On 5/16/2012 11:04 PM, wikien-l-request(a)lists.wikimedia.org wrote:
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Wed, 16 May 2012 21:14:12 -0400
> From: Gwern Branwen<gwern0(a)gmail.com>
> To: English Wikipedia<wikien-l(a)lists.wikimedia.org>
> Subject: Re: [WikiEN-l] "How the Professor Who Fooled Wikipedia Got
> Caught by Reddit", _The Atlantic_
> Message-ID:
> <CAMwO0gyGeZA0j0ksTz3uNjdUH3zBGY3gqHH7KeTynSz82UFVxA(a)mail.gmail.com>
> Content-Type: text/plain; charset=UTF-8
>
>
>
> Incidentally, I have been finishing an experiment involving the
> removal of 100 random external links by an IP; I haven't analyzed it
> yet, so I don't know the outcome, but this gives us an opportunity!
>
> Would anyone in this thread (especially the ones convinced Wikipedia's
> editing community is in fine shape) care to predict what percentage or
> percentage range they expect will have been reverted?
>
> Or what percentage/percentage range they would regard as an acceptable
> failure-to-revert rate?
>
I just went through 19 random pages (9 of them didn't have any ELs, so I
didn't count them, and I found three articles in which the last EL was
not a useful link. One of them was a spam link to a (non-WMF)
wikiproject, one was a link to a find-a-grave page with a photo of the
subject (unneeded because we already had a photo of the subject), and
the third was a link to the presidential library in which a specific
judge's papers are archived. (That last would be relevant in an article
about the judge, but not so much for the article about the district
court for which he was the chief justice for ten years; I actually went
and added the link to the article on the judge, which didn't have such a
link.) That looks like a 30% fail rate. We'll see how many of them get
reverted, but I suspect that it won't be many, because I didn't go
through and randomly remove ELs, and I edited logged in; for some
reason, people who have been administrators for four years with over
18,000 edits tend to get reverted far less than IP editors. (Go figure.)