In April 2011, the community reached a large consensus by RfC for a 6 month trial limiting the creation of articles in mainspace by non-autoconfirmed users. In June 2017 after discussions with the English Wikipedia community, the Wikimedia Foundation agreed that such a trial would be implemented. On 14 September 2017, the Wikimedia Foundation implemented the Autoconfirmed article creation trial (commonly referred to as ACTRIAL) and it ended on 14 March 2018 after 6 months. The results of the research can be found at meta:Research:Autoconfirmed article creation trial and Wikipedia:Autoconfirmed article creation trial/Post-trial Research Report. Three of the major themes of the report were that:

1. New user activity and retention is largely unaffected
2. Creation of pages by new users shifted to draft space from article space
3. Fewer low-quality and inappropriate pages have been created in article space

The independent consultant's research generally shows that the concerns (hypotheses) that were raised before the trial about the restrictions having a negative impact on retention of new editors did not occur, the trial and the research conducted has received generally positive feedback.

Following the trial, and in view of the promise to the community that the rule will not be permanently installed without a further discussion, it is proposed that as from 3 May 2018, (or sooner if a consensus is reached):

Creation of new articles in the main space is restricted to accounts that have reached autoconfirmed status (four days registered and at least ten edits). TonyBallioni (talk) 13:39, 19 March 2018 (UTC)[reply]

1. Support as proposer. While ACTRIAL was not a cure all, it was in my view, an enormous success. For the first time in at least two years, both AfC and NPP are at their sustainable levels: AfC only has 4 drafts at the 60+ day mark, and all of NPP is below the Google index point: we have effectively eliminated the NPP backlog, while maintaining AfC at an adequate level to serve the needs of good faith contributors.
   To me, probably the most significant statistic from the survey was the we reduced the median number of articles deleted a day by 227 while reducing the number of A7s by 84% and the number of G11s per day by 73%. This is probably the single biggest time saving change we've had on-wiki, and while it did reduce the number of patrol action per day, freed up admins and new page reviewers to do other needed work for the encylopedia.
   Finally, I will close by pointing out that anyone who has been near the new pages feed or the CSD categories in the past few days has seen the clear increase in unencyclopedic content starting immediately after ACTRIAL ended, showing it to be necessary to restart as soon as possible. I want to close with this quote from, Shrikantarts', the first article created by a new editor after ACTRIAL ended: Hello, Myself [Name]. Art has always been an inspiration to me. That inspiration credits goes to my Brother and my mom, who stood up always when I started falling apart. I don't think I can make a better case than that. TonyBallioni (talk) 13:39, 19 March 2018 (UTC)[reply]

   I'm sure this number will continually fluctuate over the course of this RFC, but as of this particular point there are zero pending drafts at 60+ days. Primefac (talk) 14:00, 19 March 2018 (UTC)[reply]

   Ah, thanks, even better :) I'd gotten my number a bit before launching this! TonyBallioni (talk) 14:01, 19 March 2018 (UTC)[reply]

2. Support. Obviously. Kudpung กุดผึ้ง (talk) 13:43, 19 March 2018 (UTC)[reply]
3. Support – Trial phase showed considerable net positive outcomes for the project, with none of the feared negatives. — JFG ^talk 13:45, 19 March 2018 (UTC)[reply]
4. Support without qualification. The outcomes of the trial far exceeded expectations in confirming that this remedy actually does work, and has no perceptible undesirable side-effects. Justlettersandnumbers (talk) 13:49, 19 March 2018 (UTC)[reply]
5. Support For sure! Dial911 (talk) 13:52, 19 March 2018 (UTC)[reply]
6. Support. Just like I support waiting periods before buying guns. It's only restricting impulsive creations, not creations generally. It only makes sense that someone right off the street can't just create an article right away. ‡ Єl Cid of ᐺalencia ^ᐐT₳LKᐬ 13:56, 19 March 2018 (UTC)[reply]
7. Strong support Probably the most positive comment I can make about turning autoconfirmed creation off is that it gives me an excuse to go and find more admin candidates to work the NPP and CSD queues. Ritchie333 ^{(talk) (cont)} 13:57, 19 March 2018 (UTC)[reply]
8. Support The postives in terms of the reduction in poor-quality creations far outweight the very minor negative of adding a small extra step during outreach events. SmartSE (talk) 14:00, 19 March 2018 (UTC)[reply]
9. Support - At the end of the day, letting someone create an article with their very first edit isn't just time wasting for us, but there's really just no totally non-BITEY way of telling someone you're deleting their work, no matter how obviously inappropriate it may be, and that goes doubly for users where their very fist interaction ever with another editor is a big scary looking template. It's really just setting people up for failure. AfC has it's own set of challenges that need looked at, but it's still a better process, more likely to provide meaningful feedback, and less likely to summarily communicate "Go back to Facebook. You and your work are equally worthless, and we don't much care for you." GMG^talk 14:04, 19 March 2018 (UTC)[reply]
10. Support Doesn't appear to be any negative effects on say editor retention, but has massive improvements in the quality of the articles, and gives more time for patrollers to focus on afding spam than A7ing articles that took a minute to create. Creating an article requires time, and understanding of wikipedia, which is why the delay makes sense (beyond the empirical evidence of mass improvements in article quality). And, as GMG says, less bitey to prevent someone from creating a bad article than giving a big delete notice Galobtter (pingó mió) 14:06, 19 March 2018 (UTC)[reply]
11. Support - At the level just below that which needs deletion is that which, though promising, isn't ready for mainspace. It bears repeating that it's more encouraging for new users to have their material promoted from draftspace than demoted to draftspace. ACTRIAL wasn't just a protection against trash, it was a less confrontational path for new users to have their material accepted. Cabayi (talk) 14:08, 19 March 2018 (UTC)[reply]
12. Support per Galobtter. Double sharp (talk) 14:20, 19 March 2018 (UTC)[reply]
13. Support per TonyBallioni and Galobtter. JTP ^{(talk • contribs)} 14:28, 19 March 2018 (UTC)[reply]
14. Support New editor attraction and retention were not negatively impacted, and creation of obviously poor new articles was reduced. The only negative unforeseen consequence of note seems to be that the AfC queue has gotten longer. That is and acceptable trade-off for reducing the speedy delete and PROD queues. Eggishorn (talk) (contrib) 14:30, 19 March 2018 (UTC)[reply]
15. Strong Support Great results, even though the AfC backlog is increasing. L293D (☎ • ✎) 14:44, 19 March 2018 (UTC)[reply]
16. Support. The difference between the admin workload and quality of new articles with and without the autoconfirmed requirement is like night and day. It also makes paid for spam stand out sufficiently enough that it can be detected with simple behavioral heuristics. To anticipate/reply to some objections:
    Any solution to this problem must deny proactively, i.e. before the article is created. Otherwise, volunteer time must be wasted in finding, researching, debating and deleting promotional, social networking and/or non-notable crap. Speedy deletion of this rubbish is still a waste of volunteers' time. Automatic flagging of crap articles is not an acceptable solution and the raised CORP notability standards are necessary, but not sufficient. Similarly, NOINDEXing unpatrolled articles helps but there are many new page patrollers out there that cannot recognise paid-for articles. Userspace is NOINDEX but there's still plenty of junk added each day (if it's empty, see my deletion log). Patrollers also become less effective when flooded with garbage.
    Anti-abuse tools in MediaWiki are sufficiently derelict that a subject matter filter can only currently be implemented by the edit filter. (It is near impossible to get any useful software development out of the WMF in any reasonable amount of time.) While I really like the idea of a spammer who types out their article then getting denied when saving it (wastes their time = increases their costs), it's not a good experience if there's false positives and collateral damage. A subject matter filter can be circumvented by a bait and switch approach. Therefore, it must be run on every attempted edit by a non-autoconfirmed user and thus performance questions arise. A solution that denies proactively must run in real time. (I believe ORES runs ex-post.)
    I suspect that we're going to need all alternative measures AND a permanent autoconfirmed restriction in order to keep the spammers out, given how persistent and widespread the abuse can be. A consequence of Wikipedia getting more popular is the need to minimize the amount of time we spend processing each junk page in order to maintain quality. The restriction achieves this. Finally, I had the privilege of deleting the first page created by a non-autoconfirmed user, Shrikantarts', and yes, its creator made no effort to understand what an encyclopedia is before he edited one. MER-C 14:45, 19 March 2018 (UTC)[reply]

1. Oppose A four-day delay is directly disruptive to outreach events such as this which are usually scheduled for a single day during which new editors are trained and encouraged to start new articles about neglected topics. Andrew D. (talk) 13:53, 19 March 2018 (UTC)[reply]

   @Andrew Davidson: are the outputs of these usually suitable articles? Outreach events certainly could still create Draft: articles - and a speedy review/move process could be incorporated. — xaosflux ^Talk 14:04, 19 March 2018 (UTC)[reply]

   @Andrew Davidson:, perhaps if you were to read this you might take note of all the workarounds for editathons. That said, plenty of editathons result in overtime for our New Page Reviewers. Kudpung กุดผึ้ง (talk) 14:13, 19 March 2018 (UTC)[reply]

Regarding GLAM related events and workshops, the solution to that is to make sure an admin is around to promote such accounts to confirmed immediately. Or do you think Tyler Bonner (American drug trafficker) is an acceptable encyclopedia topic? Ritchie333 ^{(talk) (cont)} 13:58, 19 March 2018 (UTC)[reply]

@Ritchie333: showing up means you are not likely to be a random vandal - but it doesn't mean you know how to actually write an article (that is, that initially creating a Draft isn't best). In regards to event flags, once the Program and Events dashboard account creation process is a bit better, I'd be in support of it also having a "+confirmed, 10 day expiration" option built in. — xaosflux ^Talk 14:04, 19 March 2018 (UTC)[reply]

Yes, I think that is the solution, and it's been discussed on the ACTRIAL talk page. I'd support it, and have offered to help draft an RfC on it should this be successful. I did not include it in this proposal, however, because it had previously been rejected by the community twice in August in different forms, so I thought it would be better to have its own RfC rather than have two potentially controversial changes being tested on the same page. TonyBallioni (talk) 14:07, 19 March 2018 (UTC)[reply]

Yup. Writing an article isn't that easy - the topic needs to be notable, need to find sources and avoid close-paraphrasing, write it in encyclopaedic style etc. Somewhat offtopic but I think these events would be better if they focused on improving existing articles. I too would be supportive of a +confirmed - but I think we should make sure event coordinators with that ability reasonably monitor created articles to make sure they are atleast ok (and don't dump a reviewing load on NPP) Galobtter (pingó mió) 14:14, 19 March 2018 (UTC)[reply]

I'd love to have an admin on hand for every event I ran, but logistically that's difficult. For sure, as Tony and others have suggested, some accommodation for GLAM events could be made - I found the workarounds suggested for use during the trial to be only partially useful in maintaining event quality. Lirazelf (talk) 14:15, 19 March 2018 (UTC)[reply]

Lirazelf, yep. Agreed. I don't think anyone is anti-GLAM or anti-outreach, and at the very least, I'm more than willing to put in the effort to get a new proposal and draft guidelines up and running for how we can get Xaosflux's +confirmed, 10 day expiration suggestion a reality. I know some individuals involved in outreach wanted that to be part of this RfC, but given that similar proposals in August failed, and that others whom I trust on and off-wiki suggested that having them as distinct RfCs would be better, I left this as a simple proposal on the creation requirement itself. I'm fully expecting there to be another RfC to make accommodations for GLAM and outreach efforts if this is implemented. TonyBallioni (talk) 14:25, 19 March 2018 (UTC)[reply]

I'd support that "+confirmed, 10 day expiration" too, despite the two previous RfCs (one of which I believe I even opposed myself). Kudpung กุดผึ้ง (talk) 14:17, 19 March 2018 (UTC)[reply]

I don't see the point of an expiration at 10 days when an active user would gain autoconfirmed at 4 days, but the general idea is a good one. Cabayi (talk) 14:33, 19 March 2018 (UTC)[reply]

• Question - Do our outreach events focus on creating new articles or is this just where participants insist on going? Online, we like to encourage new editors to start by improving existing articles. I assume this should be the case in person as well. ~Kvng (talk) 14:41, 19 March 2018 (UTC)[reply]