Template talk:Registered editors by edit count

Question

I'm confused by the first line on this table: "If you have made 1 edit, then you rank in the top 50% of all editors". Does that mean 50% of "editors" have made 0 edits? Or is there a huge tie that 50% of editors have only made one edit? Thanks! GoingBatty (talk) 19:11, 16 November 2021 (UTC)

Yes, the table is, um, suprising at least. I don't see anything about never-editing registered users at Special:Statistics, but the implication is that more than 21 million people (half of 42,562,219) signed up for an account and then never edited. Now, that's a sleeper cell!. Of course, there could be only 10 million who did that, but they each made two accounts with which they never edited. IAC, 21 million never-used accounts is a lot of forgotten passwords! Or maybe our interface is soooo bad, it scared would-be editors away alone through the sign-up process.

Now if you'll excuse me, I have to go check my edit count to see in what level of the elite I am. — JohnFromPinckney (talk / edits) 01:20, 17 November 2021 (UTC)

Of the 42.5 million registered accounts, the majority never successfully published a single edit. Some of these are indeed software problems (or software "non-problems", like anyone who wanted to post a link that's on the Wikipedia:Spam blacklist). Others are people who created an account but never intended to edit (e.g., they just wanted to use the watchlist, or they really edit at another Wikipedia, and the Wikipedia:Unified login system auto-created an account when they read a page here).

"Half" is actually me preferring round numbers; only 28.68% of registered users at the English Wikipedia have ever successfully published a single edit. I could have equally written that anyone who has published a single edit is in the top 30%. WhatamIdoing (talk) 03:38, 18 November 2021 (UTC)

If you'd like to expand the table:

• 18.3% have made two or more edits.
• 13.4% have made three or more edits.
• 10.7% have made four or more edits.
• 8.9% have made five or more edits.

If you've made just five edits, then you're in the top 10%. Even if you prefer to exclude all of the zero-edit accounts, then a mere five edits puts you in the top 30% of all editors for all time. WhatamIdoing (talk) 03:48, 18 November 2021 (UTC)

Also, @JohnFromPinckney, I know you meant it as a joke, but you are #3450 at Wikipedia:List of Wikipedians by number of edits, which means you have made more edits than 99.92% of all registered editors, ever. And yet there are a few editors who might treat you like you have "only" made almost 30K edits. WhatamIdoing (talk) 03:55, 18 November 2021 (UTC)

Can we not say "more than X% of accounts" or similar? You can't be an editor without having done an edit! Best Wishes, Lee Vilenski ^{(talk • contribs)} 12:10, 4 January 2022 (UTC)

I'm not sure that's true. How would you classify someone who attempted to edit, but gave up because it was too confusing? WhatamIdoing (talk) 16:39, 4 January 2022 (UTC)

Hm, well, someone with 0 edits is not an editor as they've never edited. You could call them "registered as an editor" or something, but that isn't the same thing. So the 1 edit thing is indeed (extremely) confusingly worded, and imo also clearly wrongly worded.

It might be best to rejig the table very slightly to show "more than" 90%, 99%, 99.9%, 99.99%, etc, as the other entries don't add much, and that first row in particular just starts hares running without adding anything useful. Just my tuppence'orth. All the best, Chiswick Chap (talk) 18:43, 23 February 2022 (UTC)

The table has about twice as many rows as it did when I started. I assume that whoever added the other rows thought that it would add some value. WhatamIdoing (talk) 21:49, 23 February 2022 (UTC)

The column "That's more than..." is unnecessary.

Hi! Sorry if you already discussed it (I couldn't find any archive here): Isn't the "That's more than..." column in this template a bit unnecessary (just 100% - x)? I mean, it's funny and you can see that's the intention with the "!", but I feel that the column "then you rank in the..." is more than enough. Or what do you think? --VictorBenitoGarciaRocha (talk) 06:00, 27 December 2021 (UTC)

@VictorBenitoGarciaRocha, you're right: It's just simple math. But it's also my experience that many people can't do simple math. If you tell someone "half the people with this disease die", they have a very different emotional reaction than if you tell them "half the people with this disease survive". Some people have an easier time grasping one form vs the other. WhatamIdoing (talk) 16:38, 4 January 2022 (UTC)

@WhatamIdoing, I can do simple math and when I read "10,000 edits: top 0.025% of editors; top 10,920 of all editors; 99.975% of editors" I was fine (just mildly surprised) until I reached the last figure. 99.975% of editors have done what?.Perhaps I hadn't had enough coffee but the effort of studying the column headings and patching them and cells together into sentences was clearly quite beyond me on first reading. Looking at it now, either a plain "10,000: 0.025%: 10,920; 99.975%" or a consistently fulsome "10,000 edits: top 0.025% of all editors; top 10,920 of all editors; more edits than 99.975% of all editors" would have worked better for me. NebY (talk) 06:39, 7 June 2022 (UTC)

Why is 10,000 formatted differently?

The entry for 10,000 edits says "top 10000 of all editors". All other entries in the table are in percentages (ie: top 0.25% etc), why is this one not in a percentage? RudolfRed (talk) 02:47, 5 January 2022 (UTC)

@RudolfRed, it appears that 10,000 registered accounts have made 10,000 edits, which is a fun coincidence that won't be noticed in the percentage approach. This was noted in the edit summary. WhatamIdoing (talk) 21:52, 23 February 2022 (UTC)

@WhatamIdoing: Thanks for the reply and the explanation. RudolfRed (talk) 22:02, 23 February 2022 (UTC)

Align left

I suggest adding the option to align left. Thinker78 (talk) 04:19, 22 August 2022 (UTC)

If you know how to do it, I've no objections. WhatamIdoing (talk) 02:23, 4 April 2023 (UTC)

Misleading

The figures in this template are wrong, to varying degrees of wrong.

The first wrongness is associating "users" with "registered editors". A "user" is someone with an entry in the USER table in the database. They have an account here. But they aren't an "editor" till they have made one edit. There are various reasons why people might have accounts here (some mentioned above, such as users on other wikis who just happen to read an en-wp page, or who create an account to get a watchlist, etc). So I think if we are going to measure editors off against each other to say "you are in the top x of all editors", we should start with actual editors who make edits, not accounts.

But going with the accounts = editor wrongness for now, the second wrongness is that the numbers in the third column in the template are derived from the percentages in the second column, multiplied by the current number of user accounts. But the percentages in the second column are out of date and appear to have always been rather approximate. One of them (the figure for 1 edit) is so wildly wrong I don't know how it got there. This approximate or flawed percentage is then multiplied by an accurate (if misleading) number of user accounts, to give what appears to be an accurate number of users in that category. It isn't.

The "actual" columns in table below is created by running a query on the user table. The "current" columns are what we currently show in the template.

User accounts by edit count

Edits	Actual #	Current #	Actual %	Current %
0 edits	32,372,751		71.4%
1 edit	12,951,300	22,660,000	28.6%	50%
10 edits	2,293,272		5.06%
10 edit & 4 days (autoconfirmed)	2,282,214	2,266,000	5.04%	5%
100 edits	321,314	453,000	0.709%	1%
500 edits	94,591		0.209%
500 edits & 30 days = extended autoconfirmed	94,207	113,000	0.208%	0.25%
1,000 edits	58,418	45,000	0.129%	0.1%
10,00 edits	11,146	11,300	0.0246%	0.025%
25,000 edits	4,466*	4,500	0.0099%	0.01%
45,000 edits	2,301*	2,200	0.0058%	0.005%
90,000 edits	924*	900	0.0020%	0.002%
140,000 edits	458*	450	0.0010%	0.001%
200,000 edits	260*	226	0.00057%	0.0005%
350,000 edits	98*	90	0.00022%	0.0002%
500,000 edits	50*	45	0.00011%	0.0001%

In second column of the above table, numbers with a * are taken from Wikipedia:List of Wikipedians by number of edits which excludes bot accounts; otherwise from a query of the USER table, which will include bots.

If instead we base our values on users who have made at least one edit, we get the following table

Editors by edit count

Edits made	Number of editors	% of all editors
1	12,951,300	100%
2	8,286,137	64%
3	6,080,075	47%
4	4,841,383	37%
5	4,035,686	31%
6	3,477,597	27%
7	3,064,243	24%
8	2,748,341	21%
9	2,497,069	19%
10	2,293,272	18%
11	2,125,429	16%
20	1,301,053	10%
30	922,805	7.1%
40	719,537	5.6%
45	649,204	5.0%
50	590,949	4.6%
60	501,686	3.9%
70	437,079	3.4%
80	388,898	3.0%
90	351,584	2.7%
100	321,314	2.5%
200	183,247	1.4%
300	135,095	1.0%
400	110,129	0.85%
500	94,591	0.73%
600	82,911	0.64%
700	74,438	0.57%
800	67,969	0.52%
850	65,258	0.50%
900	62,751	0.48%
1,000	58,418	0.45%
2,000	36,320	0.28%
3,000	27,432	0.21%
4,000	22,332	0.17%
5,000	18,938	0.15%
6,000	16,562	0.13%
7,000	14,733	0.11%
8,000	13,287	0.10%
9,000	12,088	0.093%
10,000	11,146	0.086%
18,500	6,525	0.050%
20,000	5,609*	0.043%
30,000	3,649*	0.028%
40,000	2,633*	0.020%
50,000	2,038*	0.016%
60,000	1,614*	0.012%
70,000	1,323*	0.010%
80,000	1,101*	0.0085%
90,000	924*	0.0071%
100,000	791*	0.0061%
112,000	648*	0.0050%
200,000	260*	0.0020%
300,000	127*	0.00098%
400,000	83*	0.00061%
500,000	50*	0.00039%

In second column of the above table, numbers with a * are taken from Wikipedia:List of Wikipedians by number of edits which excludes bot accounts; otherwise from a query of the USER table, which will include bots. -- Colin°^Talk 11:07, 7 April 2023 (UTC)

The better-to-understand first table needs an update (i.e., there are now 260 users with 200,000 edits, not 226), and I really can't figure out what your second table is showing, and shouldn't become the "official" table due to its inclusion of bot accounts. But yes, the table would be nearer editorial accuracy by removing the "editors" who have made no edits. Randy Kryn (talk) 11:16, 7 April 2023 (UTC)

I've retitled the second table. Don't see why the first is "better to understand". What's the point in easily understanding something that is just wrong and misleading. This is like a table ranking scores in O Level Geography that includes pupils who never even sat O Level Geography and then wrongly claiming you are in the top 5% of Geography students.

This is important as we get people claiming if you have made 1,000 edits then you are 1 in a 1,000 (or in the top 0.1% as the table says). But you aren't. You are in the top 0.45% and 1 in 222. To get to be in the top 1,000, you have to make 8,000 edits. And if you've made 5 edits, you aren't in the top 5% of editors but in the top 31%. There's a big difference between being in the top 20th and the top third.

As for bot accounts, what makes you think the current table excludes them? The commentary on the page only says to examine the Wikipedia:List of Wikipedians by number of edits for the "highest activity levels", and otherwise says it got its data originally from the same queries on the USER table as I did. Where do you draw the line on editors using some degree of automation for their activity here, vs editors who manually make every edit? The current table, by including "editors" who aren't even editors at all, makes a far worse statistical error. -- Colin°^Talk 13:09, 7 April 2023 (UTC)

I've updated the bottom portion of both tables with data from Wikipedia:List of Wikipedians by number of edits and marked affected numbers with a *. You can see it doesn't change the percentage figure much till you reach the bottom. -- Colin°^Talk 14:08, 7 April 2023 (UTC)

Good work, you've analyzed these well. Separating automatically run bot edits from human editors should be done (bots can often do, for example, 6,000 edits a day, and shouldn't be credited to the edits of the user who created them). Can tell that the chart lists bot edits by comparing the 260 editors at 200,000 to the 385 mentioned in the chart. Randy Kryn (talk) 13:55, 7 April 2023 (UTC)

I've made a further edit to simplify both tables and use the non-bot data for all rows with 20,000 edits or more. I see now the original table wasn't so far out for those figures, but was much further out-of-date/wrong for the lower activity editors. I think the second table is the approach we should use for this template, which is all about "editors" not account holders or user records. I included lots of rows just because I had the data, but in practice, we'd probably want to abbreviate the table a bit. -- Colin°^Talk 14:15, 7 April 2023 (UTC)

In ratios

• 1:10 editors make 20 edits
• 1:20 editors make 45 edits
• 1:100 editors make 300 edits
• 1:200 editors make 850 edits
• 1:1,000 editors make 8,000 edits
• 1:2,000 editors make 18,500 edits
• 1:1,0000 editors make 70,000 edits
• 1:2,0000 editors make 112,000 edits

-- Colin°^Talk 14:27, 7 April 2023 (UTC)

Shouldn't the last two lines be "1:10,000 editors make 70,000 edits" & "1:20,000 editors make 112,000 edits"? -- llywrch (talk) 21:47, 18 April 2023 (UTC)

Whether someone who is logged in but not (yet) successfully published an edit (on this wiki) is "an editor" depends on how you define editor.

I have previously wondered about displaying both (the {{NUMBEROFUSERS}} version and a "successfully completed" version), and Wikipedia:Village pump (technical)#Fancy formatting describes some ways to do that without making the template display completely unwieldy. WhatamIdoing (talk) 03:32, 11 April 2023 (UTC)

BTW, there's a note in the /doc page saying that the table would probably need to be updated this year. I'd be happy to have someone else do that. In terms of keeping it reasonably true to its purpose (which is to remind editors like us, with our tens and hundreds of thousands of edits that someone with "only" a hundred edits or so is part of the unusual elite, rather than part of the disposable, unwashed masses), whoever wants to do that might find it useful to look at the original version, which was shorter and simpler. WhatamIdoing (talk) 03:39, 11 April 2023 (UTC)

The definition of editor

Following from the above discussion, let's see if we can agree how to define "editors" for the purpose of this template information (which is transcluded at Wikipedia:List of Wikipedians by number of edits). For example, if we say your edit count is in the top 1% "of all editors" what does that mean? There are two meanings we could use:

1. An editor is someone who has got an account on English Wikipedia. (They have a record in the USER table in the database).
2. An editors is someone who has made at least one edit. They have edited Wikipedia.

Definition 1 is what the template is currently using. The problem is most of such account holders never make any edits at all. It seems editors on other Wikis get an account here if they look at a page. Or someone might have created an account in order to get a watchlist or to do some research. Or created an account and tried to post something that the edit filter blocked and gave up. Or created an account as a student in a class doing Wikipedia, but then couldn't be bothered to do any homework. Or created an account and then forgot their password. Etc, etc..

I think that's going to confuse a lot of people who will think that, well, you need to edit to be an editor. It turns out only 28.6% of accounts here have ever made even one edit (the template claims that's 50% but it is wrong). So it really alters the stats if you include these no-edit accounts. I think it would be more sensible if we moved over to definition 2.

(Pedants will note that it is possible for someone to have more than one account. It isn't really feasible to do any stats at that level, because the is no common way that people declare this, assuming they do declare it.) -- Colin°^Talk 16:00, 11 April 2023 (UTC)

By definition and WP:COMMONSENSE an editor has to have made an edit. Seems an easy call. Randy Kryn (talk) 22:26, 11 April 2023 (UTC)

I'd say that by definition and common sense, an editor is anyone with the Edit user right. WhatamIdoing (talk) 23:35, 11 April 2023 (UTC)

Here is the Wikipedia definition of Editor:

"Editing is the process of selecting and preparing written, photographic, visual, audible, or cinematic material used by a person or an entity to convey a message or information. The editing process can involve correction, condensation, organisation, and many other modifications performed with an intention of producing a correct, consistent, accurate and complete piece of work."

Then here is WP:EDITOR: "Wikipedians are volunteers who contribute to Wikipedia by editing its pages, unlike readers who simply read the articles."

Nothing in either definition fits someone who has not edited a particular publication or site. Randy Kryn (talk) 23:43, 11 April 2023 (UTC)

I agree totally with Colin and Randy here (!). JoelleJay (talk) 21:24, 16 April 2023 (UTC)

It turns out only 28.6% of accounts here have ever made even one edit (the template claims that's 50% but it is wrong).

Not really? The template claims that if you have made one edit, you are in the top 50%, or the top 22,670,000 of all 45M editors, and that this is more than 50% of all editors.

It would be more precise to say that if you have made one edit, you are in the top 30%, or the top 15M of all editors, and that this is more than 70% of all editors, but I believe that you will agree with me that all people in the top 30% of any population are also in the top 50% of that same population. WhatamIdoing (talk) 23:38, 11 April 2023 (UTC)

WhatamIdoing, regardless of one's definition of editor, I have not been able to find any data to support either the 50% figure or the 22,670,000 figure (which seems to be just 45M/2). Let me say it fails verification! The current template is driven entirely by a set of percentages that were worked out to be approximately correct years ago, but maybe too approximate today. The exception to that is the one-edit row, which seems to be simply a mistake. A copy/paste error or something. -- Colin°^Talk 07:17, 12 April 2023 (UTC)

Colin, I think you're better at math than this. Let's say that you rank at the 70th percentile (0 is the bottom, 100 is the top). Are you:

(a) in the top 50%, or

(b) not in the top 50%? WhatamIdoing (talk) 02:07, 13 April 2023 (UTC)

Are you suggesting 12,951,300 out of 45,366,750 is "top 50% of editors" because 50 > 30 and so "top 30% of all editors" => "top 50% of all editors" as well. Or am I being really dumb and the maths really does come to 50% and no less. Because what I think you are arguing is that "Elon Musk is in the top 50% of people, in terms of wealth" is an entirely reasonable claim to make because maths. A statement can be "wrong" for more reasons than just maths. It clan be misleading. -- Colin°^Talk 10:29, 13 April 2023 (UTC)

I agree that it is imprecise, but I'm telling you that it's accurate. Elon Musk actually is in the top 50% of people in terms of wealth. That statement (about Musk) is true, even if it is a potentially pointless way to describe his specific situation (unless, I suppose, you're trying to explain the concept of a range to some students: "The top 50% includes a lot of school teachers and nurses and salesmen, but it also includes people like Warren Buffet and Elon Musk").

If we wanted to be more complex, the users who have made only one edit fill the entire range from the 71.4th percentile to the 81.7th percentile. They're all in the top 30%, but they are none of them exactly at the 70.000000th percentile.

I've no real objection to changing the table so that it says 1 edit = in the top 30%, and maybe it would also be good to add that 2 edits = in the top 20%. I similarly have no objection to adding a separate calculation based only on those who managed to make an edit. I linked the code earlier, so that people can switch back and forth between the two views. WhatamIdoing (talk) 16:58, 13 April 2023 (UTC)

I'm a bit annoyed, tbh, WAID. It is wrong just as sticking a photo of Musk in the middle of a chart of wealth from poorest to richest is wrong and then claiming, when challenged that well, yes, he is the second richest man in the world, but he's also in the top 50%, so putting him there is mathematically just fine... and now you've moaned about it, I'm going to stick him down at the bottom of the chart, because he's also in the top 99% too. I don't feel this is helping us decide how to explain editor activity on this template, and I don't agree about your definition of accurate, which carries qualities of precision and exactness, not just pedantic correctness. Let's move on. -- Colin°^Talk 19:12, 13 April 2023 (UTC)

Accuracy and precision are completely different things.

The thing is that I'm not sticking anyone in the middle of the chart.

Sticking them in the middle of the chart looks like this:

🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🔴🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡

What I've written looks like this:

🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🔴🔴🔴🔴🔴🔴🔴🔴🔴🔴🔴🔴🔴

The single-edit user – or Elon Musk – isn't "in the middle"; it's anywhere in that large red range.

What you appear to want for the single-edit user looks either more like this:

🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🔴🔴🔴🔴🔴🔴🔴

or like this:

🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🔴🔴🔴🟡🟡🟡🟡

What I don't recommend is this:

🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡🟡①🟡🟡②🟡③⑤⑫

because the single-edit user might not be where the circled number 1 is; that user could be in that position or either of the two to the right of it. There are too many single-edit users to represent them in a single small space. WhatamIdoing (talk) 23:48, 13 April 2023 (UTC)

The definition of editor rests at someone making at least 1 edit. The only editors on the chart should be those who've made one edit or more. That's it, that's all we should count. The key points would be what percentage only make one edit and how many make two, and then 10, etc. Figuring it at that rate can someone break down those numbers in percentages: one edit (100% to __), then two edits, then ten. Thanks. I have an edit in mind which would use those numbers. Randy Kryn (talk) 02:42, 14 April 2023 (UTC)

I have no objection to also providing numbers for users who have made at least one edit, but I want to have a set of numbers for users who have not yet managed to make that first edit. I've seen some of the user testing around this: making the first edit is hard (especially in wikitext), and people give up. WhatamIdoing (talk) 02:53, 14 April 2023 (UTC)

That would be the accurate way to count editors. A Wikipedia editor would be someone who has signed up for an account and made at least one edit (not counting the millions of IP's who have edited, a vase number). Users are different, either they have tried to edit or have not tried, no way to tell how many try and fail, but if they have an account they are at user level. But the main chart of editors should list only those who have edited, per accuracy of the term "editor". I would like to add a sentence in WP:ABOUT telling readers where they would rate if they make two edits and ten edits (also letting them know that if they make one edit they are considered a Wikipedian editor). Giving readers of that page that information what should occur, per human nature, is many people will make two or ten edits just for the bragging rights ("I'm top X%") and, importantly, many of those who make ten edits will likely make more. But before that could be added the chart should be as accurate as possible and used somewhere in main space, which seems to be the goal of this discussion and chart work. Randy Kryn (talk) 03:31, 14 April 2023 (UTC)

No, that would be the accurate way to count the group of people that you choose to call editors. We don't have to have the same definition. I consider myself to be an editor at all the WMF-hosted wikis, even at the ~400 where I haven't published an edit yet. You (apparently) think you stop being an editor until you click the big blue button at a new wiki. That's okay. Neither of us are actually wrong.

I collected this information and created this template because I thought it would be useful and informative for editors and WMF staff (NB: not the mainspace) to know just how hard it is for people to get through the first edit. 70% of them don't even manage to make a first edit. 80% of them don't manage to make two edits. 95% of them don't get past five edits.

If you want to tell a different story, that's okay. If you want to set up something to put in an article, then go ahead. There's unlimited space in the Template: namespace. But I am asking that you not hide the information that I've collected, just so you can tell the story that interests you. WhatamIdoing (talk) 06:28, 14 April 2023 (UTC)

Well the "information" on this template currently doesn't tell anyone that "70% of them don't even manage to make a first edit" because it says 50% <insert sticking tongue out at you emoji here>. And from discussions on this page, there appear to be many ways in which someone might have an account, a USER record, but not actually make any edits. It doesn't necessarily mean anyone "failed" to make that first edit. Wrt being an "editor" at ~400 wikis, just because we have OAth, or because they come under a WMF umbrella, I do think we have established that WhatamIdoing's personal definition of editor is exactly that: WhatamIdoing's personal definition of editor. I think the page/table could well indicate that only about 30% of accounts on Wikipedia have ever made an edit, but can't really go into reasons why this might be so. But it really does not make any statistical sense to rank editors if one includes accounts that aren't editors. For the same reason it doesn't make sense to consider percentiles of grades in a school exam and include students who never even took the subject. I'm sorry Jonny, I know you were in the top half of your class in Geography, but we need to include the History, Sociology and Psychology classes too so you've been put in the bottom quartile for Geography O Grade. It is just daft. Where do we draw the line on these "might have been an editor" people? If WMF are interested in the difficulty of making that first edit with an account, should we include everyone who ever clicked "Create account" and then decided that was too much hassle to think of a user name and password and never submitted the page? Maybe people want that "Login with Google/Facebook/Twitter" option? Many organisations cleanse their databases of old unused accounts, sometimes for EU data protection reasons, so the fact we have all these null accounts kicking about in the en-wp database is kinda odd. If you want to know about challenges of creating accounts and making that first few edits, you only need this month's data, and maybe even web logs, you certainly don't need decades of data about users creating accounts in 2007.

I hoped we could get bigger input, after Randy Kryn's post on talk Jimbo, but it seems Jimbo making a fool of himself has rather distracted the community. Without other views, we've reached a stalemate. I'm not interested in meaningless stats or silly arguments about something not technically being wrong. Unwatching. -- Colin°^Talk 07:37, 14 April 2023 (UTC)

I have agreed multiple times to change the 50–50 framing to the more precise 70–30, but nobody seems to be interested enough to do that. WhatamIdoing (talk) 16:24, 14 April 2023 (UTC)

Something I'd be interested in is the percentage of 0-edit editors who are not indefinitely blocked. I'm sure there's a learning curve to go from zero edits to one for people who are totally unfamiliar with the markup, but there seem to be an extremely large number of 0-edit accounts that never manage to make an edit because they are immediately username-blocked upon registration or cu-blocked as sleeper socks before their master has need for them. Folly Mox (talk) 19:54, 14 April 2023 (UTC)

@Folly Mox, I don't think that it's a significant number. Starting at Special:ListUsers/Folly Mox (as a pseudorandom point in the alphabetical list of editors), I find only a lot of zero-edit accounts, a small number of blocked editors, nobody blocked with zero edits. This search won't identify people who are subject to Wikipedia:Global locks, but I don't expect that to make much difference.

(If you don't already have the script installed, then try Special:Preferences#mw-prefsection-gadgets "Strike out usernames that have been blocked" to make it easier to spot the blocked accounts in that list.) WhatamIdoing (talk) 01:41, 18 April 2023 (UTC)

What's the proper way to handle IP editors? Do we count each IP address that made an edit as one editor, or should we only count named users who made at least one edit? GoingBatty (talk) 03:17, 12 April 2023 (UTC)

We aren't counting IP edits at all, and as I noted above, we don't handle editors with multiple accounts. And for the high-volume edit stats, we filter out bot accounts. If we are being pedantic we could call the thing we are counting edits for is "human editor accounts".

None of our definitions are what I think people would expect, which is an actual human being (counted just once even if they have multiple accounts or edited by IP) who actually edits Wikipedia (made at least one edit). Ask anyone in the real world what is an "editor at Wikipedia" and they would give Randy's definition, and nobody, really nobody, would mention database records or user right flags, unless you happened to ask a WMF developer :-). But we can't count human beings as the system doesn't ask for your national insurance number or passport number or some other ID. The opposite extreme of just counting every USER record, whether they've edited or not, might be a reasonable approach if it wasn't for the weirdness that about 70% of such records have no edits associated with them. It makes the results radically different. The template currently claims you are in the top 1% after making 100 edits and the top 0.1% after making 1,000 edits. But if we look at editors who edit, then you need to make 300 edits to be in the top 1% and 8,000 edits to be in the top 0.1%. -- Colin°^Talk 07:17, 12 April 2023 (UTC)

Most of these complications (e.g., editors with multiple accounts and bots) are so rare that they are just rounding errors. There are 312 bots at this wiki; that's one for every ~125,000 registered users and one for every ~40,000 accounts with at least one edit. Some of them make very few edits (e.g., User:Joe's Null Bot). Others make a very large number. It isn't until you get up into the very high contribution levels that the bots make any difference to the count. We're literally talking about the difference between saying that 28.5748950375% of registered accounts have made an edit vs 28.5744033607% of non-bot registered accounts have ever made an edit. It's less than one-ten thousandth of a difference, and in both cases, we should round it to a plain old 30%. WhatamIdoing (talk) 02:17, 13 April 2023 (UTC)

I agree the bots don't make a difference to low edit counts, only to the very high edit count end of the table, where they become a significant minority of the total users in those groups. If we counted IP edits, though, and could magically work out the human behind them, that would change the low edit counts. What proportion of edits are made from IPs vs accounts? -- Colin°^Talk 10:37, 13 April 2023 (UTC)

This might be known, but the numbers aren't comparable, so I don't think it's worth looking up. WhatamIdoing (talk) 17:00, 13 April 2023 (UTC)

• The English Wikipedia statistic 'registered editors' (i.e. Paine Ellsworth has a related subpage) reads out as misleading if the definition of the term 'editors' includes at least one edit. Signed, Truth in Stats Liberation Coalition (come to our annual parade and bring a friend). Randy Kryn (talk) 13:16, 13 April 2023 (UTC)

  I think we are somewhat stuck with people using the term "editor" in a very loose manner when talking generally alongside "user", "member", "account" and "Wikipedian". I note that Special:Statistics has "User statistics" and the relevant number is "Registered users (list of members)". Of course one can be a "user" of Wikipedia without registering an account and one can also "edit" Wikipedia without registering an account.

  Where it matters is pages like this, where we should be comparing and ranking registered editors against registered editors (you have an account, and you edit), not against USER table records or against anonymous IPs. -- Colin°^Talk 13:52, 13 April 2023 (UTC)

I have added the following comment to the bottom of the table: For the purposes of this table, an "editor" is a person who has an account on the English Wikipedia. If you change the data in the table to use a different definition of "editor", please update the comment as well. Thanks! GoingBatty (talk) 13:29, 14 April 2023 (UTC)

For continuity, I think it is better to keep the existing definition used as the denominator when calculating the percentages and to document it (as GoingBatty has done). I think percentages with a different denominator could go in an additional column or separate table (Wikipedia:Village pump (technical)/Archive 204 § Fancy formatting has a little discussion of some of the choices). isaacl (talk) 16:08, 14 April 2023 (UTC)

I wonder whether it would be helpful to move the template to Template:Registered users by edit count. Perhaps changing that word would reduce the reaction to the word editors. (That said, I do think that implementing {{switcher}} to provide both sets of numbers would have some advantages.) WhatamIdoing (talk) 16:27, 14 April 2023 (UTC)

I was thinking that since there's no issue with having tables using different definitions being available on different pages, the only thing really at issue is what table should be present on this template with this specific name. Personally I feel for continuity it would be better to keep the table with its current definition in the template with the name "Registered editors by edit count". I think it would be confusing from a historical perspective if the template were renamed and a new template with the name "Registered editors by edit count" were created with a table using a different definition. (If the current template name were not used to host a new table and was just the usual redirect to the new name, that would avoid confusion.) isaacl (talk) 05:18, 15 April 2023 (UTC)

• This is beyond pedantic in the worst way. Reading the word "editor" as "someone who can edit" is a perfectly legitimate way of understanding the word. Demanding that we read it to only mean someone who has edited already seems like unnecessarily restrictive. The broader definition of "someone that can edit" is perfectly reasonable. --Jayron32 16:33, 14 April 2023 (UTC)

  I think the discussion about PREFIXDASH easily beats this one.  ;-) WhatamIdoing (talk) 23:02, 14 April 2023 (UTC)

  Summoned from the village pump and I tend to agree. Is a doctor not a doctor until they have seen their first patient? Is a soldier not a soldier if they never get deployed to a combat zone?

  Maybe rephrase the desired stat along the lines of “your edit count is in the top 1% of those who have ever edited Wikipedia”? Barnards.tar.gz (talk) 21:24, 17 April 2023 (UTC)

  I guess I'm an Arabic, Tagalog, and Serbian Wikipedia editor, then. JoelleJay (talk) 03:54, 18 April 2023 (UTC)

I updated the table to (a) use more current numbers and (b) offer both options (see the radio buttons underneath the table), for people who are curious about comparing registered editors vs successful contributors. WhatamIdoing (talk) 01:30, 24 April 2023 (UTC)

Also: Please check the math. The tables are a bit awkward to edit, so if you see a mistake, either fix it yourself or ping me. WhatamIdoing (talk) 01:30, 24 April 2023 (UTC)

@WhatamIdoing: The number of editors and contributors in table 1 and 2 are inconsistent. Is that intended/is there a different definition of editor vs. successful contributor other than the latter has made an edit? Skarmory (talk • contribs) 23:20, 24 September 2023 (UTC)

The only difference is whether the account has made an edit. Some people prefer comparing themselves against people who were successful in making at least one edit, so both calculations are offered. WhatamIdoing (talk) 01:54, 25 September 2023 (UTC)

Then the number of editors in the "all users" and "all contributors" columns in table 1/2 respectively should be the same, unless I'm misunderstanding, and they aren't down to the 100,000 edit mark. I don't quite understand how the template works, but I think the number being pulled for table 2 is wrong (it's using 12952000, but 13800000 doesn't produce the right results either (though it's a lot closer), nor does 14000000, so I don't know what's going on). Skarmory (talk • contribs) 04:27, 25 September 2023 (UTC)

About 45 million people have successfully used Special:CreateAccount. That is table 1.

About 15 million people have successfully used the big blue "Publish" button. That is table 2. WhatamIdoing (talk) 05:03, 25 September 2023 (UTC)

I think you're expecting way more precision out of this than is warranted. I've just updated the second table to say that there are 13 million editors. In reality, as of a few minutes ago, there were 13,187,716 such users, rather than 13 million, and 28.54% of all registered accounts have successfully posted 1+ edits, rather than 30%, but this table is not meant to be precise, and it is not meant to require frequent updating. It is supposed to be an approximation. A heavily rounded number is not wrong or inaccurate; it's just not precise. The next time we update it, maybe there will be 14 million editors. WhatamIdoing (talk) 05:16, 25 September 2023 (UTC)

Fair enough. I've hopefully successfully updated table 2 to automatically use 30% of users as the input as opposed to a manually entered number, which should cut down a bit on the inconsistency and the amount of human upkeep needed, as long as it doesn't break anything. I've never touched actual template expressions; I see no reason for this to break, but who knows if I did something stupid (it looks good for now, at least). There's still a little inconsistency, but short of being able to draw and input the raw numbers for each category, I don't think it can be cut out with how this template works. Skarmory (talk • contribs) 03:13, 2 October 2023 (UTC)

H1 2022 vs all time

I posted this over at WT:MED who were looking at stats for students who took part in Wiki Ed classes in Spring 2022. I thought it might also be useful on this talk page to see that our assumptions about editor edit count aren't necessarily fair if you are looking at relatively new editors. I've taken some of the above rows and supplemented them with looking only at editors who registered in the first half of 2022.

Editors by edit count

Edits made	Number of editors all time	% of all editors all time	Number of editors H1 2022	% of all editors H1 2022
1	12,951,300	100%	255,607	100%
2	8,286,137	64%	158,907	62%
3	6,080,075	47%	114,716	45%
4	4,841,383	37%	88,978	35%
5	4,035,686	31%	72,437	28%
10	2,293,272	18%	39,875	16%
20	1,301,053	10%	21,199	8.3%
50	590,949	4.6%	8,166	3.2%
100	321,314	2.5%	3,872	1.5%
200	183,247	1.4%	2,046	0.80%
500	94,591	0.73%	866	0.34%
1,000	58,418	0.45%	388	0.15%
2,000	36,320	0.28%	171	0.067%

You can see that new editors who make few edits, up to 10 edits say, are a broadly similar percentage in H1 2022 compared to all time. But after that things diverge and you are

• only 80% as likely to make 20 edits
• only 70% as likely to make 50 edits
• only 60% as likely to make 100 edits
• only 50% as likely to make 200 edits
• only 33% as likely to make 1,000 edits
• only 25% as likely to make 2,000 edits.

-- Colin°^Talk 15:30, 7 April 2023 (UTC)

Well, this is certainly interesting and informative. Thank you Colin. Randy Kryn (talk) 04:29, 11 April 2023 (UTC)

Interesting, informative, and not surprising. It took me about six months to make my first 200 edits (if you don't count the fact that I didn't edit at all for several months before creating my account). Colin's early months were similar to mine, and @Randy Kryn took about a year and a half to reach that point. We shouldn't be surprised to discover that most new editors don't edit as much in a few months as long-time editors have ever done. WhatamIdoing (talk) 04:56, 11 April 2023 (UTC)

It is good if the stats are not surprising, because otherwise the stats might be wrong or our assumptions are wrong. There is also a chance that if you looked at editors who joined H1 2009 and analysed their edit count by April 2010, you'd see a different pattern to the above. If, perhaps, we were all more enthusiastic then, or less likely to be discouraged by being reverted, or more vandal accounts were created then. But getting such stats isn't a straightforward query. -- Colin°^Talk 07:56, 11 April 2023 (UTC)

In case there's a grad student in search of a thesis idea, I'd probably pick 2006 as the baseline year. That's just when we started automating everything, and (if memory serves) part of the period covered by that famous paper by @Halfak on what it really takes to make an edit when a bot decides that your edit is bad. WhatamIdoing (talk) 23:45, 11 April 2023 (UTC)

Discrepancy in statistics for Extended Confirmed editors

Currently this page says there are 115,000 users who have made at least 500 edits, but Special:Statistics says that the number of people with Extended Confirmed status is 67,507. The difference is 47,493 editors. EC status requires 30 days of existence in addition to 500 edits, but surely we don't have 47,493 people who have been here less than a month and have all made 500+ edits each, right? Can anyone explain why there's a large discrepancy? Clayoquot (talk | contribs) 17:53, 9 October 2023 (UTC)

@Clayoquot: I *think* that the software only thinks about this status when an edit is made: "Ooh, here is an edit by someone who lacks EC - do they now satisfy the 30/500 requirement?". So all the editors who made 500+ edits years ago and then retired, long before the EC status was invented, have never been given EC. -- John of Reading (talk) 18:03, 9 October 2023 (UTC)

That makes sense. Thank you! Clayoquot (talk | contribs) 00:08, 10 October 2023 (UTC)

This year

The next obvious expansion for this table, if we assume these tables aren't complicated enough, is to run the numbers for a single year. It looks like we'll have somewhere around (or above) 700K registered editors making an edit this year. WhatamIdoing (talk) 02:30, 24 October 2023 (UTC)

If you look only at the editors who are extended confirmed + made an edit during the last 30 days, then there are about 24,000 of them if I didn't screw up quarry:query/79972. WhatamIdoing (talk) 03:40, 27 January 2024 (UTC)

@WhatamIdoing: The quarry link you provided shows me "User not found". GoingBatty (talk) 07:02, 27 January 2024 (UTC)

Perhaps I've mangled the link formatting. Try https://quarry.wmcloud.org/query/79972 WhatamIdoing (talk) 18:28, 27 January 2024 (UTC)

@WhatamIdoing: That linked works! Your code looks like your intent is to exclude bots, but I see MystBot and HooptyBot on your list. GoingBatty (talk) 21:00, 27 January 2024 (UTC)

It's hardly my code; I swiped it from another query. I take all blame but no credit for it.

I don't really want the list of names at all, but I don't know how to make it give me just the number. I only need it to get within a reasonable rounding error. I don't see any sign the HooptyBot was ever flagged as a bot, and MystBot (an interwiki bot) lost its bot flag in 2014. WhatamIdoing (talk) 21:11, 27 January 2024 (UTC)

A few numbers that might interest folks who watch this page:

• In 2023, 812,635 registered editors made at least one edit during the calendar year.
• In 2023, 393,505 registered editors made at least five edits during the calendar year. (This means that 48.4% of editors made 5+ edits, so the median number of edits, if you made any edits last year, is almost certainly five edits.)
  • That query took almost an hour to run, so I'm not running other numbers today.
  • No, that's how many users made at least one edit that calendar year and at least 5 edits total ever. —Cryptic 23:51, 6 February 2024 (UTC)
    • Oh, right! Because the >=5 isn't limited by the timestamp. Now I'm doubly glad that I didn't run that for any other numbers (though if I run this for 10 edits, it would approximate the number of autoconfirmed accounts who made any edits [or other actions?] last year, which might be fun to know anyway). WhatamIdoing (talk) 02:20, 7 February 2024 (UTC)
• In 2023, 1,968,290 new accounts were registered during the calendar year.
• In 2023, 478,218 editors both registered a new account and made at least one edit here during the calendar year. (This means that 58.8% of the registered editors who made any edits last were new editors.)

Many thanks to Cryptic for doing most of the work to make these numbers possible. WhatamIdoing (talk) 23:35, 6 February 2024 (UTC)

Comment re "contributors"

I suggest basing the last row on "users" rather than "contributors" for consistency. —⁠ ⁠BarrelProof (talk) 23:17, 31 May 2024 (UTC)

 Done. Thanks, @BarrelProof. WhatamIdoing (talk) 01:27, 1 June 2024 (UTC)

Boy, that was fast! Thanks! I see that the numbers didn't change. I was assuming "contributors" was equivalent to "users who edited after creating an account", so that the number based on "contributors" would be different than the number based on "users", so the number would change by about a factor of three. The fact that the numbers didn't change makes me doubt that interpretation. —⁠ ⁠BarrelProof (talk) 05:27, 1 June 2024 (UTC)

I think it was just a typo. A million edits is 99.99 99 7% of all users, and we don't calculate that number for all contributors. WhatamIdoing (talk) 20:21, 1 June 2024 (UTC)

(Or perhaps it's the other kind of typo, and that number should be in the other table?) WhatamIdoing (talk) 20:22, 1 June 2024 (UTC)