered, this is the problem of WP:ITSUSEFUL, I am afraid. The author believes that this is a helpful addition to Wikipedia, as it will educate the readers, alas, not everything that is useful is also NOTABLE... Piotr Konieczny aka Prokonsul Piotrus&#124; reply here 12:42, 9 March 2022 (UTC)
 * Indeed, I believe this is a useful addition. Yet, note that on top of that, I gave evidence of notability in the field of Computational social choice. Please refer to them. Motzarcik (talk) 12:32, 10 March 2022 (UTC)
 * Lamona, I've updated the reference list in the article. I hope this will be now clearer how this is referenced in the literature and how this relates to other concepts and topics (see the "properties" section). The citation numbers for the works that define the method in different settings are actually higher (12, 33, 3, 13), giving 61 in total. This is a lot for papers in Computational social choice. For all the works that are now referenced in the article, these are (12, 33, 3, 13, 86, 39, 220, 7, 51, 66, 81) (I put the numbers in the order in which they are used in the article). Motzarcik (talk) 12:08, 10 March 2022 (UTC)
 * Motzarcik I looked at all of those I can access but do not see "method of equal shares" in them. If this is a named method, I would expect to see it in them. Taking this one for example, if you search on "equal shares" you get zero hits in the source. That would make the statement: "MES with adjusting initial budget, PAV and Phragmen's voting rules can all be viewed as an extension of the D'Hondt method to the setting where the voters can vote for individual candidates rather than for political parties" original research because that conclusion is not reached in the article. Three of them have information about Rule X. At best, this article could be titled "Rule X" and only those sources that describe Rule X would be referenced. Lamona (talk) 14:15, 10 March 2022 (UTC)
 * Lamona, thanks for the comments: Let me first explain the issue with the name. In this Wiki article, there is the statement "In early papers the method has been also referred to as Rule X", provided as footnote 1. Indeed, in paper [2], in the version that has been published (https://openreview.net/pdf?id=5rm0b_fsNZ) the authors wrote that they decided to name the rule "Method of Equal Shares". Indeed, in the original paper that first defined the rule ([3]), the authors wrote "which we give the preliminary name Rule X", suggesting that this is a temporary choice. Note that "Method of Equal Shares" and "Rule X" is exactly the same rule, and the authors of [3] are the superset of the authors of [2], thus the authors who initially defined the method are those who decided to change the name. I've seen three papers in review which are using the name MES already. Indeed, we can use in this article the name "Rule X", but I think it would be better to respect the decision of the authors, also given that the name "Method of Equal Shares" is more informative (intuitively, "equal shares" appear twice: first, each voter is assigned an equal share of the budget, and second, once a candidate is selected, each voter who voted for the candidate is burden with an equal share of the cost of the candidate). Perhaps, saying that the method is also called Rule X in the first line of the article (not in the footnote) would avoid further confusion. What do you think? Regarding the D'Hondt method: in [3] (https://arxiv.org/pdf/1911.11747.pdf), Proposition 2 says that Rule X (MES) satisfies priceability, and Theorem 3 says that priceability implies the fact that the method behaves as the D'Hondt method in the apportionment setting. This is exactly the same interpretation as the one taken in [12]. Thus, the statement "MES with adjusting initial budget, PAV and Phragmen's voting rules can all be viewed as an extension of the D'Hondt method to the setting where the voters can vote for individual candidates rather than for political parties" is not original research, but has been proved in independent sources. However, you are right that I did not add the mentioned reference in the right place, which I am now correcting. I hope this is now clear. Please let me know. Motzarcik (talk) 17:04, 10 March 2022 (UTC)
 * We still need sources that refer to this as the "method of equal shares". I didn't find any, so if you have some, please link to them here. Lamona (talk) 22:28, 10 March 2022 (UTC)
 * Sigh That article nowhere has the term "method of equal squares" or even "equal squares". To have an article that is presumably about a method called "equal squares" you need sources that call it "equal squares". That article does mention Rule X but is not significantly about Rule X. Lamona (talk) 22:29, 11 March 2022 (UTC)
 * That seems wrong on two counts. Excluding sources because they happen to use the old name for something seems like a daft idea, that would have serious impact upon some countries, for starters.  &#9786;  And for an article that supposedly isn't about Rule X, it seems to devote a fair number of pages, and an entire appendix, to proving multiple theorems about Rule X.  Uncle G (talk) 02:34, 12 March 2022 (UTC)
 * Uncle G, you misunderstood me, perhaps because this thread is so long. We need some evidence that MES is the new name, and that needs to be shown in sources. So it's not because it's the new name, it's because we don't have evidence that it is the new name. I'm suggesting that unless we have such evidence, the article should be called "Rule X" because that is what the sources listed are about. It is possible that there are sufficient sources for Rule X, I really don't know, but the article appears to have a fair amount of WP:OR, which would need to be removed. Also, the article that Motzarcik lists as evidence that Rule X has been changed to be MES and they are the same, actually says: "Recently, Peters and Skowron [2020] introduced an aggregation rule for approval-based committee elections that they called Rule X. In that setting the rule satisfies a combination of appealing proportionality properties. Here, we extend it to the more general model of participatory budgeting, that is, to the model with arbitrary costs and utilities. We will call this rule the Method of Equal Shares (in short, MES)" which sounds to me that they are not one and the same, but one is derived from the other. So I ask: cannot the MES article stand on its own, without being sourced primarily to articles about Rule X? Given that the cited article that coins the term "MES" is from 2022 it may be WP:TOOSOON. Lamona (talk) 16:46, 12 March 2022 (UTC)
 * Lamona: Rule X is exactly the same as MES. The paper that you refer to has also its arxiv version, which is older (2020), and where the authors use the name Rule X: https://arxiv.org/pdf/2008.13276.pdf. The two articles contain definitions of the rules, so you can check that this is indeed the new name that is used by the authors (I guess, the authors got a feedback, that the name Rule X is not informative and decided for a change). The authors write that they extend the rule to the setting of PB, because if costs of projects are equal to one this is exactly the setting of committee elections, and the two definitions are then exactly the same. This is common that the same object is referred to differently in different literature. For example, Turing machine was first called "a-machine" in the original work. Sequential proportional approval voting is sometimes called greedy proportional approval voting or reweighed approval voting. Still we use sources using both names. Of course, we can call the article "Rule X", but I do not see a clear reason why we should change the name that was suggested by the authors. Lamona, what is the criterion for using the name? If I write to the authors with a question, and if they decide to update the first article in ArXiv (the one that uses the name Rule X) to use the new name, will you be satisfied? Motzarcik (talk) 18:34, 12 March 2022 (UTC)
 * Lamona, I've changed the page to include the following explanation in the first sentence "(in early papers the method has been also referred to as Rule X[2,3,4], but since 2022 the authors started using the name "method of equal shares"[1])". The link to [1] is indeed the link to the version that uses the name MES, and that says that this is indeed Rule X. For the paper which you say only mentions Rule X (https://arxiv.org/pdf/1911.11747.pdf), note that except for introducing the definitions, Theorems 7, 8, 9 are purely about Rule X. Proposition 1, Proposition 2, and Theorem 3 are about the axiom of priceability, which is satisfied only by Rule X and Phragmen's sequential rule (Proposition 2), so these results are also about Rule X (in particular, Theorem 3 implies that Rule X extends the D'Hond't method). Theorems 4 and 10 show that welfarist rules (like Proportional Approval Voting (PAV)) do not satisfy some of the axioms of proportionality. Only Theorems 5 and 6 are about PAV. This paper is very much about Rule X (it introduces the rule, and introduces axioms, by which it compares Rule X to PAV---and more generally to welfarist rules---and to the Phragmen sequential rule). Motzarcik (talk) 14:42, 12 March 2022 (UTC)
 *  Relisted to generate a more thorough discussion and clearer consensus.
 * I think the references are sufficient to show notability. Even though the same author - Skowron - appears in most references, there are still many other authors from different research groups. This is sufficient to show that the topic has general relevance to the field. Erel Segal (talk) 13:10, 18 March 2022 (UTC)


 * The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review). No further edits should be made to this page.