Wikipedia talk:Autoconfirmed article creation trial/Archive 6

Archive 1

←

Archive 4

Archive 5

Archive 6

Is it working?

Or is it too early to tell?I saw above someone said the backlog was slowly dropping, but is that because creation is down or because many NPPers are trying to cut it down so we can all be back by X-mass? Thanks L3X1 (distænt write) 16:11, 9 October 2017 (UTC)

I wouldn't know about the NPP backlog, but AfC isn't doing well lately. Enterprisey (talk!) 16:41, 11 October 2017 (UTC)

NPP backlog is down to 12,734 from ~15,500. We're also getting a lot of positive feedback from reviewers on the quality of articles and from admins on CAT:CSD being much smaller.

I'm actually not that concerned with the AfC backlog as of yet: it actually stayed stable after ACTRIAL was implemented, and I believe shrunk at one point. Its gone up recently, but also AfC tends to have swings, which can sometimes be attributed to a single active editor stopping for a few days (NPP has this too, but AfC tends to be more impacted by this from what I can see.) TonyBallioni (talk) 16:48, 11 October 2017 (UTC)

My impression is indeed that the new article quality became higher, but it would be great to have some quantitative estimates.--Ymblanter (talk) 19:27, 11 October 2017 (UTC)

I'll second what TonyBallioni says about the data. AfC is well within normal backlog range - the increase in AfC I was expecting did not materialize. So there appear to be less junk pages being created overall. Can anyone figure out if the overall number of new pages is down? There must be some mix of editors with a little experience create less junk and throw away account creators are posting less junk. Legacypac (talk) 19:53, 11 October 2017 (UTC)

Far less junk arriving. I see this because I don't find many to CSD or delete. What I do see is a lot more paid editing, but this does not mean an increase - I'm seeing them now because they stand out more and are not hidden in a crowd of unmitigated trash. Yes, the number of new pages is certainly down, but so will be the number of deletions. Kudpung กุดผึ้ง (talk) 00:29, 12 October 2017 (UTC)

Is everyone finding they accept a higher percentage and faster? Legacypac (talk) 14:54, 12 October 2017 (UTC)

I've been busy both IRL and on WP with other things the past month, so I've been doing less work in the feed itself, but yes. This has been my experience. TonyBallioni (talk) 14:57, 12 October 2017 (UTC)

Yep. I have no idea of what percentage I was skipping/declining before, but it certainly seems like less trash in the queue.L3X1 (distænt write) 16:50, 12 October 2017 (UTC)

I'm finding more trash while patrolling the template-space. Used to be one a day or so, now several a day is not uncommon. I suspect these are users being directed to work in their sandbox, then getting confused by the prefill code in combination with VisualEditor. – Train2104 (t • c) 03:26, 13 October 2017 (UTC)

There is noticeably more crap in userspace as well, but it's only about as bad as it was when userspace was indexed (i.e about 2x baseline). MER-C 04:10, 13 October 2017 (UTC)

The userspace makes sense with the landing page directing people to the option of creating a sandbox. TonyBallioni (talk) 04:12, 13 October 2017 (UTC)

Are you patrolling userspace from the NPP tool? I know there is already a HUGE collection of junk in stale userspace. https://en.wikipedia.org/wiki/Category:Stale_userspace_drafts U5 is ok, but not broad enough. We need to expand G13 to cover userspace as well. Legacypac (talk) 04:57, 13 October 2017 (UTC)

There are, according to the cat, about 32,898 stale userspace items. I don't think we'll ever be able to work through that while the focus is now on detecting paid editing while we have fewer new pages arriving due to ACTRIAL. It's a question of priorities. You can try to start a RfC to expand G13, but experience shows that it's difficult to get a consensus for new CSD crieria. I would support it as would probably many others - we need DGG's expert opinion on this. If we got it, it would need a bot. Kudpung กุดผึ้ง (talk) 05:55, 13 October 2017 (UTC)

There are a couple of abuse filters that catch the worst of the spam. I agree that G13 needs to cover userspace. MER-C 08:23, 13 October 2017 (UTC)

Would support a G13 in userspace as well, but due to the variety of uses available for userspace, I wouldn't support 6 months. Perhaps user inactive for one year or indefinitely blocked would work. – Train2104 (t • c) 11:30, 13 October 2017 (UTC)

I patrol the userspace and find that most of what I CSD is either promo pages which fail G11 and U5 or vandalism pages from sock accounts.L3X1 (distænt write) 13:20, 13 October 2017 (UTC)

We created some dashboards that might be useful for this discussion:

• Daily Non-redirect Pages Created in the Main namespace
• Daily Pages Created in the Draft namespace

And there's also this: https://tools.wmflabs.org/apersonbot/afc-pending/index.py. We should have some preliminary data related to the research questions available in the next few weeks. Kaldari (talk) 17:46, 19 October 2017 (UTC)

9-14-17, a day that will forever live in infamy, if your a spammer :). The charts are clear, after the 14th the drafts have spiked to nearly 3x their usual amount, and mainspace creations are ~40% lower. L3X1 (distænt write) 21:39, 19 October 2017 (UTC)

It's clear that ACTRIAL has succeeded in reducing the number of unwanted new articles. Do we have any feel yet for how many wanted new articles we are losing? Certes (talk) 09:34, 20 October 2017 (UTC)

And now I wish CSD A7 worked on drafts. —PaleoNeonate – 15:13, 20 October 2017 (UTC)

If I’m reading the charts correctly ACTRIAL has driven the daily new Articles from about 1000 a day to about 600 a day while the new Drafts have gone from about 100 a day to about 250 a day. That suggests a net drop of about 250 pages a day of garbage we don’t have to deal with at all and another 150 a day that gets reviewed in Draft instead of Mainspace. I like that a lot. Legacypac (talk) 22:20, 24 October 2017 (UTC)

"That suggests a net drop of about 250 pages a day of garbage" While the data shows that we get approximately that many less articles being created on Wikipedia (anywhere), it doesn't tell you that they were garbage. We might be losing the good faith articles because contributors didn't care or aren't technically literate enough to figure out where they were supposed to be making them, whereas the dedicated spammers still make it through to Draft. Or you might be right and we're only losing garbage. At any rate, I think further analysis is needed on whether the average quality of ACTRIAL draft articles is higher or lower than new articles created by non-autoconfirmed users before the trial, because there's still a chance that we're filtering the wrong users. Sam Walton (talk) 09:28, 26 October 2017 (UTC)

Samwalton9, while I agree with you that more analysis is needed, I do think the early signs are that this is excluding mainly the inappropriate articles (or that at least the quality of pages making it to main space is up). Many of the CSD cats are empty now, and I think I might have placed 3 or 4 non-G5 or G12 tags since ACTRIAL has started. All anecdotal, but they are positive signs.

I do want to address the dedicated spammers point you brought up because it seems to be worrying a lot of people lately that they're gaming ACTRIAL, and I've been asked about it several times off-wiki: ACTRIAL was never meant to deal with the UPE professional firms and their longtime freelancers who used throwaway accounts. We knew they'd find ways to create articles anyway. What it does do is make them significantly more easy to spot and deal with on a behavioral basis, and it also frees up hands that would otherwise be dealing with autobiographers and garage bands to review the better looking commissioned works. Neither of those are directly linked to page creation, but are very real benefits we are seeing come from ACTRIAL. TonyBallioni (talk) 13:45, 26 October 2017 (UTC)

Actually the data does tell us about 250 less garbage pages are getting created because the data before ACTRIAL found about 80% of new pages by new users were compeltely deleted. Many of the remaining 20% were rewritten or redirected. An absolute drop in the number of pages created by new users is a huge success. Legacypac (talk) 17:46, 26 October 2017 (UTC)

BLP PROD is going extinct

Since ACTRIAL started I have not had to tag a single article with a BLP PROD. A quick look at Category:BLP_articles_proposed_for_deletion shows that only a handful of articles are BLP PRODed each week. This is a major change to NPP and I thought it would be worth sharing here. This might also be something to do with the new article wizard, which is much more user friendly (new autoconfirmed users are more likely to use it) and makes it clear that BLPs will be deleted without references. — Insertcleverphrasehere ^{(or here)} 21:13, 20 November 2017 (UTC)

Yes, the backlog is going back up, but I am seeing this and many other improvements with new pages currently. I'm hopeful that the increase is just a fluctuation and will go back down, but there are other benefits that are very easy to see. TonyBallioni (talk) 21:15, 20 November 2017 (UTC)

Yep, in early September (before ACTRIAL), I managed to rack up 20 BLP PRODS, so in my personal experience this is a massive change to the quality of new articles. We still need more help over at New Page Patrol though, as we are struggling to keep the backlog stable (much less reduce it). The backlog has gone up by about 500 in the past couple weeks (13,500 currently). I think that there were a lot of users (myself included) that reviewed a lot of additional articles right after ACTRIAL started, and have since moved on now that the initial excitement has waned. At one point we were reducing the backlog by ~1000 articles a week, but now it is mainly long term regulars that are reviewing most of the pages. We need 30 or so other editors to jump in and join us for a few weeks and we would crush the backlog, but it is difficult to get people involved. — Insertcleverphrasehere ^{(or here)} 21:24, 20 November 2017 (UTC)

Subverting ACTRIAL

If folks involved in the trial didn't notice it please see this ANI thread.Jytdog (talk) 02:41, 7 December 2017 (UTC)

Commented. This looks like an old trick that people used to try with disambiguation pages. The user was actually already autoconfirmed at one point, so it wasn't needed. It was an attempt to get around NPP more than get around ACTRIAL. TonyBallioni (talk) 02:49, 7 December 2017 (UTC)

editathons & new editors

Starting a new section branch of "Effect on outreach activities" above... As a (very) long-time editor, librarian, and someone who runs editathons to get new academic editors (both students and faculty) involved on a regular basis (1-2 events a semester or more) -- I can confirm that Actrial has made outreach events harder. It's discouraging and confusing to new editors and a huge hassle for those training them, and results in new pages piling up at new page review and languishing as drafts. As a result, I think Actrial is a terrible, terrible idea -- an opinion I have formed over the course of training about 100 well-educated people this semester in multiple events. -- phoebe / (talk to me) 22:21, 20 November 2017 (UTC)

Incidentally, I am not an admin, despite having served on the board of the WMF and having been editing for 15-ish years -- not because of anything terrible, and not because I haven't been nominated, but because I've done and wanted to do other kinds of work. So the notes above about getting an admin to do autoconfirms make a lot of assumptions that just aren't true. -- phoebe / (talk to me) 22:24, 20 November 2017 (UTC)

Unfortunately, various fixes that were proposed for allowing editathon organizers to confirm editors manually didn't get much traction. I still think that some solution in this direction is better than trying to say that ACTRIAL is not a good idea. Realistically, New Page Patrol was drowning in a sea of untreated sewage before ACTRIAL started, and now we are at least treading water comfortably. — Insertcleverphrasehere ^{(or here)} 23:54, 20 November 2017 (UTC)

People keep saying that this is a problem (including me) but because this page archives messages after 21 days it just gets hidden :( John Cummings (talk) 09:15, 5 December 2017 (UTC)

We have to look at the greater good. I definitely feel NPP is not the sewage-line, it was two months back, and it's a lot, lot better now with quality articles.So terrible, terrible is not the correct word.

Also, what do you mean by new pages piling up at new page review? That's entirely unrelated with ACTRIAL! As to the problems faced in editathons /outreach activities etc., we can create a special page and advertise it among the organizers, which lists all drafts generated in the process and we can voluntarily prioritize their reviews etc.

Also, why can't you just move sandboxes etc. to mainspace for them when it is ready for publication? And, you can also move drafts out of AfC, if you feel they are suitable for mainspace, at your own discretion!

To end, the community consensus on the matter is that editathon participants should not be automatically exempted from ACTRIAL.So, there's no scope of any magic-moment:--Yo! I created an article! without an expenditure of minimal resources, unless ACTRIAL ends but there's scope for such elation in other wiki-activities.Regards:)Winged Blades ^Godric 09:50, 5 December 2017 (UTC)

I don't see why editathon participants should be accorded any preferential treatment. They are obviously not vandals and spammers, but the facilitators are there for the very purpose of ensuring that the participants only create appropriate articles. That said, history has shown that far form all editathons are as well organized and conducted as those by my good friend RexxS and have on occasions caused considerable turmoil for the New Page Reviewers. Some education programs have ended in very costly WMF catastrophes and requiring a massive, structured clean up by unpaid volunteers. Some of us have long memories. Kudpung กุดผึ้ง (talk) 15:38, 7 December 2017 (UTC)

Kudpung it's not really necessary to set up a page and log the drafts. We have a tool called Programs & Events Dashboard that does this automatically. here is example output from one I attended/facilitated through Cascadia Wikimedians User Group. ☆ Bri (talk) 22:57, 14 December 2017 (UTC)

Bri, I'm not saying all editathons are badly organised but I'm still smarting from the results of a recent one for women in South Africa where we here at NPP were put in the embarrassing situation of having to tell many of them that their good faith creations were about totally unencyclopedic topics. The Programs & Events Dashboard is fine but I doubt that our New Page Patrollers are going to be watching it and lurking in the wings ready to clean up again when editathons go awry - which of course they shouldn't, but do and will, and that why phoebe's assertions above are unfortunately quite incorrect. Kudpung กุดผึ้ง (talk) 23:42, 14 December 2017 (UTC)

I’m going to sell autoconfirmed accounts

Since ACTRAIL is the biggest newbie biting scheme in history im going to set up a dark web market selling accounts so people can edit instantly. I’m also getting a stash of 500 edit accounts too. ACTRIAL (talk) 10:16, 30 January 2018 (UTC)

How kind of you to warn us. Joking aside, I am aware that it is possible to purchase mature accounts which have clocked up enough months and transactions to become a trusted seller for eBay etc. Before rubber-stamping ACTRIAL, please check for any evidence of this sort of black market emerging, and consider the consequences of replacing Wikipedia's traditional accessibility and fairness by a pay-to-edit system. Certes (talk) 10:30, 30 January 2018 (UTC)

Slipping Through the Cracks

I tagged two articles for Articles for Deletion that had Articles for Creation tags but whose author seems not to have passed the criteria for being auto-confirmed. Is there a way to report these issues? I wondered briefly whether there might have been a WMF intervention to turn ACTRIAL off, but I see that it is still in progress. So how do I report apparent slips through the cracks? (Insects are small enough that they often sneak through cracks.) Robert McClenon (talk) 04:09, 6 February 2018 (UTC)

I see what happened. It's not a bug; it's a feature or a misfeature. The editor figured it out well. They edited the two articles in draft space, doing nothing but making edits in draft space, for four days, and achieved autoconfirmed status, and then moved the pages into article space. They gamed the system by clean mathematics. Robert McClenon (talk) 04:15, 6 February 2018 (UTC)

Well, luckily NPP is here to pick up the slack from those that are gaming the system. Good on ya. — Insertcleverphrasehere ^{(or here)} 04:21, 6 February 2018 (UTC)

In this case, it was AFC that caught it. The articles were tagged as having originated as AFC drafts. In any case, they have been nominated for AFD. This may be a case of undisclosed paid editing. Robert McClenon (talk) 11:28, 6 February 2018 (UTC)

Backlog

I expect everyone has received the December NPR newsletter. Eleven years ago Jimbo Wales suggested Wikipedia contributors should shift their focus from the number of articles and instead work on improving their quality.

By preventing through ACTRIAL the possibility for first-time users to create articles directly in mainspace, we have successfully reduced the workload at the New Pages Feed by about 80%. But that does not mean we can retire. For some unexplained reason the backlog has not reduced one iota, unreviewed articles - mainly junk - is being released to Google after remaining unpatrolled for 90 days . And we are now faced with a new challenge: detecting an increase in artspam and other paid editing by users who do their 10 edits and patiently wait 4 days.

In a volunteer collaborative environment, participants can make their own minds up as to where they want to work and the extent of the work they do. One of the biggest challenges to effective New Page Reviewing is incentive and reward, but it’s not like Today’s Article for Improvement or Collaboration of the Week or the WikiCup. Those are all informal projects. NPP is an essential, but dreary core function that simply has to be done by someone, and all the time 24/7. Kudpung กุดผึ้ง (talk) 23:34, 12 December 2017 (UTC)

We can't reduce the backlog, even when we reduce the inflow? Have the NPPers gone on vacation? Legacypac (talk) 18:32, 30 January 2018 (UTC)

Legacypac, the backlog is currently at 3,692 pages. We just completed a backlog drive in January which really drove it down. TonyBallioni (talk) 18:43, 30 January 2018 (UTC)

I suspect there are several thousand pages that are not so bad to attract a CSD/AfD but not so go as to justify "patrolled". The older the creation date the more likely the unpatrolled page fits in this group. Any solutions? Legacypac (talk) 01:31, 8 February 2018 (UTC)

Legacypac It's true that most are borderline articles, some of which need AfD but most will pass NPP. That being said, you still find the occasional copy-paste copyvios and stuff even at the back of the backlog, so we can't just rubber stamp them. The backlog has been pretty much static for the past two weeks, I am hoping that our Newsletter that just went out will inspire a reduction of the backlog (we only need to reduce it by 100 articles per day to meet the zero line by March 14th). — Insertcleverphrasehere ^{(or here)} 12:15, 8 February 2018 (UTC)

Well, we haven't managed to reduce the backlog below about 3.5k, however, by focusing on the back of the backlog, we have managed to reduce the back of the backlog significantly below the 90 day google index point. As such, there are currently no unreviewed indexed articles (google searchable). The backlog has also been stable or falling for the last month or so while we chipped away at the stubborn articles at the back. Considering that we started ACTRIAL with a backlog of over 17 thousand, I think we can call ACTRIAL a resounding success. I am more than a little concerned that returning to the pre-ACTRIAL system is going to undo a lot of our hard work, but accept that this was the price of getting it to happen in the first place. Given that we are barely holding steady at the moment, it is going to take a lot of extra effort by reviewers to be able to even tread water when we go back to the old system in a few days, much less continue to make progress on the NPP backlog. — Insertcleverphrasehere ^{(or here)} 20:51, 11 March 2018 (UTC)

If we could just get another 3 months… L3X1 ◊distænt write◊ 21:21, 11 March 2018 (UTC)

Will WP:ACTRIAL end? It seems like a resounding success. WP:AfC submissions have quadrupled from 40-50 a day to 200-250 a day and there is a big backlog there but the general flood of crap has been seriously reduced. Legacypac (talk) 22:02, 11 March 2018 (UTC)

Yes — JJMC89 (T·C) 22:09, 11 March 2018 (UTC)

So how do we make ACTRIAL into ACPERM? Discuss where? Legacypac (talk) 22:12, 11 March 2018 (UTC)

​ There is supposed to be 'month of discussion' during which the trail will be discontinued, which I assume means a formal 1 month RfC. I'm not sure where this is going to be though. It should be started promptly though. — Insertcleverphrasehere ^{(or here)} 02:32, 12 March 2018 (UTC)

I've been planning on drafting an RfC soonish, but I want to give the WMF people a chance to finish up the research bit. That being said, I expect an RfC will start before the end of March and early April at the latest. TonyBallioni (talk) 02:43, 12 March 2018 (UTC)

Discussion month?

Per the process outlined in the lead of this page, when the 6-month phase of ACTRIAL ends we should have a 'month of discussion'. I assume that this means a formal RfC. Who is preparing the RfC? The WMF or a group of admins? Just looking on some clarity on what the process is going to be. — Insertcleverphrasehere ^{(or here)} 01:27, 12 March 2018 (UTC)

Seconded. Jjjjjjdddddd (talk) 02:49, 12 March 2018 (UTC)

As I mentioned above, I was planning on preparing an RfC (likely with Kudpung since he and I were the two most active people dealing with the initial implementation). I think this should happen by the end of March or early April. I want to give the WMF time to finalize their research numbers, but given that all the feedback I have received has been positive, and many people who were not involved have suggested to me just letting it run on, I think that the community is generally happy with the change, and will likely approve making it permanent. TonyBallioni (talk) 02:53, 12 March 2018 (UTC)

Concurring with Tony, I think we should wait for the WMF to publish their ultimate report. That said, most people are in agreement that ACTRIAL has been an unprecedented success, and that most of the WMF's albeit legitimate hypotheses were proven inaccurate. What we now need to do is concentrate on combating the paid editors who are inserting their sold articles on their 13th edit. I'm not sure that patrollers are giving enough attention to this. Kudpung กุดผึ้ง (talk)

Given that the discussion RfC will probably last 30 days, if we wait until early April to start the RfC we will have about 2 months of non-autoconfirmed article creation, which will kinda suck, but we can live with it if it is necessary. However, I would like some clarification from Nettrom, because from what I can see HERE], the results have already been compiled as of yesterday, including the answers to the research questions. The original discussion where the duration was decided on never intended the month after ACTRIAL to be part of the trial, but rather to be a discussion period and that "A discussion will take place over the 30 days immediately following the end of the trial". As I read the closure of that discussion, we are meant to spend the following month discussing the effects, not waiting on the WMF for more info. I'm a bit confused by a need for such excessive delay. — Insertcleverphrasehere ^{(or here)} 03:28, 12 March 2018 (UTC)

Partially as a courtesy to the people who have spent money doing this trial for us (I know that there isn't much love lost between people who work in NPP and the WMF, but they did pay for us to have a professional conduct research, and I think it is important we acknowledge it). The other part is that a hastily put together RfC rarely achieves results, and it's better to write one when all the information is known. As I said, I expect to launch one by early April at the latest, and I'm fine with going ahead without the WMF final analysis if needed, but I'd prefer not to. TonyBallioni (talk) 03:42, 12 March 2018 (UTC)

All of that is fair, and I have nothing but respect for the work that Nettrom has done. I do notice that most of the 'final results' at the ACTRIAL research page are for the first few months of the trial, and I wonder if these are going to be updated with the stats from the full trial, or if they consider these questions substantially answered? I'm fine with delaying as long as there are good reasons for waiting. — Insertcleverphrasehere ^{(or here)} 04:05, 12 March 2018 (UTC)

[1] shows exactly why we don't go back to the old way. The number of copyvios, Promo, G1 and all kinds of other deletions fell off dramatically. Legacypac (talk) 04:37, 12 March 2018 (UTC)

Of course ACTRIAL reduces the number of unwanted additions. Every step towards locking down Wikipedia filters out more rubbish and also some good new articles. The real question, which no one seems to be answering, is how many babies have been thrown out with the bathwater. Certes (talk) 11:44, 12 March 2018 (UTC)

@Certes I forget where, but there was some previous research that did a breakdown on what happened to articles created by autoconfirmed/non-autoconfirmed/etc. From what I remember something like 80% of non-autoconfirmed submissions ended up getting deleted under the old system. Given that we have had a lot of articles come through AfC during ACTRIAL, I'd say that's where most of the 20% of the previously good ones have ended up. Sure we would have lost some, but not much compared to what we have gained (a stable NPP with editors that don't feel like vomiting constantly). The only argument that I can think of for not permanently continuing ACTRIAL is that it has had a significant effect on AfC, with a significantly increased workload over there. However, while AfC has had an increase of around +80 extra submissions per day, NPP has had a decrease of over 300 submissions per day. Net positive by far. New articles are not the be all end all anyway, most of the vital articles on Wikipedia are crap quality, and need improvement. But we have some of our most experienced editors being overworked babysitting trolls and COI editors at NPP? How does that make sense? Our new user retention hasn't even really changed with ACTRIAL (according to the preliminary stats on the research page), so it seems like new editors realise that they can't just create a new page straight away, so they go work on other stuff and improve the wiki in other ways. — Insertcleverphrasehere ^{(or here)} 12:19, 12 March 2018 (UTC)

Perhaps now is the time to begin revisiting this. Kudpung กุดผึ้ง (talk) 14:11, 12 March 2018 (UTC)

About 80% of non-autoconfirmed creations were fairly quickly deleted. In contrast, about 80% of autoconfirmed creations survive a meaningful length of time. AfC's backlog is up righg now due to the increase of submissions and the loss of several prolific reviewers. However, the increase in AfC backlog is many times smaller than the decrease in NPP backlog due to ACTRIAL. I doubt we are lossing many (any?) good pages. The serious new editor will either get 10 edits/4 days or use AfC. We have long restricted IP page creations and yet very few pages at AfC come from IPs suggesting that the serious IPs who want to create new page adapt and set up an account. Legacypac (talk) 16:02, 12 March 2018 (UTC)

Thank you for the informative reply. I feared that good-faith anons who see "please register to create your article" followed by "thank you for registering; we still won't let you create your article" might give up, but I'l be very happy if that proves not to be the case. Certes (talk) 16:09, 12 March 2018 (UTC)

Based on the research on meta, there doesn't seem to have had any impact on recruiting and retaining new editors (see the data on this. TonyBallioni (talk) 17:20, 12 March 2018 (UTC)

I think it's often forgotten in these discussions just how trivial it is to become autoconfirmed: 4 days and 10 edits. Most new editors will accrue that before they even think of creating an article. For the few that don't, I believe they're directed to start an article as a draft, rather than told that they can't create one entirely. – Joe (talk) 17:22, 12 March 2018 (UTC)

Autoconfirmed article creation during the discussion month

Is there a reason why AC has to be switched off between the end of the trial and when a final decision is reached? We've made tremendous progress in eliminating the NPP backlog over the last few months, it seems a shame to jeopardise that when we are on the cusp of eliminating the it entirely. Also, if the consensus is in favour of AC, we will end up switching it off and on again in a couple of months, which seems unnecessarily confusing for new users. On the other hand, if the consensus is against, we will only have to switch it off once. – Joe (talk) 17:17, 12 March 2018 (UTC)

​Yes. The original discussion that set the terms for the trial clearly state that "The trial will not be extended." (point 6). — Insertcleverphrasehere ^{(or here)} 11:46, 13 March 2018 (UTC)

Thanks for the link. It seems rather silly to be beholden to a seven-year-old RfC, if nobody wants it switched back on now (?), though. – Joe (talk) 12:50, 13 March 2018 (UTC)

Some people are very keen for it to be switched back on, but I also feel that any mandate for it expired long ago. Certes (talk) 13:06, 13 March 2018 (UTC)

Silly to go by that part of the 7-year-old RfC, but not silly to use that 7-year-old RfC as the basis for this trial to begin with? (I still think it's weird that no substantial discussion/RfC was held to at least say "hey, so, still want to do this thing we wanted to do way back when?"... not that it really matters at this point :) ) — Rhododendrites ^talk \\ 20:57, 13 March 2018 (UTC)

It was raised on Jimbo-talk and I think VPP and a few other places, and the general consensus by those commenting was that the old RfC was still valid or at the very least that this had been discussed consistently enough for years that there was still a general consensus that it needed to happen and that we didn't need to test that again. I am for disabling it until a new RfC is held, however, as that is what those of us who were active in getting this to run promised the community: we would abide by the 6 month time frame, and hold an RfC at the end. Whether or not we need to have an entire month off before starting an RfC, I'm less clear on (the WMF seems to have put together some good numbers for analysis that might make that unnecessary).

The ultimate point being: the community agreed to this in 2011 as a trial and Kudpung and I were absolutely clear last year when we were really pushing for this that it would be a trial, and that it would be cut off. I don't think you have to wonder what my views are here in terms of making it ACCREATE or something similar permanent, but we do owe it to the community to consult them again before turning on the technical implementation forever. TonyBallioni (talk) 21:05, 13 March 2018 (UTC)

It should definitely be turned off to see what happens, though if the flow of bad content is unbearable, and consensus is clearly for making it permanent, I don't see any reason it will have to stay off for a month. power~enwiki (π, ν) 21:01, 13 March 2018 (UTC)

• IMO, the one month will be required to fully digest the results and to see if the influx of new articles reverts to the pre trial situation. Kudpung กุดผึ้ง (talk) 05:40, 15 March 2018 (UTC)

It's almost over

I can't wait to be able to create pages again. If ac is made permanent I suggest Wikimedia only accept donations from ac accounts. 84.92.223.20 (talk) 10:24, 13 March 2018 (UTC)

IPs still won't be able to create articles. — Insertcleverphrasehere ^{(or here)} 11:40, 13 March 2018 (UTC)

Humans can create articles. Creating an account has advantages. Create a named account here. --SmokeyJoe (talk) 05:48, 15 March 2018 (UTC)

Obviously "gut feeling" is not really enough to base anything concrete, but I just looked at CAT:A7 and CAT:G11, having been used to it being fairly sparse over the past few months, and it's exploded with articles that are either blatant spam, POV pushing or just utter junk. Waaaah! Ritchie333 ^{(talk) (cont)} 15:39, 15 March 2018 (UTC)

And blatant new spammers already being blocked. Kudpung กุดผึ้ง (talk) 16:43, 15 March 2018 (UTC)

WP:ACREQ started

I've started a page at WP:ACREQ as a centralized place to assemble the case for making Auto-Confirmed REQuired for new mainspace page creations. This should save us typing the same case over and over and be useful when we run the RfC. Legacypac (talk) 02:20, 17 March 2018 (UTC)

A showpiece for ACREQ

I found this gem in the new pages by new editor feed. It had been handled by other NPPers already as the tags show. I google translated it because I understand enough spanish to know it was a problem and have preserved this in Draft space temporarily for illistration purposes. [] The feed is again full of crap from brand new accounts that should never have been allowed to be created. Legacypac (talk) 19:52, 18 March 2018 (UTC)

Is this example representative? Richard Nevell (talk) 20:25, 18 March 2018 (UTC)

Yes pretty much. Page after page by non-auto confirmed users goes down - it's like shooting fish in a very full barrel. I just sent several contentless pages to draft. Another one we are wrangling with copyvio by a persistent COI editor on a non-notable actress. Often the hardest part is deciding which of the 5 reasons that apply to use to seek deletion under. Legacypac (talk) 20:34, 18 March 2018 (UTC)

This kind of thing frequently pops up in discussions who have a dim view of Wikipedia. Something along the lines of 'Wikipedia is terrible, look at how poor this article is'. In something of this size, it's not too hard to find the examples of poor practice. Even showing them the vast numbers of good articles (even tens of thousands of Good and Featured articles) doesn't win them over. If there is value in a process, if there are diamonds in the rough, it is important to find a way to make sure we don't lose those. Richard Nevell (talk) 20:42, 18 March 2018 (UTC)

Here is the feed Special:NewPagesFeed select unreviewed and newcomers. It is full of pages tagged for deletion or that need to be tagged. Legacypac (talk) 20:37, 18 March 2018 (UTC)

There is a way to get the diamonds. Let people get a little bit of experience before they create new pages. The potential benefit of getting a good a valuable new topic from a new editor that refuses to use Draft and refuses to get auto confirmed is not worth the pain to established editors that have to sort through garbage day in and day out. Legacypac (talk) 21:03, 18 March 2018 (UTC)

'Refusing' makes too many assumptions about people's motivations when editing. Richard Nevell (talk) 21:20, 18 March 2018 (UTC)

I find that a lot of people who write pages here are only here to make their page, an then they leave. You see a lot of that in draftspace. Jjjjjjdddddd (talk) 21:10, 18 March 2018 (UTC)

User:Richard Nevell how often do you work NPR or do Admin deletion work? Legacypac (talk) 21:36, 18 March 2018 (UTC)

Very rarely, why? Richard Nevell (talk) 21:38, 18 March 2018 (UTC)

I recognize you are a respected editor but from your comments it is painfully obvious you have no recent experience at NPR or deletion which is why I asked. If you spent some time doing clean up of new acct garbage you would quickly come around to understanding how pie in the sky idealistic these arguments really are. Everyone that is involved in NPR or AfC knows exactly where Wikipedia needs to draw some lines. Until you spend a few hours (as verified by User:Richard Nevell/CSD_log on NPR your comments on ACTRIAL's usefulness is uninformed noise. Heck even if you look at my log - assuming you are an Admin and can see deleted pages - user:Legacypac/CSD_log you might change your mind. Preserving the ability of random idiots to create new pages like the one I linked is a really really bad idea. Legacypac (talk) 22:49, 18 March 2018 (UTC)

I'm sorry you feel that way Legacypac. I had not claimed to be an expert in the NPR process, my area of expertise in this regard is outreach. I would sincerely hope that any discussion could incorporate a broad range of experience to shape what happens and that concerns would not be casually dismissed because the lose voices are perceived as not having done the time. I understand the NPR process and the editors that support it have been under a lot of pressure. There are plusses and minuses to each course of action available. One may outweigh the other, and that should be up for discussion, but implying there are no downsides is sadly inaccurate. Good evening to all involved in this discussion. The night is getting late but I hope to contribute to this evolving issue later in the week. Richard Nevell (talk) 23:45, 18 March 2018 (UTC)

No one is dismissing concerns from outreach editors - but common sense solutions for those concerns are dismissed as unworkable. I see no effort to provide alternative solutions to the crush of bad new pages from outreach people. We can't justify keeping a huge problem to avoid a small one with easy work arounds. Legacypac (talk) 00:23, 19 March 2018 (UTC)

• @Richard Nevell, ACTRIAL had no significant impact on the rate of new article creation. This is clear in the WMF's report (linked further up the page. Likely most of the 20% of good articles by new users made their way through AfC, and the decreased workload on NPP probably was freeing up time for experienced editors to create new articles themselves. Before you ask, there was no significant effect on new editor retention either. The quality of new articles however, has gone up significantly. — Insertcleverphrasehere ^{(or here)} 22:07, 18 March 2018 (UTC)

• Thanks Insertcleverphrasehere (talk · contribs), would you be able to provide a direct link so I can make sure we're on the same page? Richard Nevell (talk) 22:14, 18 March 2018 (UTC)

• Richard Nevell, the WMF research is here. — Insertcleverphrasehere ^{(or here)} 22:55, 18 March 2018 (UTC)

• Thanks, Richard Nevell (talk) 23:36, 18 March 2018 (UTC)