Wikipedia talk:Article size

Arbitrary break: review of proposed changes and responses

Considering how long this section has gotten, I figured I'd create a separate subsection per the talk page guideline with a summary of the proposed changes and previous responses if anyone else wants to close the larger part of this section and continue the discussion down here.

Proposals

1. Raise the three highest word count ranges (WCRs): Proposed by Ladtrack and changes to their "What to do" column recommendations proposed by Mburrell. Its specific changes are:
   1. Increase the above-15,000 WCR to above-23,000 words, and change its "What to do" recommendation to "An interested editor could divide or trim this article. Otherwise, leave it alone, it's fine."
   2. Increase the above-9,000 WCR to above-14,000 words, and change its "What to do" recommendation to "One could consider dividing or trimming the article, though the scope of the topic can sometimes justify the added reading materials. If not looking for an editing project, leave it alone, it's fine."
   3. Increase the above-8,000 WCR to above-12,000 words, and change its "What to do" recommendation from "May be divided or trimmed if one is looking for a project. Otherwise, leave it alone, it's fine."
2. Reorganize the lower WCRs: Proposed by WhatamIdoing, along with changes to the "What to do" column recommendations for all WCRs. Its specific changes are:
   1. Keep the above-15,000 WCR as is, and add "About 1 in 1,000 articles are this long, and most of them are stand-alone lists. See also Special:LongPages" to its "What to do" recommendation.
   2. Increase the above-9,000 WCR to above-10,000, and add "Only 1 in 250 of Wikipedia's articles reach this length. Will take most people almost an hour to read. The longer the article, the fewer people who will read the whole thing" to its "What to do" recommendation and change its wording from "justify the added reading material" to "justify the unusual length".
   3. Consolidate and lower the above-8,000 and below-6,000 WCRs into one above-5,000 WCR with its "What to do" recommendation as "Divide or trim only if that makes a better article in terms of organization, encyclopedic style, etc., and not merely due to the being longer than most well-developed articles. 98% of all Wikipedia articles, and 67% of all Featured articles, are shorter than this. Will take most people 20–30 minutes to read."
   4. Add below-1,000 and below-350 WCRs with "The median length of a non-stub article as of 2024. Will take most people five minutes to read" and "The median article size as of 2024. Will take most people two minutes to read" as their "What to do recommendations" respectively.
   5. Lower the below-150 WCR to below-100, and change its "What to do" recommendation to "One of Wikipedia's shortest articles. Please try to expand it, or consider merging it with a related article. See also Special:Shortpages."

Responses

• Of the 17 editors who have left comments in the discussion, I count 5 editors (Nikkimaria, Z1720, Noleander, Trainsandotherthings, Borsoka) who clearly argued to keep the guideline as is, with 12 editors expressing support for one of the two proposals at least in part or in the rationale for the proposal.
• For Proposal #1 (the Ladtrack-Mburrell proposal), I count 6 of 17 editors (Ladtrack, Mburrell, EEng, Hawkeye7, Komonzia, SchroCat) that have explicitly expressed support for it or its rationale, while 10 of 17 editors (Nikkimaria, Z1720, Moxy, CommonKnowledgeCreator, Noleander, HJ Mitchell, Kusma, Trainsandotherthings, Grnrchst, Borsoka) have explicitly opposed it's changes or rationale.
• For Proposal #2 (the WhatamIdoing proposal), I count 6 of 17 editors (WhatamIdoing, Moxy, HJ Mitchell, CommonKnowledgeCreator, Kusma, Grnrchst) that have clearly expressed support for this proposal or for parts of it. However, 5 editors (Ladtrack, EEng, Nikkimaria, Mburrell, Komonzia) have not expressed any opinion about it since their comments were left 3 weeks to 13 weeks before this proposal was made.

So, as far as I can tell, there is a consensus among the editors who have left comments here to change to guideline, but not the changes included in Proposal #1. While Proposal #2 does not have a clear consensus against it, there are points of disagreement among editors about it.

Points of disagreement on Proposal #2

1. WCR reorganization: Since this proposal was posted on 21 February, 5 editors (Noleander, HJ Mitchell, Z1720, Trainsandotherthings, Borsoka) expressed support for keeping all of the WCRs as they currently are, while 4 editors (WhatamIdoing, Moxy, CommonKnowledgeCreator, Grnrchst) expressed support for the general reorganization, and 1 editor (Kusma) suggested lowering the WCRs.
2. Lowest WCR: 2 editors (CommonKnowledgeCreator, Grnrchst) suggested changing the lowest WCR to below-250 for consistency with the stub guideline, to which 1 editor (WhatamIdoing) expressed opposition based on other recommendations in the stub guideline and the merging information page's list of reasons for merging articles.
3. Average Wikipedia article word count statistics: While 3 editors (WhatamIdoing, Moxy, HJ Mitchell) expressed support for including such statistics in the "What to do" recommendations, 2 editors (Hawkeye7, CommonKnowledgeCreator) expressed skepticism about doing so when considering that most articles are stubs and because the word counts of lengthy articles are more likely due to an implicit consensus or silence rather than an explicit consensus. Note: The current revision of the guideline also has a "Measures of article size" table below the size guideline that notes that the median article length is 350 words and that 100 words is the upper limit of the lowest word count quintile—which are proposed WCRs under Proposal #2.
4. Featured article (FA) and Good article (GA) word count statistics: 1 editor (CommonKnowledgeCreator) expressed support for including FA/GA statistics in the "What to do" recommendations for the two highest WCRs (above-15,000 and proposed above-10,000) rather than the middle WCR to provide specific examples of where an explicit community consensus has made exceptions to the guideline that emphasize to the editor reading the WCRs realizes that they are recommended best practices rather than required standards, which another editor (WhatamIdoing) opposed. However, while not explicitly expressing support or explicitly opposing including the FA/GA statistics in other WCR "What to do" recommendations, 4 editors (HJ Mitchell, Kusma, Hawkeye7, WhatamIdoing) said in their comments that they believe that the word count limits should be seen as recommendations rather than requirements and that the community is already flexible with the guideline.
5. Estimated reading times (ERTs): 5 editors (WhatamIdoing, Moxy, HJ Mitchell, CommonKnowledgeCreator, Grnrchst) expressed support for doing this, while 3 editors (Ladtrack, EEng, Hawkeye7) argued that reader attention span should not be a consideration for the guideline's recommendations. Note: While Proposal #2 includes "The longer the article, the fewer people who will read the whole thing" as part of the "What to do" recommendation its proposed above-10,000 WCR, the ERTs are not based on attention span but the average reading speed for a typical adult without diminished comprehension. -- CommonKnowledgeCreator (talk) 04:17, 24 February 2026 (UTC)

@Ladtrack, EEng, Nikkimaria, Mburrell, and Komonzia: Apologies for the ping, but I wanted to see if you have opinions about the proposal listed above that posted by WhatamIdoing on 21 February 2026 since none of you have left subsequent comments here since then. -- CommonKnowledgeCreator (talk) 04:47, 24 February 2026 (UTC)

I'm fine with the what-to-do changes mentioned under Proposals. I'm not convinced about the specific number changes proposed. Nikkimaria (talk) 04:55, 24 February 2026 (UTC)

The two proposals (1,2) above are suggesting three orthogonal changes:

a) Changing the size numbers

b) Adding details & stats to each size-range

c) Reducing the number of size-ranges from five to four by consolidating two of them

Those are three distinct choices, and I suspect that most us will have differing answers to them. Personally, I'd choose "No, Yes, Dont Care". Nikkimaria (immed above) is "No, Yes, Yes" (I think). I can't imagine anyone objecting to (b) though there's bound to be differences of opinion about which statistics. To come to closure, it might be best to limit the discussion to (a). Or at least keep the three orthogonal decisions separate. Noleander (talk) 05:03, 24 February 2026 (UTC)

I think the interplay between a and c makes things a bit more complicated. An argument given above is that the gap between 8k and 9k is small. a could solve that. c could solve that. Do we need both? Nikkimaria (talk) 05:30, 24 February 2026 (UTC)

Well, if you look at my proposal on 21 February, I didn't actually suggest "c". @Noleander, I proposed increasing the number of size ranges to six.

The current list is: 150 – 6,000 – 8,000 – 9,000 – 15,000.

My proposed list is: 100 – 350 – 1,000 – 5,000 – 10,000 – 15,000.

I have never heard anyone try to defend the 8K vs 9K gap. It's weird enough that I've thought about checking the history (typo? someone WP:GAMING a dispute?) but never bothered. WhatamIdoing (talk) 06:48, 24 February 2026 (UTC)

@Nikkimaria: To clarify, do you support merging the above-8,000 and above-9,000 word count ranges? And if so, would you prefer the range to be above-8,000 or above-9,000? @WhatamIdoing: I apologize if my summary of the previous section is the source for Noleander's confusion about the reorganization proposal. -- CommonKnowledgeCreator (talk) 19:13, 25 February 2026 (UTC)

Sure, but it seems like that wasn't what was proposed. Nikkimaria (talk) 23:45, 25 February 2026 (UTC)

Like I said, that probably owes to my poor summary of the proposals. Do you have preference whether the word count range is above-8,000 or above-9,000? -- CommonKnowledgeCreator (talk) 00:08, 26 February 2026 (UTC)

@Nikkimaria, when you say "what was proposed", are you talking about the proposal to change the table to say 150 – 6,000 – 12,000 – 14,000 – 23,000?

(Also, I've just noticed that Wikipedia:Article size#Readability says 10,000 words. The different sections should match, no?) WhatamIdoing (talk) 00:32, 26 February 2026 (UTC)

I was actually talking about your comment immediately above the one one I was responding to, but I think we're overflowing with proposals a bit here. Nikkimaria (talk) 00:33, 26 February 2026 (UTC)

There's really only two proposals. There is a consensus against Proposal #1. There are points of disagreement about aspects of Proposal #2. What we're attempting to do is gauge what the editors who do not oppose it think about different aspects of it. -- CommonKnowledgeCreator (talk) 03:32, 26 February 2026 (UTC)

I don't have a strong preference between those options in isolation - it would depend on what else if anything is getting changed. Nikkimaria (talk) 00:35, 26 February 2026 (UTC)

Does "it would depend on what else if anything is getting changed" mean something closer to "It would depend on whether we add new levels for 350 and 1,000" or does it mean something closer to "It would depend on whether we add percentages of FAs at each level"? WhatamIdoing (talk) 03:54, 26 February 2026 (UTC)

It would depend on whether there are any changes to numbers, whether alterations to the existing numbers, additions of new numbers, or removals of current numbers. For me the percentages is an unrelated question. Nikkimaria (talk) 04:56, 26 February 2026 (UTC)

The two proposals (1,2) above are suggesting three orthogonal changes I don't think we're proposing making the article size guideline to include perpendicularity or statistical independence (nor is it clear to me how this would be done). :) -- CommonKnowledgeCreator (talk) 01:26, 25 February 2026 (UTC)

The methodologies used to derive both proposals are more sound than what is currently being used. So far, I have been slightly more in favour of Proposal #1 (increasing the limit with some clear case-by-case treatment or justifications given) than Proposal #2 because that's just my own personal pain point, but I'm aware that we have to cater for most readers, not readers like me, so I am not against Proposal #2 which does not raise the upper word count threshold.

My own pain point is for articles about war (where I mentioned there are sometimes too many facets to cover within a reasonable word count), it can be a good illustrative test case for what should be done at the top end of the word counts. In that context - should they be split into separate articles about "political reactions / political background" vs "battlefield and front line story"? Or maybe stick to splitting it based on time? Will that lead to too much fractured editing? Is it better to have the same topic covered in one place?

Personally: I would tolerate all of the topics in one place, with a large word count, if organised by sections. But I think I'm very much in the minority - as might most people who would comment here - we are mostly editors and not just readers.

I can see World War II was mentioned above as one case. That specific war, since it was a world war, can be split geographically as well as by topic. The same doesn't count for Yom Kippur War and War in Donbas. Similarly, the Yom Kippur War officially only lasted two weeks, so it might not be amenable to splitting across time.

- 🔥Komonzia (message) 11:17, 24 February 2026 (UTC)

The vast majority of my content creation work is in various aspects of warfare, and at FAC. (77 successful warfare FACs.) Yet I am strongly against an increase in the word limit guidelines. I think Noleander's comments are relevant here, and that summary style and appropriate sub-articles can eliminate the need for more than 9,000 words in all - or very nearly all - cases. To give some examples from current FAs - and so inevitably not addressing the specific cases above - Silesian Wars and Second War of Scottish Independence both come in at fewer than 5,000 words and even Punic Wars is just a shade over 8,000. The first and third of these might be considered candidates for higher word counts as each consists of several separate wars. Gog the Mild (talk) 13:33, 24 February 2026 (UTC)

• I am against WhatamIdoing's proposal because it essentially raises the "probably should be trimmed" threshold from 9,000 words to 10,000 words: essentially a 10% increase in article size. Since many editors are against increasing the size limits of articles, I think this change would be against that consensus. I am not against adding the "1 in X" wording and the other wording changes. I would also support stronger language than "recommended": too many editors respond that their article is the exception to the article size rule, citing the text as currently written (examples at Manchester's FAR, Frank Sinatra's GAR, Horatio Nelson's GAR). Having language that states that large articles are a rare exception would be helpful for this guideline. Z1720 (talk) 20:40, 24 February 2026 (UTC)

  I am not against adding the "1 in X" wording and the other wording changes. I would also support stronger language than "recommended": too many editors respond that their article is the exception to the article size rule, citing the text as currently written... Having language that states that large articles are a rare exception would be helpful for this guideline. This is why I proposed including FA and GA word count statistics in the first place, and proposed noting that only 8 FAs and 16 GAs are longer than 15,000 words and that 767 FAs and GAs in total are longer than 10,000 words in the "What to do" recommendations.

  I am against WhatamIdoing's proposal because it essentially raises the "probably should be trimmed" threshold from 9,000 words to 10,000 words: essentially a 10% increase in article size. Since many editors are against increasing the size limits of articles, I think this change would be against that consensus. Considering that the "What to do" recommendations for the top two word count ranges in the proposal and the current guideline say "Almost certainly should be divided or trimmed" and "Probably should be divided or trimmed", I view only the above-15,000 word count range as the only real limit imposed by the guideline. I'd have to look at them all again, but I don't remember editors who expressed opposition to "raising the limits" mostly referencing the "What to do" recommendations language in their comments, so it was only clear to me that most editors were opposing Proposal #1. Moreover, I believe changing the current above-9,000 word count range to the proposed above-10,000 word count range would be more sensible because it would account for basically the entire word count distribution of FA and GA articles; 96% of FAs and 99% of GAs are shorter than 10,000 words. -- CommonKnowledgeCreator (talk) 01:03, 25 February 2026 (UTC)

• When I review GAs and FAs, I will usually comment on article size when it is above 10,000, and make a judgment call when it is between 9,000 and 10,000. This is because editors will say, "It's barely over the 9,000 'almost certainly be trimmed' guidance, so we shouldn't make a big deal of it". For example, at Manchester's FAR, editors argued that the article could be just under 10K words. After I conducted a copyedit, the article stands at 8,225 words. If the limit was raised to 10K, which is a 10% increase from what it is now, editors will argue that articles at 10,250 are "close enough" and resist calls to trim. I would rather have lower limits so that editors have to have better arguments than "close enough" when the article is at 9,500 words. Z1720 (talk) 01:36, 25 February 2026 (UTC)

  @Z1720, if the proposed list were 100 – 350 – 1,000 – 5,000 – 9,000 – 15,000, would that be acceptable to you? WhatamIdoing (talk) 01:51, 25 February 2026 (UTC)

• @WhatamIdoing: For me, 9,000 is OK, 8,000 is preferred. I do think that 8,000 and 9,000 should be merged together, which is what your proposal does. Z1720 (talk) 02:33, 25 February 2026 (UTC)

  What makes you prefer 8K to 9K for that article around the 99.5th percentile?

  In my dataset, there are 20 articles out of 10,000 that fall into the gap between 8K and 9K. There are just 12 that fall into the gap between 9K and 10K. Does it really matter where the line is drawn, when only 1 in ~500 articles could be affected? WhatamIdoing (talk) 03:48, 25 February 2026 (UTC)

  I think Z1720's rationale for not increasing the second word count range to above-10,000 is a sufficient justification for not doing so, in addition to the express opposition to raising the word count ranges from 7 other editors in this discussion (Nikkimaria, Noleander, HJ Mitchell, Kusma, Trainsandotherthings, Borsoka, Gog the Mild). If having above-9,000 words as one of the guideline word count ranges helps keep FA and GA candidates under 10,000 words during the review process, then that should be enough of a reason to not increase the second word count range. While Z1720 and Borsoka have expressed support specifically for an above-8,000 word count range, and Kusma called for the lowering word count ranges across-the-board, considering that 7 editors (Ladtrack, Mburrell, EEng, Hawkeye7, Komonzia, SchroCat, Jo-Jo Eumerus) would prefer the word count ranges to be higher, I think that's a reason for not lowering the second word count range to above-8,000, but nonetheless continue to consider eliminating the current above-8,000 word count range since Z1720's willingness to support that part of the reorganization proposal puts its support at 5 editors (WhatamIdoing, Moxy, CommonKnowledgeCreator, Grnrchst, Z1720) with 3 remaining opposed (Noleander, Trainsandotherthings, Borsoka). -- CommonKnowledgeCreator (talk) 17:48, 25 February 2026 (UTC)

  So... I haven't found a rationale beyond WP:ILIKEIT in the comments from @Z1720. There's an argument against having any cutoff, because no matter what the cutoff is, some editors are going to say that whatever they wrote is close enough. He says that he personally prefers 8K to 9K, but it's just a personal preference. WhatamIdoing (talk) 22:38, 25 February 2026 (UTC)

@WhatamIdoing: Others have summarised good points as to why to keep it at 8,000, and past discussions also mention why a lower number is better:

• As demonstrated in a previous discussion on this topic, several academic journals recommend 8,000, some as a lower limit and some as a maximum (Examples: Past & Present, Sage Journals, Early Modern Women Interdisciplinary Journal, Global Labour Journal, BGSU SSCI Journal Publishing Guide, Futures). Some of the sources use 8,000 as the lower range: if Wikipedia keeps it at 8,000, then someone claiming an article can be larger at 10,500 will have more difficulty using this guideline to justify the length.
• According to several sources that were previously cited in past discussions on this topic (such as and ) the attention span of an average adult learner wanes after 15-20 minutes. Keeping the attention span of a reader becomes more difficult with a longer article (as clicking onto a new article offers new stimuli that can help with attention).
• From what I read above, the consensus in this discussion is to not raise the limit (and consensus may be building towards lowering it instead). I think it would be against that consensus to then raise the limit to 10,000.
• While I want more detail in my articles, like William Lyon Mackenzie, there have been several books written about him that I cited in the article. I am interested in the topic, so I read them all. The average reader is not as interested as I am. If they were, they would click on the sources I cited and start reading. Wikipedia articles should be introductions to the topic, not complete documents with all the details editors can find.

These would suggest that article limits should go down, not raise them to 10,000. Z1720 (talk) 23:25, 25 February 2026 (UTC)

• Academic journal articles should not be a model for Wikipedia articles since Wikpedia is an encyclopedia rather than an academic journal that publishes original research.
• Editors who want longer articles have suggested that reader attention span should not be a consideration in the article length since most readers don't read articles from beginning to end and not necessarily due to attention span. The estimated reading times that have been proposed being included are based on the average reading speed of a typical adult without diminished comprehension.
• As the editor who has been reviewing the comments left by the 19 editors here, I don't see a clear a consensus in favor of lowering or raising the word count ranges (or any specific changes really), just a consensus to change the guideline rather than keep it as is. -- CommonKnowledgeCreator (talk) 23:48, 25 February 2026 (UTC)

  Z120, here are my thoughts:

  1. This is two reasons; the first is that some others use 8K – including some using 8K as their minimum word count – so we should make 8K our maximum. The second is that no matter what the threshold number is, some Wikipedia editors will want to exceed it and claim that theirs isn't. Your second sounds like someone designing a weight-loss diet so that when the dieter inevitably cheats, they'll end up in the "right" place. (This works, BTW: if you tell people to follow an extreme Atkins diet, and they'll end up on a reasonable low-carb diet.)
  2. If the average adult reads 250 wpm and loses interest after 20 minutes, then you should be supporting my proposal to lower the mid-range item from 6K to 5K, because 250 wpm × 20 minutes = 5,000 words.
  3. From what I read above, the consensus is against raising the current 15K up to 23K. (I also oppose raising the recommended maximum.) I see very few editors engaging with my proposal, which is: to ▼ lower 6K to 5K, to ▲ raise 9K to 10K, and ▷ keep 15K exactly the same.
  4. I agree. But I don't think that my proposal is incompatible with that. It sounds, though, like you might prefer a list that runs 100 – 350 – 1,000 – 5,000 – 7,500 –10,000 – 15,000, so that you have another 'hook' for complaining that an article is longer than you prefer.

  WhatamIdoing (talk) 00:05, 26 February 2026 (UTC)

  I think this discussion should just be closed. At this point, I don't think we're going to get a consensus for any specific changes. -- CommonKnowledgeCreator (talk) 23:32, 26 February 2026 (UTC)

  I think we have little or no opposition to my proposal except for changing the 8K-and-9K line to 10K. I don't see anybody arguing to keep that small (~12%) gap, but it looks like a couple of editors are uncomfortable with 10K and would prefer that either 8K or 9K were chosen. WhatamIdoing (talk) 23:45, 26 February 2026 (UTC)

  I think you could say there was a qualified consensus for the above-10,000 word count range. 7 editors (Ladtrack, Mburrell, EEng, Hawkeye7, Komonzia, SchroCat, Jo-Jo Eumerus) have expressed a general preference for higher word count ranges, and 4 editors (WhatamIdoing, Moxy, CommonKnowledgeCreator, Grnrchst) who initially supported the reorganization proposal did not oppose the above-10,000 range. 19 editors in total have left comments in the discussion, which would imply that 11 of 19 would support such a change—which is not a very large majority. Also, after talking to Z1720 and reviewing the comments by Noleander and Gog the Mild, I'm less sure that I'm favor of it now based on their experiences with FA and GA nominations. However, since the 7 editors in favor of higher word count ranges are not really engaging with the proposal, I don't know whether they really support the change. Also, per their expressed preference, they probably would oppose lowering the below-6,000 word count range, and by my count, only 8 editors (WhatamIdoing, Moxy, CommonKnowledgeCreator, Kusma, Grnrchst, Z1720, Gog the Mild, Borsoka) clearly support doing so.

  I also think you're overlooking the other points of disagreement on your proposal:

  • 2 of the 4 editors other than yourself (myself and Grnchrst) who expressed a view about the lowest word count range (below-100) do not support that change. Also, while I initially supported the below-350 range, I oppose it now since both the below-350 and the below-100 word count ranges are percentiles in the "Measures of article size" table below the rules-of-thumb table; including them in both is redundant and especially if your not going to include a specific "What to do" instruction for the below-350 word count range and only statistics. Remember: the purpose of the rules-of-thumb table is to give the editor reading it recommended instructions of "What to do" with respect to splitting, merging, or trimming. I'd also note that your below-1,000 word count range does not include a recommended "What to do" instruction either and only statistics.
  • We still disagree about the FA and GA word count statistics.
  • Other than yourself, 2 of the 4 editors who have expressed an opinion about the average Wikipedia article word count statistics are skeptical about them (including myself and Hawkeye7). I'd add that the "Measures of article size" table below the rules-of-thumb table arguably makes these also redundant.
  • Also, support for the estimated reading times is only 5 of 8 editors who clearly support it (again, a bare majority).

  More importantly, considering the level of drop-off in participation in the discussion, limited engagement with the proposal in general, and how long it is taking to hash out what are fairly minor changes in some cases, I just don't see us getting any further. -- CommonKnowledgeCreator (talk) 01:22, 27 February 2026 (UTC)

  By your count, other than yourself, 3 out of 4 editors who expressed an opinion about the lowest word count range and the average article statistics favor them.

  One way to make progress is to pick the least-controversial change and make it. This is precisely the scenario for which WP:BRD was developed: there is a long discussion that's not going anywhere, so make a change and then see if anyone disagrees enough to revert it.

  Another is to break down the questions into smaller bits, and ask one little question at a time. For example, ask only about the 150 vs 100 question, e.g., "Do you think we should recommend that all articles with 150 words or less be considered for merging up to a larger topic? For reference, 30% of Wikipedias articles have less than 150, and 20% have less than 100." WhatamIdoing (talk) 03:29, 27 February 2026 (UTC)

  By your count, other than yourself, 3 out of 4 editors who expressed an opinion about the lowest word count range and the average article statistics favor them. This is incorrect.

  • Myself and Grnchrst expressed support making the lowest word count range equal to the stub guideline's minimum threshold for stubs (below 250 words), so neither us clearly support lowering the lowest word count range to below 100 words.
  • Other than yourself, while opposing the two highest word count ranges of your proposal, Borsoka said that they otherwise support the word count ranges.
  • Moxy indicated that they support proposal in general, but did not specify what parts of the proposal they supported.
  • HJ Mitchell expressed support for the estimated reading times and average word count statistics and opposed increasing the "limits", but did not express support for lowering the word count ranges.
  • Kusma and Gog the Mild support lowering the "limits", but did not cite the lowest word count range of the current guideline or your proposal.
  • Noleander only cited the three lowest word count ranges of the proposal as a justification for changing the three upper word count ranges.
  • Z1720 has only expressed support for lower "limits", not lowering the lowest word count range.

  So, yes, of the 4 editors who clearly have expressed an opinion about the lowest word count range, only 2 expressly support your proposed change.

  However, you're missing the larger issues with the lower three word count ranges your proposal that should be considered, such as:

  • Not including instructions in the below-350 and below-1,000 word count ranges and only statistics;
  • The "Measures of article size" table below the rules-of-thumb table effectively makes the below-350 word count range and average Wikipedia article word count statistics redundant;
  • Unclear why the lowest word count range should be different than what the stub guideline says and especially since below 100 words is a percentile in the the "Measures of article size" table as well. -- CommonKnowledgeCreator (talk) 04:32, 27 February 2026 (UTC)

  Likewise, as I said at the top of the subsection, for the average Wikipedia article word count statistics, 3 editors (yourself, Moxy, HJ Mitchell) expressed support for them, while 2 editors (myself, Hawkeye7) expressed skepticism about them. While Noleander expressed support for including "Adding details & stats to each size-range", they did not specify which ones and so cannot be counted as supporting including average Wikipedia article word count statistics. -- CommonKnowledgeCreator (talk) 04:50, 27 February 2026 (UTC)

  One way to make progress is to pick the least-controversial change and make it. This is precisely the scenario for which WP:BRD was developed: there is a long discussion that's not going anywhere, so make a change and then see if anyone disagrees enough to revert it. By this logic, what should be removed is "98.5% of all Wikipedia articles, and 76% of all Wikipedia:Featured articles, are shorter than this" from the current below-6,000 word count range's "What to do" recommendation, and add "While 489 featured articles (FAs) and 799 good articles (GAs) are this long as of February 26, 2026, 93% of FAs and 98% of GAs are shorter than this" to the current above-9,000 word count range's "What to do" recommendation since you appear to be the only editor who strongly opposes including FA/GA word count statistics and the support for including average Wikipedia article word count lengths is 3 to 2 (per my previous comment). -- CommonKnowledgeCreator (talk) 07:33, 27 February 2026 (UTC)

  But considering that you will probably revert it, I see no reason to do so. -- CommonKnowledgeCreator (talk) 07:36, 27 February 2026 (UTC)

  I probably wouldn't revert it, actually. I might refine it to something shorter, though, e.g., "Only 7% of featured articles (FAs) and 2% of good articles (GAs) are this long as of 2026". WhatamIdoing (talk) 01:18, 2 March 2026 (UTC)

  Your wording would defeat the purpose of including the FA/GA word count statistics. The reason for noting that almost 1,300 FAs and GAs combined are longer than 9,000 words is to simultaneously emphasize that the community has made a large number of exceptions to the specific recommendation by an explicit decision-making process but that the recommendation is still usually followed. Noting both would better illustrate that it is only a recommendation and not a requirement. -- CommonKnowledgeCreator (talk) 08:24, 2 March 2026 (UTC)

  If you prefer raw numbers, then it could say "About 500 out of 7,000 FAs and 1,300 out of 43,500 GAs are longer than 9,000 words, as of 2026".

  If your goal is to normalize longer articles, then I think that giving FA stats only makes a better case for this, and that raising that level to 10K would not only meet my 'round number' goal, but also make it another 1K words before anyone needs to think about it. WhatamIdoing (talk) 16:24, 2 March 2026 (UTC)

  If you prefer... If your goal is... First, you need to stop imputing motives to me that I don't have because it is effectively mischaracterizing my comments, and is part of why this conversation keeps dragging on longer than necessary. More importantly, my aim is to try to address the concerns of the editors who have expressed a preference for longer articles. One of those concerns is that editors reading the guideline frequently interpret its instructions as requirements rather than recommendations. I would argue that would counsel for the wording that I already proposed that includes both FA and GA word count statistics rather than yours because it is unclear why only the FA word count statistics should be used to illustrate this since both FA and GA word count statistics do so and would better to provide the editor reading the guideline with a larger set of exceptions to illustrate this. Moreover, most of your opposition as far as I can tell is an "I just don't like it" rationale couched in policy-sounding, source-based language that you keep reiterating but is not persuading me because you're just talking past what I'm saying. -- CommonKnowledgeCreator (talk) 18:08, 2 March 2026 (UTC)

  Not unilaterally assigning motives is the whole purpose of saying "If...". If I knew what your motives were, I'd use a different word, such as "Since...". What I'm telling you is that If this is your goal, then here's a practical way to achieve your goal. And, implicitly, if it's not, then don't do that.

  My distaste for combining absolute numbers for one "side" with percentages for the "other side" is based on knowing How to Lie with Statistics, especially in the pharmaceutical industry. Only two people had this very serious adverse effect (but don't mention that there were only three people in that first-in-humans trial, so that's a 67% rate). This expensive new treatment halves your risk of developing deadly scaryitis! (Please nobody mention that your risk was only one in a million to start with.) See also Base rate fallacy and related pages. WhatamIdoing (talk) 06:27, 3 March 2026 (UTC)

  I've taken introductory statistics in college, and I am currently taking a calculus-based probability and statistics course. I don't need a lecture about the base rate fallacy or the law of large numbers, and I certainly don't need one from you. Given your reflexive opposition to including what would be a fairly minor change to the guideline, and given how frequently you have cited the Policy writing is hard essay in past discussions, you are engaged in ownership behavior with respect to P&G pages as far as I am concerned. Given how long you've been an editor, you must be one the editors who led to the longstanding and perfectly defensible criticisms of Wikipedia's self-governance processes and talk page discussions. -- CommonKnowledgeCreator (talk) 16:43, 3 March 2026 (UTC)

  Is what I'm saying making sense? -- CommonKnowledgeCreator (talk) 03:11, 27 February 2026 (UTC)

CommonKnowledgeCreator, of the last 25 comments, 24 have been by you. Perhaps give a bit of air for other editors to comment? Gog the Mild (talk) 14:04, 27 February 2026 (UTC)

Most of the edits are copyedits. Apologies for not designating them as minor edits. Besides, air is not required for people to comment here. If people have a comment, they need only type them up and click. -- CommonKnowledgeCreator (talk) 14:36, 27 February 2026 (UTC)

I saw mention that editors that are not actively typing in on the issue are not engaged. I will say that I am reading the comments daily. However, I feel that I gave my opinion once (higher limits), and anything more is just a re-iteration, or engaging in verbose and unneeded dialog. It would be my preference if each editor stated their opinion once, then backed off and see what others say. I don't see this as a debate, where the one with the largest wall of words wins, but as a poll. My opinion is that all articles can handle a larger word count. I have never cared about featured articles or good articles, so don't care about their word or character count.

I would like to see the chatter stop and a poll be instituted. I would not mind if there was a poll for each and every segment of the character/word count, but what I would like to see is for this debate to end and a decision to be made. Mburrell (talk) 01:30, 1 March 2026 (UTC)

I agree, I gave my opinion above, but it's become a very confusing discussion. It's probably time to wrap it up .... but instead of a single poll there should be three polls: 1) whether to merge or add rows; 2) the word count of each row; and 3) what kind statistics to add into each row, if any. If those three distinct decisions are commingled within a single poll, it will be very hard to reach a consensus, because those are three distinct decisions, and every editor may have different opinions on each of the three choices. Better yet, those three polls should be spaced out to deal with one issue at a time. Noleander (talk) 01:45, 1 March 2026 (UTC)

@Mburrell and Noleander: I was under the impression that Wikipedia is not a democracy and that polling is not a substitute for discussion. I agree that closing the discussion would be appropriate, but I would argue that a separate straw poll is unnecessary since the discussion has already established that there is no clear consensus in favor of any specific change. I saw mention that editors that are not actively typing in on the issue are not engaged. This is a mischaracterization of what's been said. The issue is that editors have not generally expressed opinions about the proposal posted by WhatamIdoing on 21 February. I would argue that the reason this discussion got confusing is because of how many different changes that specific proposal included. I suspect there would also be fewer comments if editors would stop mischaracterizing each other, and if the editors making proposed changes were not dismissive of the concerns expressed by other editors and made some concessions—which they have been generally unwilling to do. -- CommonKnowledgeCreator (talk) 22:14, 1 March 2026 (UTC)

Years ago, I got in a large scale discussion with another user about splitting an article. I was so into responding, creating dialogue. What I discovered was that I had my opinion, the other user had another opinion, and nothing one person or the other said changed the other's opinion or moved the discussion along. So I certainly have tried to create dialogue, and have generally found it to be unhelpful after the first exchange of good ideas. So when does discussion just become a delay for doing action. You are correct, it clearly states no voting, that we should be heading towards consensus. However, I have seen after a long discussion that an administrator count up all the opinions and sum it up, as if that was a consensus. Therefore, it does appear to be voting or polling.

It is stated that consensus is "A consensus decision takes into account all of the proper concerns raised. Ideally, it arrives with an absence of objections, but often, we must settle for as wide an agreement as can be reached. When there is no wide agreement, consensus-building involves adapting the proposal to bring in dissenters without losing those who accepted the initial proposal." Still seems like a vote or poll to me, prettied up in fancy words. And we have to, because a crowd of users will not agree on much if anything at all. Pretty it up, and is still a vote. All I was saying is that we have been discussing this for about thirteen weeks, or a quarter of a year, and no movement has occurred. I just wanted to see if there was any appetite to move this discussion into a resolution, or if the preference is to keep discussing this for another stretch of time. No issue for me. The longer this goes unresolved, the longer it goes without being broken or changed into a more prescriptive guidance.

I will state that my objection to smaller limits is for the following reasons: We say that this is guidance, a recommendation, not a requirement, but there will be editors that will take it as a requirement. Splitting an article is one of the few actions that BRD (Bold action, Revert, Discuss) is not simple to revert, so once an article is split, it is not easy to rewind the clock. I am one of those people that don't read an article from start to finish, that prefer to visit an article for the information I am looking for, so I don't see that the size of an article as a reading limitation, but more of a technical limitation. I mostly work in the music sections, so I deal with lists primarily, which is generally used for review of information rather than being read from start to finish. However, even the articles are not so complex as to max out the size, but I can see how certain bands that are important to the industry would have larger articles. Looking at The Rolling Stones, I see that they have 14827 words (using the page size tool), and that The Beatles have 14839 words, and I am fine with that, not seeing a need to turn the article into a redirect page to the various sub-articles. The Beatles post-breakup activities section has a link to another article, but it also stands alone as a healthy section that provides sufficient information for those who just want to read the information in the main Beatles article. What should be broken out of the article and moved to another sub-article? I don't know, and I fear the one editor who decides they know and use this guidance as a command to just butcher the main article in order to comply with a prescriptive size limit, or what they read as being prescriptive.

I do agree that articles can be split, most can certainly be cleaned up, especially to remove non-neutral point of view and marketing, but I would like to see a split article discussion be started in the talk page first and hashed out how it would be done, before some editor says that an article must be split, and they just took the bold choice to split it on that editor's criteria. I would like to see a lot more caution about splitting articles, and a lot of consensus on splitting, before it was done. That is why I earlier stated that I would prefer to see that articles were left alone, unless someone was looking for a project to tackle. And finally, I will say that I have stated my opinion again, given some details, and think that is sufficient, and I will observe, having cast my vote, and otherwise be silent, but still engaged. Mburrell (talk) 23:59, 1 March 2026 (UTC)

@Mburrell, when you talk about your "objection to smaller limits", do you mean the limit that says 15K words "Almost certainly should be divided or trimmed"? Because when I think about "smaller limits", I'm thinking about the other end of the table (e.g., the recommendation to merge up 30% of articles because they're short/less than 150 words). WhatamIdoing (talk) 01:41, 2 March 2026 (UTC)

Almost right at the very top of the limit section of this discussion, I listed my views on size limits. I had stated the following for the upper limit "> 23,000 words | An interested editor could divide or trim this article. Otherwise, leave it alone, it's fine." Most of this discussion is revolving around 15,000 words. That is what I meant by smaller limits. I have no use for the smaller end of the table, except for classifying articles that are about two lines long as stubs. Mburrell (talk) 02:13, 2 March 2026 (UTC)

Thanks. So the existing quasi-limit of 15K is the "smaller" one.

My interest is primarily in the bottom part of the table, because that's where the disputes and misconceptions are ("If you haven't written 500 words already, we should delete or merge away the article, because that's totally subpar"). WhatamIdoing (talk) 02:40, 2 March 2026 (UTC)

My interest is primarily in the bottom part of the table, because that's where the disputes and misconceptions are ("If you haven't written 500 words already, we should delete or merge away the article, because that's totally subpar"). First, this is also a mischaracterization of what's been said. Hardly any of the editors who have left comments in this discussion have expressed any clear views about the lower word count ranges. Second, I would argue that this sentiment is entirely backwards. Considering that 99% of featured articles and 83% of good articles are longer than 1,000 words, that would suggest that the community does implicitly view articles shorter than 1,000 words as not typically being up to the community's standards and best practices for articles since the FA criteria and GA criteria include "comprehensiveness" and "breadth of coverage" respectively.

Moreover, I would argue that the sentiment that articles can afford to be longer on average at the lower end is more defensible when considering that shorter articles probably get fewer pageviews on average than longer ones do because of how search engine algorithms work. While the PageRank algorithm is no longer the sole metric that Google Search uses, it remains an important one, while other search engines use similar metrics of relevance. To take the Douglas MacArthur article that's been cited frequently in this discussion, I have no doubt that it got 1.4 million pageviews over the course of a year because of how long it is. The fact that 81% of Wikipedia articles are shorter than 1,000 words probably contributes to Wikipedia losing readers as measured in pageviews.

Also, as reported by TechCrunch last October, the WMF's senior product manager has noted that Wikipedia is losing web traffic due to generative artificial intelligence and its contributing to Wikipedia articles losing pageviews and probably contributing to the declining numbers of editors. One institutional change we could make to stem the decline in pageviews could be to recommend that articles shorter than 1,000 words be merged into articles about a related broader topic so that more articles are likely to be returned higher in a search engine query (since there probably isn't much we can do policy-wise to directly address the declines caused by generative artificial intelligence). I would argue that this would be more in the best interest of the encyclopedia in the long-term rather than having related content scattered across the 81% of articles shorter than 1,000 words. -- CommonKnowledgeCreator (talk) 17:06, 2 March 2026 (UTC)

• Yes: Hardly any of the editors who have left comments in this discussion, emphasis added. But I could name multiple editors off the top of my head who would cheerfully affirm their distaste for short articles and who have the RFCs and AFDs to prove it. Anyone who's paid attention to Wikipedia:Articles for creation output will realize that they often insist on an above-median word count for an article to be moved to the mainspace. This problem exists, even if it's not being discussed here.
• Mind the gap between being up to the community's standards and the best articles Wikipedia has to offer. Ordinary, average non-FA articles aren't below standard. They just aren't "the best".
• I doubt that having lots of un-read short pages has any effect on Wikipedia's total pageviews. A longer article will naturally have more opportunities for matching a search string, but that doesn't mean that search engines favor long pages. (And if they did, we'd expect SEO farms to have very long pages, because text is cheap. But we don't, so it probably doesn't.)
• I don't think that merging up articles is likely to have SEO benefits. In fact, since search engines care about relevance, if someone's searching for Entomocorus benjamini, and all we offer is a general article on Catfish, in which E. benjamini gets a sentence or two, then that would probably make the page rankings worse, because (a) it would be less closely related to the search string and (b) there are a lot more websites out there that talk about catfish, but not so many that talk about Entomocorus benjamini. PageSize says that article has 102 words, by the way. Google lists it as the #1 hit if you search for its name.

WhatamIdoing (talk) 06:42, 3 March 2026 (UTC)

To take the examples, you've cited, Catfish (4,361 words) and Entomocorus benjamini (102 words), the Pageview statistics tool clearly shows that from March 2025 through February 2026 the Catfish article ranged from almost 40,000 to 55,000 pageviews per month while the Entomocorus benjamini ranged from 41 to 134. At least for state leaders, there does appear to be a correlation between article size and pageviews. Additionally, a 2008 academic journal article researching Wikipedia article word counts does show that article word count was a superior heuristic/metric of article quality than alternative and more complex methodologies for identifying higher-quality content. So, no, length of an article does probably correlate with longer page views and probably with relevance. Also, considering that you are talking past my previous comment here as well, I am only suggesting that articles about related topics be merged, not merging indiscriminately since the former may very well increase search engine relevance. -- CommonKnowledgeCreator (talk) 16:44, 3 March 2026 (UTC)

How would you control those page view numbers for the fact that people are more likely to search for catfish than for E. benjamini? If a million people search for catfish each day, but only 0.1% come to the Wikipedia article, but only five search for E. benjamini, and 100% of them come to the Wikipedia article, that doesn't indicate that the E. benjamini article is failing to attract attention from the people who want to read specifically about that species.

How would you control for the fact that things that are interesting to the general public attract more Wikipedia editors, and therefore become longer articles? We've never run an experiment in which we create 100 similar articles, half randomized to be 3,500 words long and the other half to be 350 words long (i.e., the median length of FAs vs the median length of all Wikipedia articles) and then waited to see whether the page views were different.

(I think that the work done on mw:ORES#Article quality is more robust than that "FAs are longer than average, so long is correlated with high quality" paper. https://dl.acm.org/doi/abs/10.1145/2491055.2491063 might be a jumping-off point if you're interested.) WhatamIdoing (talk) 02:50, 4 March 2026 (UTC)

@Mburrell: Thank you for your lengthy and thoughtful comment. I think your concerns about lengthy articles being split without due consideration is a defensible one, which I think it could be served by changing the wording of the "What to do" recommendations for the above-15,000 word count range. Since you said you have no interest in FAs and GAs, maybe it should be changed to say something like, "Size alone can justify division or trimming, but splitting articles even this long should still be discussed first"? Also, to further illustrate that discussion is not totally a waste of time, I think the arguments you, EEng, and Hawkeye7 have made with respect to the typical reader not reading articles from beginning to end justifies not including the estimated reading times in the "What to do" recommendations as proposed by WhatamIdoing's proposal. After greater thought, I think the length of page views suggested by the 2011 Nielsen Norman study and the 2019 WMF study at best justifies the current guidance the Manual of Style suggests for the length of lead sections and agree that it should have no bearing on article size (although I remain not persuaded that the upper word ranges should be increased). -- CommonKnowledgeCreator (talk) 03:54, 2 March 2026 (UTC)

Thank you for your thoughtful response. I am not wed to the need for an increase to the upper word range, I just wanted to discourage butchering articles by arbitrary splitting. I do like your suggestion of "Size alone can justify division or trimming, but splitting articles even this long should still be discussed first" You have my support here. Mburrell (talk) 05:46, 3 March 2026 (UTC)

Arbor-treeish break

Could WhatamIdoing please make a single table showing the old and new word counts, and old and new text? EEng 18:02, 24 February 2026 (UTC)

No offense taken by deletion of my table. I thought it sucked anyway. :) -- CommonKnowledgeCreator (talk) 04:04, 25 February 2026 (UTC)

User:EEng, here's a side-by-side comparison. The rows are somewhat arbitrary.

WhatamIdoing (talk) 23:52, 24 February 2026 (UTC)

• If the two upper limits were reduced substantially (from 15,000 to 12,000, and from 10,000 to 8,000), I would be happy to support the proposal. According to the estimate presented above, an average reader would need more than 40 minutes to read an article exceeding 8,000 words. In practice, this renders such articles inaccessible to most of our readership. Are editors expected to write primarily for themselves and a small circle of fellow enthusiasts, or for a broader readership? Borsoka (talk) 01:57, 25 February 2026 (UTC)

  The claim that "The longer the article, the fewer people who will read the whole thing" is untrue and unfounded. Douglas MacArthur, one of our larger articles, gets 1.4 million page views per annum. Hawkeye7 (discuss) 02:37, 25 February 2026 (UTC)

  The assumption that page views correspond to readership is absolutely unsound, as many articles are opened without being read beyond their opening lines. Borsoka (talk) 04:20, 25 February 2026 (UTC)

  Well, what does that tell you? That's all most readers want. Duh. EEng 04:35, 25 February 2026 (UTC)

  If that is the case, one might conclude that nearly 99% of Wikipedia's content represents wasted effort, since most readers do not read it. I would instead aim to write articles that are wholly useful to a wider readership, making more effective use of time and energy. Borsoka (talk) 04:56, 25 February 2026 (UTC)

  No, one couldn't conclude that, because 50% of Wikipedia's articles are about one screenful long. WhatamIdoing (talk) 07:15, 25 February 2026 (UTC)

  The Douglas MacArthur article may get more than 1.4 million page views per year, but how long are those page views? Ladtrack referenced a 2019 WMF study in the first part of this discussion that found that the average page view was 25 seconds; the Nielsen Norman Group did a study back in 2011 that found that average web page visit across the internet was less than one minute. The 75th percentile for the 2019 WMF study was 75 seconds. So, if I had to guess, people reading the MacArthur article spend less than 5 minutes reading it, which would be far shorter than necessary to read the entire article. -- CommonKnowledgeCreator (talk) 04:21, 25 February 2026 (UTC)

  So apparently 5 minutes' worth is all they want. What's wrong with that? And your 25 seconds and one minute and 75 seconds -- these are for "the average web page across the internet", so what do they have to do with our articles, which are quite different from 99.9999% of other pages on the internet? It's like saying that the average feature film is 100 minutes, so operas should only be that long too. One has nothing to do with the other.

  Or, let's say that 75 seconds has some application here. What -- our articles should be no longer than can be read in 75 seconds? If that's not what you're saying, then what does 75 seconds have to do with anything? (You're not along in throwing around irrelevant numbers -- there's something about this talk page that makes people do it.) EEng 04:35, 25 February 2026 (UTC) P.S. Thanks for posting the comparison table.

  And your 25 seconds and one minute and 75 seconds -- these are for "the average web page across the internet", so what do they have to do with our articles, which are quite different from 99.9999% of other pages on the internet? The 25-second figure and the 75-second figure are for Wikipedia articles. Only the 1-minute figure is for the web pages in general.

  So apparently 5 minutes' worth is all they want. What's wrong with that? Nothing. I'm sure most readers don't read past the lead sections since they're closer to what most encyclopedia entries were like historically. But it is unclear to me why that is a justification for increasing the article size guideline's word count ranges.

  Or, let's say that 75 seconds has some application here. What -- our articles should be no longer than can be read in 75 seconds? If that's not what you're saying, then what does 75 seconds have to do with anything? It has to do with the fact that downloading Wikipedia articles still requires data and bandwidth. The longer the article, the more hyperlinks and citations it includes, the more data is required to download it. Maybe not as much data as when video and images are included, but text and hyperlinks still use data. But more importantly, it goes more to the first of the five pillars: once they exceed a certain length, the articles can no longer be reasonably considered analogous to an encyclopedia entry in any meaningful sense. -- CommonKnowledgeCreator (talk) 05:03, 25 February 2026 (UTC)

  I apologize for misreading your post -- my eye fell on the mention of Nielsen and tuned out that the other numbers were from WMF. (Fact is, it was well past my bedtime BUT GODDAM IT I HAD TO KEEP UP THE ARTICLE-SIZE FIGHT!) But with the corrected understanding that the 25-second and 75-second figures are the median and 75th-%ile English Wikipedia "dwell times" (time the article is visible in a window, before the window is closed or the user clicks away to something else), then the idea that anyone reads any but the very shortest articles from top to bottom is now not only dead and buried, but dead, encased in a titanium casket surrounded by 500 tons of radioactive concrete studded by shards of cyanide-tipped broken glass, then rocketed into the sun by an international consortium of NATO and Warsaw Pact counties coordinated by the UN. I'll write more about this later. EEng 21:01, 27 February 2026 (UTC)

  Actually, Hawkeye7, I must disagree with you. Of course longer articles are less likely to be read from beginning to end. I wouldn't be surprised, for example, if 2% of readers will read a 5000-word article from beginning to end, while only 0.5% of readers will read a 10000-word article from beginning to end. But who gives a shit? It's a rarity that anyone will read an article straight through, beginning to end, because that's not what almost any reader sets out to do. What readers do is read (or more likely skim) the lead, then glance at the TOC. After that they might skip to a section they find in the TOC, and read it, and then maybe keep reading a section or two more for a while, then if they're still interested enough they go back to the lead and read half of it more carefully, then start reading the first section, then back to the TOC, or just skim here and there, then the phone rings and they have to get that email they've been dreading writing to their thesis advisor written, then maybe the next day they find the open tab and realize that the fourth section might have something interesting for whatever they're working on, then ....

  The right question isn't: What proportion of readers will read beginning to end? The right question isn't even: What proportion of readers will find what they came to the article looking for? (The reason that's not the right question is that readers often come to a page looking for one thing, but come away having found something they weren't even looking for, but which they're happy to have found.) The right question is: What proportion of readers will come away glad they visited the page? That's a complicated question, with a lot about it that can be debated, but I know one thing for sure: it isn't answered by reducing it to "What proportion of readers will read the article from beginning to end?" I'm sorry, but that a ridiculous leap of logic with no basis whatsoever.

  And -- please inform me, Borsoka. How does an average reader would need more than 40 minutes to read an article exceeding 8,000 words translate into In practice, this renders such articles inaccessible to most of our readership? Whaaaat? Huh? Howzattt? That makes no sense. EEng 04:35, 25 February 2026 (UTC)

  I realise this observation may again be met with some theatrical astonishment, but readers who consult an encyclopaedia do not wish to spend over half an hour extracting the essential information on a given subject; excessively long articles are therefore, in practice, inaccessible to them. Those seeking more detailed knowledge will turn to a specialised study or a book. A text that is too detailed to function as a concise encyclopaedia entry, yet insufficiently rigorous to qualify as an academic study, ultimately serves neither purpose and is therefore of little practical value — a waste of its author's time. Borsoka (talk) 04:51, 25 February 2026 (UTC)

  I am indeed astonished that anyone can have such a cramped and dwarfed view of what we're all doing here.

  • The lead gives the base-level "essential information". Someone wanting more "essential information" will skim the article to find it, not take your shock-value "half hour" to plod through the whole thing word for word (which is what you'd do to get all the article's information, not the essentials).
  • Encyclopedic means comprehensive and exhaustive, but you seem to want it to mean something else.
  • You also don't seem to know what "concise" means. It doesn't mean "short" or "abridged".
  • Of course an "excessively long" article is, well, excessively long, and an article which is "too detailed" is too detailed; your wording counterproductively begs the question.
  • too detailed to function as a concise encyclopaedia entry, yet insufficiently rigorous to qualify as an academic study – Gosh, if only we had a table of rigid numbers to tell us when that's the situation!

  The right length, and right level of detail, for a given topic is often a hard question, and one that must be answered on a case-by-case basis for each subject. It sure as hell isn't answered by arbitrary numbers. EEng 05:36, 25 February 2026 (UTC)

  EEng, the bigger issue is that the majority of editors who have left comments here are clearly are not in favor of increasing the word count ranges, so there is no consensus in favor of doing so. To bring this discussion to closure, what needs to be discussed is Proposal #2 (WhatamIdoing's proposal) to resolve the points of disagreement about it. -- CommonKnowledgeCreator (talk) 12:27, 25 February 2026 (UTC)

  @EEng: Like In complete agreement. Hawkeye7 (discuss) 19:59, 25 February 2026 (UTC)

  However, to be clear, I am not arguing that all articles should be required to be shorter than an arbitrarily specified length (and have consistently said throughout this discussion that the lengths included should be seen as recommendations rather than requirements). I do believe that broad topics (such as World War II) would be better served by a longer article. I just think your concerns would be better served by including featured article and good article word count statistics in the guideline rather than increasing the word count ranges (which is why I initially proposed doing so on 20 February before this discussion exploded). -- CommonKnowledgeCreator (talk) 20:31, 25 February 2026 (UTC)

We are in agreement that appropriate length must be decided case by case, which is precisely why a single universal limit cannot be proposed. You may not have noticed, but I support a six-level system, which can hardly be characterised as "rigid". Your remaining remarks serve mainly as an example of time unnecessarily wasted. Borsoka (talk) 06:07, 25 February 2026 (UTC)

I dunno, my impression is that the disagreement is about whether the "total length of article" matters and how much. Because if as EEng posits most people read the lead and the TOC and only continue if interested, this would falsify most arguments for keeping the current guidelines presented above since that would mean that total article length is of little relevance to readers. Jo-Jo Eumerus (talk) 12:35, 25 February 2026 (UTC)

(disclaimer: I have a vested interest in larger articles, having authored a few and don't want to do the extra work during updating that comes from a topic being split across multiple articles) Jo-Jo Eumerus (talk) 12:36, 25 February 2026 (UTC)

Because if as EEng posits most people read the lead and the TOC and only continue if interested, this would falsify most arguments for keeping the current guidelines presented above since that would mean that total article length is of little relevance to readers. That would matter more if Wikipedia didn't make decisions by consensus. With your and Gog the Mild's comments, the 11 of the 19 editors who have left comments in the discussion still oppose raising the word count ranges—meaning no consensus in favor of it. -- CommonKnowledgeCreator (talk) 12:51, 25 February 2026 (UTC)

Regarding the numbers: [nerd alert] if new thresholds are introduced at the low end: 1,000, 350, 100 that suggests a logarithmic spacing (not constant spacing). Thus, intuitively, the upper spacing should be something like: 15,000, 8,000, 4,000, ... (not 15,000, 10,000, 5,000). Noleander (talk) 18:21, 25 February 2026 (UTC)

While logarithmic spacing could be built into the size guideline, it's unclear to me that it would remain a rule of thumb if we did so. The word count ranges should be based on the practical experience of editors, not a particular mathematical distribution. It is why I favor including the estimated reading times, featured article and good article word count statistics, the word count threshold for stubs (below-250) as the lowest word count range, and the below-1,000 word count range (since 99% of FAs and 83% of GAs are longer than 1,000 words), but in retrospect think that the proposed below-350 word count range (as well as including the average Wikipedia article lengths in the "What to do" recommendations) is unnecessary since there is a separate "Measures of article size" table in the guideline already. -- CommonKnowledgeCreator (talk) 19:31, 25 February 2026 (UTC)

I wouldn't be surprised if it was a Half-logistic distribution, and there is value in telling editors that n is the typical word count, so if your word count is >>>n, it probably needs to be cut. WhatamIdoing (talk) 00:27, 26 February 2026 (UTC)

I'll admit to not reading the above in its entirety. I don't see a strong reason to change what we currently have, but I am very unsure about the bottom part of the table (< 150 words – If an article or list has remained this size for over two months, consider merging it with a related article. Alternatively, the article could be expanded; see Wikipedia:Stub. I oppose raising the limits for the top part of the table (though I could see changing the current > 9,000 words – Probably should be divided or trimmed, though the scope of a topic can sometimes justify the added reading material. to > 10,000 words instead). In my experience, 8,000 words is about the right length to go for when there is extensive literature on a topic, striking the right balance between breadth and depth. Part of the issue here is simply learning how to better write Wikipedia-style articles, which is a trainable skill. It remains the case that these are optional recommendations rather than mandatory requirements. I have no strong opinions on adding statistics. TompaDompa (talk) 20:53, 2 March 2026 (UTC)