Creative Vandalism

Triptych · Unread post by **Triptych** » Thu Feb 12, 2015 3:04 pm

Peryglus wrote:
thekohser wrote:100% chance, thus far. One article in my sample was fairly quickly edited by one of its top contributors, shortly after I edited it. He apparently didn't realize that my edit was comprised of bullcrap.
I'm going to guess it was Montana Secondary Highway 503 (T-H-L)?

Nah, can't be that one because the IP has like 25 edits to its credit, including creating an article (or rather transferring it from wherever the draft was).

thekohser · Unread post by **thekohser** » Fri Feb 13, 2015 2:01 pm

Eight articles thus far in my experiment. None of the vandalisms have been reverted. Seven of the eight articles have not been touched after my edit.

If the experiment continues as such, it strongly suggests to me that "nobody's minding the store" any longer on Wikipedia's light-to-medium viewed articles. Soon, I will try my hand at a very heavily-viewed article, and we'll see how that goes.

Edit: This just in -- one of the articles is related to transportation, and one of the transportation buff editors invited my IP address to create an account, and he did some further gnoming on the article, not catching my hoax.

Flameau · Unread post by **Flameau** » Fri Feb 13, 2015 6:35 pm

I hope you realize that while you are subjecting yourself to identification with such details, but this breaching experiment is certain to succeed. You do not seem to have a poor understanding of Wikipedia, but this experiment by definition questions it. It basically boils down to "hurr hurr I inserted plausible bullshit on a low-traffic page - let's play Wikipedia". A cursory look at actual Wikipedia hoaxes shows your experiment to have much better proof - requiring absolutely no testing on your part. Desert anyone? Amelia Belia? At least you say you will remove the incidents - but your test offers nothing new.

Vigilant · Unread post by **Vigilant** » Fri Feb 13, 2015 8:06 pm

Flameau wrote:I hope you realize that while you are subjecting yourself to identification with such details, but this breaching experiment is certain to succeed. You do not seem to have a poor understanding of Wikipedia, but this experiment by definition questions it. It basically boils down to "hurr hurr I inserted plausible bullshit on a low-traffic page - let's play Wikipedia". A cursory look at actual Wikipedia hoaxes shows your experiment to have much better proof - requiring absolutely no testing on your part. Desert anyone? Amelia Belia? At least you say you will remove the incidents - but your test offers nothing new.

Don't forget the lulz...

MoldyHay · Unread post by **MoldyHay** » Fri Feb 13, 2015 9:02 pm

Flameau wrote:I hope you realize that while you are subjecting yourself to identification with such details, but this breaching experiment is certain to succeed. You do not seem to have a poor understanding of Wikipedia, but this experiment by definition questions it. It basically boils down to "hurr hurr I inserted plausible bullshit on a low-traffic page - let's play Wikipedia". A cursory look at actual Wikipedia hoaxes shows your experiment to have much better proof - requiring absolutely no testing on your part. Desert anyone? Amelia Belia? At least you say you will remove the incidents - but your test offers nothing new.

What are they going to do, ban him?

Triptych · Unread post by **Triptych** » Fri Feb 13, 2015 9:37 pm

More garbage ideas if you get stuck.

Waco traffic circle: The site on which the Waco traffic circle now stands in 1825 was the last stand of the north Texas Comanche, in which they held out two days against France's sixth regiment.

The American popular TV show "The Honeymooners" starring Jackie Gleason had a short-lived Norwegian television knock-off "De Lykkelige Nygifte," which in its 8th and last week featured a guest appearance that turned out to be the last acting role of renowned German actor Gustav Froehlich, once featured in Metropolis.

In "Taxicab" insert the parenthetical "(or Putney carriages)" after "carriages" and before "also known as" in its following sentence: "Hackney carriages, also known as public hire, hailed or street taxis, licensed for hailing throughout communities."

thekohser · Unread post by **thekohser** » Fri Feb 13, 2015 10:57 pm

Flameau wrote:...but your test offers nothing new.

Quantitative data and analysis.

A counterpunch to the prevailing wisdom that "most vandalism is reverted within minutes".

Flameau · Unread post by **Flameau** » Fri Feb 13, 2015 11:56 pm

Thekohser - I am not entirely a fan of your ethics, but we both know that is bullshit. Your quantitative data and analysis approach is a rounding error in the grand scheme of things. Not only are you wasting time on something you already know - but your approach leaves much to be desired. The only interesting aspect of this experiment is that you will undo such insertions.

In a serious question, out the hundreds of hoaxes and other bullshit insertions that complete the citogensis process what will your experiment accomplish? The Amelia Bedelia page had a hoax that got many thousands of hits (usually 100+ page views a day) for the course of a few years. Another affected page "Peggy Parish" was also discovered and corrected after more than four years. The difference? Within months the Amelia Bedelia hoax was "fact" and could be sourced to numerous publications. Including one that purports that Herman Parish confirmed the hoax in one article titled: "Amelia Bedelia author credits notebook for success".

Your attempts here represent a distraction from a deeper issue...

EricBarbour · Unread post by **EricBarbour** » Fri Feb 13, 2015 11:58 pm

Flameau wrote:Thekohser - I am not entirely a fan of your ethics, but we both know that is bullshit. Your quantitative data and analysis approach is a rounding error in the grand scheme of things.

Wikipedia is nothing but a massive collection of "rounding errors". That's part of Greg's point.

Vigilant · Unread post by **Vigilant** » Sat Feb 14, 2015 12:01 am

Flameau wrote:Thekohser - I am not entirely a fan of your ethics, but we both know that is bullshit. Your quantitative data and analysis approach is a rounding error in the grand scheme of things. Not only are you wasting time on something you already know - but your approach leaves much to be desired. The only interesting aspect of this experiment is that you will undo such insertions.

In a serious question, out the hundreds of hoaxes and other bullshit insertions that complete the citogensis process what will your experiment accomplish? The Amelia Bedelia page had a hoax that got many thousands of hits (usually 100+ page views a day) for the course of a few years. Another affected page "Peggy Parish" was also discovered and corrected after more than four years. The difference? Within months the Amelia Bedelia hoax was "fact" and could be sourced to numerous publications. Including one that purports that Herman Parish confirmed the hoax in one article titled: "Amelia Bedelia author credits notebook for success".

Your attempts here represent a distraction from a deeper issue...

Well sweetie, don't be so coy.

thekohser · Unread post by **thekohser** » Sat Feb 14, 2015 12:13 am

I was wondering who this Flameau is, with all of 25 posts in nine months on Wikipediocracy.

Then I remembered why "he so angry" at me. He's the nitwit who said:

...vandalism is down and the speed and accuracy of vandal reversions have greatly improved...

He doesn't like my experiment because it's proving him wrong.

Flameau · Unread post by **Flameau** » Sat Feb 14, 2015 12:40 am

Is that an admission that you don't understand what I am referring to? You are making a test where you cannot fail, knowing that the stylometric approach and normal detection methods will not even flag your edits. When you have an institution forging documents to make a fake page with sources - you are off the deep end into a misinformation campaign. Whether its a fictional war, clitoris research or Cameroonian maid citations, these cases of citogenesis are a small set of "knowns" that originated from such acts. I've caught authors stealing right off Wikipedia and I've caught them not by plagiarism, but the deliberate insertions of misinformation. Encyclopedias used to have fake topics as well - with Appletons' Cyclopædia of American Biography being amongst the most insidious and persistent.

I don't like what you are doing because you are not playing with any risk and are not shining a light on Wikipedia's deeper citogenesis and ingrained hoaxes that can be legitimized in the same way Appleton's fakes have been. I'm not angry at you, just disappointed that you put so much time and energy into such a worthless case study. I have higher expectations of you, especially since you are into the whole paid aspect.

Lukeno94 · Unread post by **Lukeno94** » Sat Feb 14, 2015 1:07 am

And who are you, exactly?

thekohser · Unread post by **thekohser** » Sat Feb 14, 2015 12:59 pm

Lukeno94 wrote:And who are you, exactly?

I think he might be my mother, brought forward in a time machine from 1977, considering the admonishing tone.

And it's really sad that she doesn't distinguish between its and it's.

Triptych · Unread post by **Triptych** » Sat Feb 14, 2015 3:04 pm

thekohser wrote:
Lukeno94 wrote:And who are you, exactly?
I think he might be my mother, brought forward in a time machine from 1977, considering the admonishing tone.

Heh heh heh. It'd be interesting to know who Flameau is because he or she is clearly well-versed and opinionated in this stuff and thus might be a well-know Wikipedia or even WMF figure, however of course Wikipediocracy doesn't require that of its participants and rightfully not.

I disagree with Flameau that this experiment is a foregone conclusion, that every edit slips by all the mindless-but-hyperactive patroller drones and also genuine contributors fail to notice them. I think the edits could indeed crash and burn.

Poetlister · Unread post by **Poetlister** » Sat Feb 14, 2015 8:44 pm

EricBarbour wrote:
Flameau wrote:Thekohser - I am not entirely a fan of your ethics, but we both know that is bullshit. Your quantitative data and analysis approach is a rounding error in the grand scheme of things.
Wikipedia is nothing but a massive collection of "rounding errors". That's part of Greg's point.

The problem is that you're wrong. If Wikipedia were totally the pile of steaming nonsense that some people here think it is, it would have crashed in flames years ago. In fact, the bulk of Wikipedia is OK; a significant proportion is good. As a result, many people believe that it's all good; that's what makes it so dangerous.

Flameau · Unread post by **Flameau** » Sun Feb 15, 2015 5:26 pm

Poetlister wrote:The problem is that you're wrong. If Wikipedia were totally the pile of steaming nonsense that some people here think it is, it would have crashed in flames years ago. In fact, the bulk of Wikipedia is OK; a significant proportion is good. As a result, many people believe that it's all good; that's what makes it so dangerous.

Exactly - for each well-managed and well-written article there are volumes of undeveloped articles that do comparatively little for readers. Wikipedia is prone to error and the deliberate insertion of false material by a knowledgeable person is likely to remain on the fringes of the content pool, but the content is there and improving the issues should be the priority and focus. Better tools, better management, better content production. That is the way to correct Wikipedia - not dousing it in gasoline and hoping it degrades or burns itself down.

The worst issues on Wikipedia are personified by the Gamergate mess, proving that current Wikipedia policy and management is unable to respond in a timely and reasonable manner.

Vigilant · Unread post by **Vigilant** » Sun Feb 15, 2015 6:18 pm

Flameau wrote:
Poetlister wrote:The problem is that you're wrong. If Wikipedia were totally the pile of steaming nonsense that some people here think it is, it would have crashed in flames years ago. In fact, the bulk of Wikipedia is OK; a significant proportion is good. As a result, many people believe that it's all good; that's what makes it so dangerous.
Exactly - for each well-managed and well-written article there are volumes of undeveloped articles that do comparatively little for readers. Wikipedia is prone to error and the deliberate insertion of false material by a knowledgeable person is likely to remain on the fringes of the content pool, but the content is there and improving the issues should be the priority and focus. Better tools, better management, better content production. That is the way to correct Wikipedia - not dousing it in gasoline and hoping it degrades or burns itself down.

The worst issues on Wikipedia are personified by the Gamergate mess, proving that current Wikipedia policy and management is unable to respond in a timely and reasonable manner.

Close. The worst issues on wikipedia are the people.

thekohser · Unread post by **thekohser** » Sun Feb 15, 2015 6:18 pm

I've been adding some articles that get between 1,000 and 2,000 daily page views. Should these vandalisms survive for a month, that may dispel the theory that vandalism only persists for a long time in articles that have a relatively low visibility.

Flameau · Unread post by **Flameau** » Sun Feb 15, 2015 6:33 pm

Your insertions are deliberate misinformation. It is distinguished from common vandalism. This is not the equivalent of "I go poopie now" or content blanking. As long as you make the distinction - I'll agree that your endeavor has some merit and use, but I still think it is ethically poor.

Johnny Au · Unread post by **Johnny Au** » Sun Feb 15, 2015 10:05 pm

thekohser, at least you are adding stuff ClueBot NG will never catch. Flameau is right that Mr. Kohs is ethically poor, but with some merit.

Soon enough, ClueBot NG would smarten up.

DanMurphy · Unread post by **DanMurphy** » Sun Feb 15, 2015 10:45 pm

Flameau wrote:Your insertions are deliberate misinformation. It is distinguished from common vandalism. This is not the equivalent of "I go poopie now" or content blanking. As long as you make the distinction - I'll agree that your endeavor has some merit and use, but I still think it is ethically poor.

You dolt. "I go poopie" does almost no harm. Vandalism that mechanicals like you and the actual robots can't detect is the harmful stuff. In a nutshell: "Wikipedia's mechanisms catch and detect 100% of the vandalism that doesn't matter. It just misses most of the vandalism that does matter."

Far more ethically-poor than Greg's experiment is allowing anyone to do it, while lying to the public that Wikipedia is an ethical and reliable reference work, sitting on piles of money all the while, not a penny of which is spent on improving and protecting content.

Jagged85 filled hundreds of Wikipedia articles on important historical topics with lies for 7 years. He was only kicked off Jimbo's Island after he messed around with the [b]video game articles[/b].

What's your editing handle at Wikipedia, by the way?

EricBarbour · Unread post by **EricBarbour** » Sun Feb 15, 2015 11:01 pm

Johnny Au wrote:Soon enough, ClueBot NG would smarten up.

The first version of ClueBot dates from July 2007. They've had nearly eight years, and they still haven't figured out how to handle subtle vandalism of the type Greg is inserting. Which leads me to assume they never will.

Johnny Au · Unread post by **Johnny Au** » Sun Feb 15, 2015 11:30 pm

EricBarbour wrote:
Johnny Au wrote:Soon enough, ClueBot NG would smarten up.
The first version of ClueBot dates from July 2007. They've had nearly eight years, and they still haven't figured out how to handle subtle vandalism of the type Greg is inserting. Which leads me to assume they never will.

Yes, but ClueBot NG (T-C-L) took over. NG stands for "Next Generation" just like Star Trek and Degrassi.

Lukeno94 · Unread post by **Lukeno94** » Sun Feb 15, 2015 11:31 pm

I think they're capable of doing so. However, certain members of the community would probably freak out over it trying to make the valid reverts on the more subtle vandalism, and would eventually get the bot stripped back to its current state.

Notvelty · Unread post by **Notvelty** » Mon Feb 16, 2015 12:13 am

DanMurphy wrote: What's your editing handle at Wikipedia, by the way?

Could be anyone of a number of a number of Wikipedians who really shouldn't be anywhere near any social enterprise without supervision.

It's why I haven't responded with my usual. I'm not sure the guy's capable of understanding.

Edit: Narrowing down, we know that they are a rules-focused person, with a preference for bots and a lack of preference for the current group running feature articles. Like any good Wikipedia, he knows that everything would be perfect, if only he were running it.

Hmm.. I wonder if he does any web development for.... different sorts of websites.

Triptych · Unread post by **Triptych** » Mon Feb 16, 2015 12:37 am

Johnny Au wrote:Yes, but ClueBot NG (T-C-L) took over. NG stands for "Next Generation" just like Star Trek and Degrassi.

If Cluebot NG is so smart, why don't they just have it write the articles?

Flameau · Unread post by **Flameau** » Mon Feb 16, 2015 5:47 am

Notvelty wrote:....

I don't claim to have the answers and I am certainly unable to address issues with my status. You won't find me amongst the admins and that is why the few people who know it will understand why my identity is not relevant. Sorry to disappoint, but I'm not fit or willing to be an admin. And I am not able to change anything.

Notvelty · Unread post by **Notvelty** » Mon Feb 16, 2015 7:24 am

Flameau wrote:
Notvelty wrote:....
I don't claim to have the answers and I am certainly unable to address issues with my status. You won't find me amongst the admins and that is why the few people who know it will understand why my identity is not relevant. Sorry to disappoint, but I'm not fit or willing to be an admin. And I am not able to change anything.

Rug sweeper. Interested enough to complain about people mentioning the smell, but not interested enough to, at the very least, stop filling the cesspool.

I'd tell you to piss off and play with the rest of the hand wringers, but Zoloft gets all cranky with me when I do that.
I do indeed. -Zoloft

thekohser · Unread post by **thekohser** » Mon Feb 16, 2015 8:49 pm

thekohser wrote:Eight articles thus far in my experiment. None of the vandalisms have been reverted.

Up to twelve articles now. Not a single falsehood reverted, including one that gets about 1500 to 2000 page-views per day, where I suggested that the human body is capable of generating volcanic rock.

I have to admit, even I didn't think the situation was this bad on Wikipedia, where the "crowd" mind has been so enfeebled that hundreds of readers can't even spot (or perhaps spot, but can't be bothered to revert) misinformation when it's presented to them.

DanMurphy · Unread post by **DanMurphy** » Mon Feb 16, 2015 10:16 pm

thekohser wrote:
thekohser wrote:Eight articles thus far in my experiment. None of the vandalisms have been reverted.
Up to twelve articles now. Not a single falsehood reverted, including one that gets about 1500 to 2000 page-views per day, where I suggested that the human body is capable of generating volcanic rock.

I have to admit, even I didn't think the situation was this bad on Wikipedia, where the "crowd" mind has been so enfeebled that hundreds of readers can't even spot (or perhaps spot, but can't be bothered to revert) misinformation when it's presented to them.

I would say 90% of the Wikipedia articles I happen across contain falsehoods - most probably unintentional. The more I know about a subject the likelier I am to spot the error. And that's without getting into the binary categorization insanity (some unknown Wikipedia blogger's judgement about who was the "strategic victor" in some inconsequential and forgotten battle in some interminable war.)

Sometimes you don't need to know anything - mutually exclusive info is contained in the text (for instance the so-called "infobox" says the restaurant chain has 1,058 locations, while the text says it has 632) or clearly outdated information ("so-and-so is currently serving as a senior analyst at the Council on Foreign Relations;' click link, which is dated 2008 and mentions a one-year appointment only).

Then there are the everyday howlers like "igneous granite" as I'm sure some of you remember. See those all the time too.

Vigilant · Unread post by **Vigilant** » Mon Feb 16, 2015 11:24 pm

DanMurphy wrote:
thekohser wrote:
thekohser wrote:Eight articles thus far in my experiment. None of the vandalisms have been reverted.
Up to twelve articles now. Not a single falsehood reverted, including one that gets about 1500 to 2000 page-views per day, where I suggested that the human body is capable of generating volcanic rock.

I have to admit, even I didn't think the situation was this bad on Wikipedia, where the "crowd" mind has been so enfeebled that hundreds of readers can't even spot (or perhaps spot, but can't be bothered to revert) misinformation when it's presented to them.
I would say 90% of the Wikipedia articles I happen across contain falsehoods - most probably unintentional. The more I know about a subject the likelier I am to spot the error. And that's without getting into the binary categorization insanity (some unknown Wikipedia blogger's judgement about who was the "strategic victor" in some inconsequential and forgotten battle in some interminable war.)

Sometimes you don't need to know anything - mutually exclusive info is contained in the text (for instance the so-called "infobox" says the restaurant chain has 1,058 locations, while the text says it has 632) or clearly outdated information ("so-and-so is currently serving as a senior analyst at the Council on Foreign Relations;' click link, which is dated 2008 and mentions a one-year appointment only).

Then there are the everyday howlers like "igneous granite" as I'm sure some of you remember. See those all the time too.

I'm reminded that EVERY SINGLE TIME I've been associated with an event covered by the press, the reporter has gotten at least 3 things wrong in the subsequent article.

These are the "reliable sources" that comprise the bones of wikipedia's mess.

Triptych · Unread post by **Triptych** » Mon Feb 16, 2015 11:33 pm

thekohser wrote:Up to twelve articles now. Not a single falsehood reverted, including one that gets about 1500 to 2000 page-views per day, where I suggested that the human body is capable of generating volcanic rock.

I assume you're couching what you're doing in words that won't give it away to the Wikipediots that are no doubt reading this thread, and going back to hunt for you. I think the lesson you need to heed based on Karmafist is: "get brazen, get busted."

Unread post by **Zoloft** » Tue Feb 17, 2015 5:10 am

Triptych wrote:
thekohser wrote:Up to twelve articles now. Not a single falsehood reverted, including one that gets about 1500 to 2000 page-views per day, where I suggested that the human body is capable of generating volcanic rock.
I assume you're couching what you're doing in words that won't give it away to the Wikipediots that are no doubt reading this thread, and going back to hunt for you. I think the lesson you need to heed based on Karmafist is: "get brazen, get busted."

Greg has already been 'busted' and broke out, then had dinner in the Warden's kitchen at his leisure.

thekohser · Unread post by **thekohser** » Tue Feb 17, 2015 5:12 am

Triptych wrote:
thekohser wrote:Up to twelve articles now. Not a single falsehood reverted, including one that gets about 1500 to 2000 page-views per day, where I suggested that the human body is capable of generating volcanic rock.
I assume you're couching what you're doing in words that won't give it away to the Wikipediots that are no doubt reading this thread, and going back to hunt for you. I think the lesson you need to heed based on Karmafist is: "get brazen, get busted."

If they want to look up every type of igneous mineral and search those against every article that has anything to do with the human body, all so that they can spot one instance of vandalism from a unique IP address having no other edits to Wikipedia, then I'm guessing they will spend at least 10x, if not 20x, more time than I did on that particular edit. It's up to them, if that cost/benefit ratio is sufficiently enticing to them.

P.S. It's not so obvious as "shitting a cinder block".

Hex · Unread post by **Hex** » Tue Feb 17, 2015 6:15 pm

Flameau wrote: You won't find me amongst the admins and that is why the few people who know it will understand why my identity is not relevant.

Perhaps you could explain why you chose an anime Nazi as your avatar instead?

Jim · Unread post by **Jim** » Tue Feb 17, 2015 6:23 pm

Hex wrote:
Flameau wrote: You won't find me amongst the admins and that is why the few people who know it will understand why my identity is not relevant.
Perhaps you could explain why you chose an anime Nazi as your avatar instead?

I think subtleties like that are best left to our imaginations. There's generally an innocent explanation.

Vigilant · Unread post by **Vigilant** » Tue Feb 17, 2015 6:29 pm

Jim wrote:
Hex wrote:
Flameau wrote: You won't find me amongst the admins and that is why the few people who know it will understand why my identity is not relevant.
Perhaps you could explain why you chose an anime Nazi as your avatar instead?
I think subtleties like that are best left to our imaginations. There's generally an innocent explanation.

It's a helpful brain teaser to jump start our morning after a holiday...

Unread post by **Zoloft** » Tue Feb 17, 2015 8:56 pm

Vigilant wrote:
Jim wrote:
Hex wrote:
Flameau wrote: You won't find me amongst the admins and that is why the few people who know it will understand why my identity is not relevant.
Perhaps you could explain why you chose an anime Nazi as your avatar instead?
I think subtleties like that are best left to our imaginations. There's generally an innocent explanation.
It's a helpful brain teaser to jump start our morning after a holiday...

Nazi zombie antagonist in Hellsing. But that's not what's at stake.

thekohser · Unread post by **thekohser** » Thu Feb 19, 2015 11:44 am

I finally had one of my vandal edits reverted. I was getting worried that the Wikipediots might not get any at all, so I hit a botany article that gets more than 500 page views on most days, and placed a couple of howlers right in the lede paragraph. They were detected and reverted after about 75 minutes... so, maybe 25 page views. Good job, Wikipedia!

thekohser · Unread post by **thekohser** » Tue Feb 24, 2015 3:13 pm

Anyone care for an update on my project?

20 articles vandalized, beginning about three weeks ago (an average of one new vandalism per day). Generally, I have moved from infrequently-viewed articles to more frequently-viewed articles. Most of the vandalism attempts have been "buried" in the middle of an article. Some attempts are "sourced" (to reference materials that don't actually say what the vandalism purports), and other attempts are simply unattributed to anything. Most of the content sounds "reasonable" to someone who only knows how to read English, but would strike any intelligent person as "silly" if they stopped to think about what is being said, or knew the exact definitions of the more complex words being used.

Thus far:

1 article vandalism reverted (after 5 hours, about 100 page views)
1 article vandalism reverted (after 80 minutes, about 30 page views)
1 article vandalism reverted (after 50 minutes, about 100 page views), then reverted again after a 2nd attempt (after nearly 12 hours, 1475 page views)
17 article vandalisms intact (about 40% of which have been subsequently edited by someone else, thus "covering up" the most-recent-edit phenomenon)

How will that sit with Wikipedians, that 85% of "properly contrived" article vandalisms left by different drive-by IP addresses may go completely unfixed for days or weeks on end? That sort of slaps down the commonly-offered praise of Wikipedia, that "most errors are quickly reverted", doesn't it?

Jim · Unread post by **Jim** » Tue Feb 24, 2015 3:23 pm

thekohser wrote:Anyone care for an update on my project?

20 articles vandalized, beginning about three weeks ago (an average of one new vandalism per day). Generally, I have moved from infrequently-viewed articles to more frequently-viewed articles. Most of the vandalism attempts have been "buried" in the middle of an article. Some attempts are "sourced" (to reference materials that don't actually say what the vandalism purports), and other attempts are simply unattributed to anything. Most of the content sounds "reasonable" to someone who only knows how to read English, but would strike any intelligent person as "silly" if they stopped to think about what is being said, or knew the exact definitions of the more complex words being used.

Thus far:

1 article vandalism reverted (after 5 hours, about 100 page views)
1 article vandalism reverted (after 80 minutes, about 30 page views)
1 article vandalism reverted (after 50 minutes, about 100 page views), then reverted again after a 2nd attempt (after nearly 12 hours, 1475 page views)
17 article vandalisms intact (about 40% of which have been subsequently edited by someone else, thus "covering up" the most-recent-edit phenomenon)

How will that sit with Wikipedians, that 85% of "properly contrived" article vandalisms left by different drive-by IP addresses may go completely unfixed for days or weeks on end? That sort of slaps down the commonly-offered praise of Wikipedia, that "most errors are quickly reverted", doesn't it?

On the face of it, yes. And, in fact, yes. Nevertheless it will be explained away by virtue of the fact that it is "clever trolling" (like no other clever trolls exist or need to be defended against), or "just trying to prove a point" - well, duh.

To a degree it just confirms what we already "knew", but I stuck my elbow in my kid's bathwater for a long time. She's 12 now, and I still yell out "make sure that's not too hot".

The Adversary · Unread post by **The Adversary** » Tue Feb 24, 2015 7:42 pm

Vigilant wrote:
DanMurphy wrote: I would say 90% of the Wikipedia articles I happen across contain falsehoods - most probably unintentional. The more I know about a subject the likelier I am to spot the error. And that's without getting into the binary categorization insanity (some unknown Wikipedia blogger's judgement about who was the "strategic victor" in some inconsequential and forgotten battle in some interminable war.)

Sometimes you don't need to know anything - mutually exclusive info is contained in the text (for instance the so-called "infobox" says the restaurant chain has 1,058 locations, while the text says it has 632) or clearly outdated information ("so-and-so is currently serving as a senior analyst at the Council on Foreign Relations;' click link, which is dated 2008 and mentions a one-year appointment only).

Then there are the everyday howlers like "igneous granite" as I'm sure some of you remember. See those all the time too.
I'm reminded that EVERY SINGLE TIME I've been associated with an event covered by the press, the reporter has gotten at least 3 things wrong in the subsequent article.

These are the "reliable sources" that comprise the bones of wikipedia's mess.

True on both the above.

In addition: I once asked to see my MD´s journal on me (yes, we can demand that in my country).
It said: XX smokes moderately, never drinks alcohol, and have just returned from several months to West Africa.

Well, I haven´t smoked a cigarette (or anything else!) for decades. I love a glass (or more) of white wine, and I have never been in West Africa. (Had just returned from East Africa, though..)
And I hadn´t lied to my MD.

Ah, those details, details...

Lukeno94 · Unread post by **Lukeno94** » Tue Feb 24, 2015 10:28 pm

There's a couple of articles that include my real name, stating that I scored for a local football club, when in fact I wasn't even playing in the match (although I am registered with them).

Unread post by **Zoloft** » Fri Feb 27, 2015 1:13 am

thekohser wrote:Anyone care for an update on my project?

20 articles vandalized, beginning about three weeks ago (an average of one new vandalism per day). Generally, I have moved from infrequently-viewed articles to more frequently-viewed articles. Most of the vandalism attempts have been "buried" in the middle of an article. Some attempts are "sourced" (to reference materials that don't actually say what the vandalism purports), and other attempts are simply unattributed to anything. Most of the content sounds "reasonable" to someone who only knows how to read English, but would strike any intelligent person as "silly" if they stopped to think about what is being said, or knew the exact definitions of the more complex words being used.

Thus far:

1 article vandalism reverted (after 5 hours, about 100 page views)
1 article vandalism reverted (after 80 minutes, about 30 page views)
1 article vandalism reverted (after 50 minutes, about 100 page views), then reverted again after a 2nd attempt (after nearly 12 hours, 1475 page views)
17 article vandalisms intact (about 40% of which have been subsequently edited by someone else, thus "covering up" the most-recent-edit phenomenon)

How will that sit with Wikipedians, that 85% of "properly contrived" article vandalisms left by different drive-by IP addresses may go completely unfixed for days or weeks on end? That sort of slaps down the commonly-offered praise of Wikipedia, that "most errors are quickly reverted", doesn't it?

I assume this isn't yoursI call it ComCast vs ClueBot:

Code: Select all

November 2013

Hello, and welcome to Wikipedia. This is a message letting you know that one of your recent edits to Miguel de Cervantes has been undone by an automated computer program called ClueBot NG.

    Lick my nuts mr. post man ClueBot NG makes very few mistakes, but it does happen. If you believe the change you made was constructive, please read about it, report it here, remove this message from your talk page, and then make the edit again.
    For help, take a look at the introduction.
    The following is the log entry regarding this message: Miguel de Cervantes was changed by 173.12.28.82 (u) (t) ANN scored at 0.951084 on 2013-11-12T15:43:31+00:00 . Thank you. ClueBot NG (talk) 7:43 am, 12 November 2013, Tuesday (1 year, 3 months, 15 days ago) (UTC−8)

Please refrain from making nonconstructive edits to Wikipedia, as you did at Miguel de Cervantes with this edit. Your edits appear to constitute vandalism and have been reverted or removed. If you would like to experiment, please use the sandbox. Administrators have the ability to block users from editing if they repeatedly engage in vandalism. Thank you. Wikipelli Talk 7:47 am, 12 November 2013, Tuesday (1 year, 3 months, 15 days ago) (UTC−8)

October 2014

Hello, and welcome to Wikipedia. This is a message letting you know that one or more of your recent edits to Separation of powers under the United States Constitution has been undone by an automated computer program called ClueBot NG.

    ClueBot NG makes very few mistakes, but it does happen. If you believe the change you made was constructive, please read about it, report it here, remove this message from your talk page, and then make the edit again.
    For help, take a look at the introduction.
    The following is the log entry regarding this message: Separation of powers under the United States Constitution was changed by 173.12.28.82 (u) (t) ANN scored at 0.954534 on 2014-10-21T14:44:58+00:00 . Thank you. ClueBot NG (talk) 7:45 am, 21 October 2014, Tuesday (4 months, 7 days ago) (UTC−7)

December 2014

Hello, and welcome to Wikipedia. This is a message letting you know that one or more of your recent edits to The Lovely Bones has been undone by an automated computer program called ClueBot NG.

    ClueBot NG makes very few mistakes, but it does happen. If you believe the change you made was constructive, please read about it, report it here, remove this message from your talk page, and then make the edit again.
    For help, take a look at the introduction.
    The following is the log entry regarding this message: The Lovely Bones was changed by 173.12.28.82 (u) (t) ANN scored at 0.99234 on 2014-12-10T18:28:40+00:00 . Thank you. ClueBot NG (talk) 10:28 am, 10 December 2014, Wednesday (2 months, 18 days ago) (UTC−8)

December 2014

Hello, and welcome to Wikipedia. This is a message letting you know that one or more of your recent edits to T.I. has been undone by an automated computer program called ClueBot NG.

    ClueBot NG makes very few mistakes, but it does happen. If you believe the change you made was constructive, please read about it, report it here, remove this message from your talk page, and then make the edit again.
    For help, take a look at the introduction.
    The following is the log entry regarding this message: T.I. was changed by 173.12.28.82 (u) (t) ANN scored at 0.868341 on 2014-12-15T17:19:10+00:00 . Thank you. ClueBot NG (talk) 9:19 am, 15 December 2014, Monday (2 months, 13 days ago) (UTC−8)

January 2015

Please refrain from making nonconstructive edits to Wikipedia, as you did at Cornhole with this edit. Your edits appear to constitute vandalism and have been reverted or removed. If you would like to experiment, please use the sandbox. Administrators have the ability to block users from editing if they repeatedly engage in vandalism. Thank you. Lugia2453 (talk) 8:21 am, 6 January 2015, Tuesday (1 month, 21 days ago) (UTC−8)

    If this is a shared IP address, and you did not make the edits, consider creating an account for yourself so you can avoid further irrelevant notices.

February 2015

Hello, and welcome to Wikipedia. This is a message letting you know that one or more of your recent edits to Wikipediocracy has been undone by an automated computer program called ClueBot NG.

    ClueBot NG makes very few mistakes, but it does happen. If you believe the change you made was constructive, please read about it, report it here, remove this message from your talk page, and then make the edit again.
    For help, take a look at the introduction.
    The following is the log entry regarding this message: Wikipediocracy was changed by 173.12.28.82 (u) (t) ANN scored at 0.874487 on 2015-02-26T13:39:47+00:00 . Thank you. ClueBot NG (talk) 5:39 am, Today (UTC−8)

JCM · Unread post by **JCM** » Fri Feb 27, 2015 2:12 am

Part of the problem here is I think the likely fact that these edits seem to have been to, as has been said, infrequently watched articles, and they seem to have been made in a way which doesn't scream "vandalism." I'm personally not sure that it actually proves anything about how frequently and quickly "most" vandalism is reported and reverted, because, unfortunately, I tend to think the majority of real vandalism is of the blanking the page or other forms of obvious vandalism that ClueBot and editors look for. There is no way to actually test the accuracy of the "most" statement without a clearer idea of what vandalism is really done, and I don't think that an individual who has personally stated he is performing a few, only a dozen or so, probably comparatively instances of low-level vandalism is necessarily going to give information that can reasonably be generalized.

I know that there are, basically, a large number of bugfuck extremists of all kinds, including devotees of fringe science or fringe religion or pseudohistory and suchlike, and, honestly, it stands to reason that most of them will be likely to vandalize comparatively high-visibility articles in predictable ways, like for instance saying chiropractic is not a pseudoscience and has received medical respect, or that God was using a 24-hour stopwatch when he was creating the universe in 7 days, or stuff like that.

Popular perception, and I think probably the records of other high traffic websites of this sort, would seem to support the idea that most vandals is, well, dumb, and make fairly obvious counterproductive edits to pages or topics which tend to get a lot of such activity. Finding their vandalization isn't hard. Unfortunately, there's no way to tell how many instances of vandalizing happen in the less frequently viewed topics, and on that basis no way to say how much is really "vandalism." Hell, I could misquote the old Hastings encyclopedias and add a fraudulent citation to them, and I doubt anyone would find it for years. Particularly if it is sourcing something which isn't that objectionable or obviously wrong. But how often does that happen? Well, I very seriously doubt anyone has ever done a study to determine that, and such limited experiments as this one probably won't meet many standards of significance. But could this sort of thing be done regularly, particularly mis-citation vandalism? Hell yes, and I think it actually is done frequently, particularly on the more contested topics.

Notvelty · Unread post by **Notvelty** » Fri Feb 27, 2015 2:41 am

JCM wrote: snip

You seem to be under the impression that Greg thinks the problem is "people damaging the pretend encyclopaedia" or, to put it in insipid wikipedian-speak "vandalising the 'pedia".

It's not. The problem is not people putting smutty pictures on some collection of blog posts. The problem is bullshit being presented as fact - vandalising knowledge.

Johnny Au · Unread post by **Johnny Au** » Fri Feb 27, 2015 2:59 am

Zoloft wrote: I assume this isn't yoursI call it ComCast vs ClueBot:

Code: Select all

November 2013

Hello, and welcome to Wikipedia. This is a message letting you know that one of your recent edits to Miguel de Cervantes has been undone by an automated computer program called ClueBot NG.

    Lick my nuts mr. post man ClueBot NG makes very few mistakes, but it does happen. If you believe the change you made was constructive, please read about it, report it here, remove this message from your talk page, and then make the edit again.
    For help, take a look at the introduction.
    The following is the log entry regarding this message: Miguel de Cervantes was changed by 173.12.28.82 (u) (t) ANN scored at 0.951084 on 2013-11-12T15:43:31+00:00 . Thank you. ClueBot NG (talk) 7:43 am, 12 November 2013, Tuesday (1 year, 3 months, 15 days ago) (UTC−8)

Please refrain from making nonconstructive edits to Wikipedia, as you did at Miguel de Cervantes with this edit. Your edits appear to constitute vandalism and have been reverted or removed. If you would like to experiment, please use the sandbox. Administrators have the ability to block users from editing if they repeatedly engage in vandalism. Thank you. Wikipelli Talk 7:47 am, 12 November 2013, Tuesday (1 year, 3 months, 15 days ago) (UTC−8)

October 2014

Hello, and welcome to Wikipedia. This is a message letting you know that one or more of your recent edits to Separation of powers under the United States Constitution has been undone by an automated computer program called ClueBot NG.

    ClueBot NG makes very few mistakes, but it does happen. If you believe the change you made was constructive, please read about it, report it here, remove this message from your talk page, and then make the edit again.
    For help, take a look at the introduction.
    The following is the log entry regarding this message: Separation of powers under the United States Constitution was changed by 173.12.28.82 (u) (t) ANN scored at 0.954534 on 2014-10-21T14:44:58+00:00 . Thank you. ClueBot NG (talk) 7:45 am, 21 October 2014, Tuesday (4 months, 7 days ago) (UTC−7)

December 2014

Hello, and welcome to Wikipedia. This is a message letting you know that one or more of your recent edits to The Lovely Bones has been undone by an automated computer program called ClueBot NG.

    ClueBot NG makes very few mistakes, but it does happen. If you believe the change you made was constructive, please read about it, report it here, remove this message from your talk page, and then make the edit again.
    For help, take a look at the introduction.
    The following is the log entry regarding this message: The Lovely Bones was changed by 173.12.28.82 (u) (t) ANN scored at 0.99234 on 2014-12-10T18:28:40+00:00 . Thank you. ClueBot NG (talk) 10:28 am, 10 December 2014, Wednesday (2 months, 18 days ago) (UTC−8)

December 2014

Hello, and welcome to Wikipedia. This is a message letting you know that one or more of your recent edits to T.I. has been undone by an automated computer program called ClueBot NG.

    ClueBot NG makes very few mistakes, but it does happen. If you believe the change you made was constructive, please read about it, report it here, remove this message from your talk page, and then make the edit again.
    For help, take a look at the introduction.
    The following is the log entry regarding this message: T.I. was changed by 173.12.28.82 (u) (t) ANN scored at 0.868341 on 2014-12-15T17:19:10+00:00 . Thank you. ClueBot NG (talk) 9:19 am, 15 December 2014, Monday (2 months, 13 days ago) (UTC−8)

January 2015

Please refrain from making nonconstructive edits to Wikipedia, as you did at Cornhole with this edit. Your edits appear to constitute vandalism and have been reverted or removed. If you would like to experiment, please use the sandbox. Administrators have the ability to block users from editing if they repeatedly engage in vandalism. Thank you. Lugia2453 (talk) 8:21 am, 6 January 2015, Tuesday (1 month, 21 days ago) (UTC−8)

    If this is a shared IP address, and you did not make the edits, consider creating an account for yourself so you can avoid further irrelevant notices.

February 2015

Hello, and welcome to Wikipedia. This is a message letting you know that one or more of your recent edits to Wikipediocracy has been undone by an automated computer program called ClueBot NG.

    ClueBot NG makes very few mistakes, but it does happen. If you believe the change you made was constructive, please read about it, report it here, remove this message from your talk page, and then make the edit again.
    For help, take a look at the introduction.
    The following is the log entry regarding this message: Wikipediocracy was changed by 173.12.28.82 (u) (t) ANN scored at 0.874487 on 2015-02-26T13:39:47+00:00 . Thank you. ClueBot NG (talk) 5:39 am, Today (UTC−8)

That user has been blocked for three months.

Unread post by **Zoloft** » Fri Feb 27, 2015 4:10 am

Johnny Au wrote:
Zoloft wrote:<snip>
That user has been blocked for three months.

Allow me a slight grammar change for clarity:
That user has been blocked today for three months.

thekohser · Unread post by **thekohser** » Fri Feb 27, 2015 4:23 am

Zoloft wrote:I assume this isn't yoursI call it ComCast vs ClueBot:

You assume correctly. C'mon, I'm better than this, aren't I?

However, it does look like the IP is used by a customer of "mine"... probably a restaurant or other retailer with some customers of their own, amusing themselves.