Wikipedia editor Calamondin12 knew this looked bad.
The article about "Jack Robichaux" had been flagged as an orphan, meaning it had few or no inbound links. What's more, the article featured a racist comment, and the only source provided did not mention anyone by that name. Worst of all, the article had been on Wikipedia for more than ten years.
This was the article in its entirety:
Jack Robichaux was a serial rapist in the 19th century, who plagued the township of New Orleans. Most of his victims were overweight females. He was a Creole, although police initially suspected that the assailant was black by his choice of victims. His talents as a jazz musician were praised throughout New Orleans, until his crimes became public knowledge.
Christopher Waldrep & Donald G. Nieman (2001). Local matters : race, crime, and justice in the nineteenth-century South. University of Georgia Press. ISBN 0-8203-2247-4.
"This appears to be a hoax," Calamondin12 wrote on August 27. "In light of the article's claims, it's worrisome that this would have been able to evade scrutiny for so long."
Calamondin12 had stumbled on the longest-running hoax in Wikipedia history, lasting 10 years and 1 month by the time it was deleted on September 3, 2015.
It was, of course, another piece of evidence that Wikipedia is fallible. Wikipedia's list of its own longest-running hoaxes shows hundreds that have lasted more than a year. At the same time, the latest victory shows how much better editors have gotten at spotting hoaxes, both new and old.
This is the story of how the hoax was created and how Wikipedia's editors finally brought it down.
Frat boys beat the free encyclopedia
Pranking Wikipedia was harder than expected.
Bill Maas and Van Robichaux, best friends and fraternity brothers at Washington University in St Louis, had tried on multiple occasions to edit random articles, but they kept getting caught. "We really just wanted to get fake information on Wikipedia, almost as a challenge. We tried on many different articles, but always failed," Maas said in an email.
For at least two years, Maas and friends edited the page for professional football player Bill Maas, trying to add a line about how he shared a name with college football player Bill Maas. Their changes would always get deleted.
"We were actually very impressed at how quickly any edits we would make, even if they seemed innocuous or even real, would get changed back," Maas wrote. "I was actually prohibited from making edits a couple of times due to this."
They had more luck when they created an entirely new page.
On July 31, 2005, Maas and another friend created the "Jack Robichaux" page. The name was a reference to Van Robichaux, full name John Vance Robichaux III, who often went by Jack. The page was nothing more than a header and a single sentence: “Jack Robichaux: He was a serial rapist in the 19th century, plaguing the township of New Orleans.”
A few days later, one of the pranksters added the rest: "Most of his victims were large females. He was a Creole, not a black, as many suspected by his choice of victims. His famous jazz talents were praised throughout New Orleans, until it was revealed that he was a criminal."
As Maas later explained: "That joke was a reference to racist cops and detectives making decisions that would ruin an investigation due to bigoted assumptions. I assumed cops were even worse back then."
When the article didn't get deleted right away, the guys decided to do all they could to preserve it.
" Van liked having the fake article online, so he tried to figure out a way for it to no longer be a 'stub' so we needed a source," Maas said. On Wikipedia, stubs are defined as "articles too short to provide encyclopedic coverage of a subject."
University of Georgia Press
Van, originally from New Orleans, found an old history book in a local library. "We looked up the book online, and there wasn't any posted quotes or information about it anywhere," Maas said. "We knew we could 'quote' from it without it being able to be checked."
So Maas added the source: "Local matters: race, crime, and justice in the nineteenth-century South. Edited by Christopher Waldrep & Donald G. Nieman. 259 pages. University of Georgia Press (March 1, 2001). ISBN 0820322474."
And so the page stood for a year and a half, with only minor changes — adding tags and the like — from a few hapless Wikipedia editors.
It came to life again in February 2007 when someone added something about Robichaux being a star football player at Texas Tech. That change was deleted two days later, flagged as vandalism to a legitimate article.
In May 2007, the name of the invented serial rapist was changed to Tim Bulger and then to Tom Poccia — another friend of Maas' from Theta Xi. These acts of "vandalism" were reverted within days.
"Then, we essentially forgot about it," Maas said.
Somehow the hoax had passed the initial checks against misinformation, and now it looked like it might last forever.
Hoax-hunters strike back
Even ten years later, however, Wikipedia's collaborative truth filter was running.
The article on Jack Robichaux had been flagged in February 2013 as an orphan. Eventually, this designation led it to be reviewed by one of Wikipedia's unofficial hoax hunters.
"I'm part of the hoax-hunting squad on Wikipedia," Calamondin12's user page brags. "Among the multiple hoaxes I have participated in uncovering are the following: [Bodhi stones, Collins Slip, and Operation Pax Romana]."
Calamondin12 had no doubt after seeing "Jack Robichaux" that it was a hoax. On August 27, 2015, at 4:11 p.m. the user flagged the post: “long-lived hoax; no references found for this case; not found in book cited; not to be confused with actual jazz musician John Robichaux .”
That same afternoon, user Swister Twister sat down at his computer and set about his usual routine: Scan categories of articles for any suspicious activity. When checking the "suspected hoaxes" section, he was excited to see the newly flagged post from Calamondin12.
SwisterTwister did some searching of his own and also turned up no evidence of a real-life Jack Robichaux. At 5 pm he posted his verdict: "All signs suggest this is fabricated with my searches finding nothing but mirrors and no connection at all with the book. What's more is this sparsely edited article has existed since July 2005 when it was started by an IP from Los Angeles who also made a few edits to Theta Xi and not only are the majority of editors IPs, there hasn't even been much change since 2005 which is another serious sign. Yet another interesting tidbit is that the article has never gotten Louisiana attention and is orphaned from any other articles."
SwisterTwister also alerted administrative editor DGG.
Forty minutes later, Calamondin12 posted further observations, concluding: "The article has also been extensively vandalized over the years. It's possible that the hoax originated as some sort of fraternity prank. Whatever the origin, it needs to go."
Four more users chimed in over the next few days. DGG finally deleted the post on September 3, shuttering the longest-running hoax in Wikipedia history after ten years and one month.
Wikipedia editors we spoke to believe the free encyclopedia is better than ever at fighting misinformation.
"Our vandalism rate is very low," said one senior editor, who wished to remain anonymous for privacy reasons. "It didn’t used to be, but it is very very low ... because we have people watching."
Wikipedia's English contribution has stabilized at more than 3,000 very active editors (making more than 100 edits) every month, a number that has tripled since 2005. The number of active editors (making more than 5 edits) every month hovers around 30,000, a six-fold increase from 2005.
Back in 2005, it was simply easier for posts to fall through the cracks. But today there are editors watching feeds of every edit that gets made and every new page that's created, as well as hoax-hunters who go out looking for them.
Calamondin12's user page describes a method for finding hoaxes: "Usually, I search under the orphaned articles category (articles that have no inbound links from other Wikipedia articles) and start randomly looking at articles. This category includes tens of thousands of articles, including many that are non-notable and a few outright hoaxes. Because these articles don't have inbound links, few users ever see them, and they can often escape scrutiny indefinitely. Not all hoaxes are orphans (in a few cases, a crafty hoaxer has added a link in another article), but the great majority of hoaxes fall into this classification."
SwisterTwister described a similar method over email: "I'm actually one of the many users who avidly search for said fabricated articles using the page 'Special:RandomInCategory.' The 'RandomInCategory' allows searching for articles in categories which have been tagged for issues [such as orphans]."
About Robichaux, he added: "It was obviously fabricated."
Though it's clearly possible to track down the originators of many Wikipedia edits, finding the perpetrator isn't what matters to the hoax hunters of Wikipedia. For users like SwisterTwister, the real issue at hand is "the threats [fake entries] have to other accurate and factual articles." Keeping the integrity of Wikipedia at a high standard is the true purpose.
With the case of Jack Robichaux officially considered closed, the hunt is to find more falsehoods among Wikipedia's millions of pages.