Anyone, whether directly involved or not, may add evidence to this page. Create your own section and do not edit in anybody else's section. Please limit your main evidence to a maximum 1000 words and 100 diffs and keep responses to other evidence as short as possible. A short, concise presentation will be more effective; posting evidence longer than 1000 words will not help you make your point. Over-long evidence that is not exceptionally easy to understand (like tables) will be trimmed to size or, in extreme cases, simply removed by the Clerks without warning - this could result in your important points being lost, so don't let it happen. Stay focused on the issues raised in the initial statements and on diffs which illustrate relevant behavior.

It is extremely important that you use the prescribed format. Submitted evidence should include a link to the actual page diff in question, or to a short page section; links to the page itself are insufficient. Never link to a page history, an editor's contributions, or a log for all actions of an editor (as those will have changed by the time people click on your links), although a link to a log for a specific article or a specific block log can be useful. Please make sure any page section links are permanent. See simple diff and link guide.

This page is not for general discussion - for that, see the talk page. If you think another editor's evidence is a misrepresentation of the facts, cite the evidence and explain how it is incorrect within your own section. Please do not try to re-factor the page or remove evidence presented by others. If something is put in the wrong place, leave it for the Arbitrators or Clerks to move.

Arbitrators may analyze evidence and other assertions at /Workshop. /Workshop provides for comment by parties and others as well as Arbitrators. After arriving at proposed principles, findings of fact or remedies, Arbitrators vote at /Proposed decision. Only Arbitrators may edit /Proposed decision.

Summary of checkuser findings

1. For the period of time encompassed by the available checkuser logs, Samiharris (talk · contribs) has edited exclusively from open proxies used by proxify.com.
2. According to user agent information recorded in the checkuser logs, Mantanmoreland (talk · contribs) and SamiHarris use very different computer setups. The value of this observation as evidence of anything is limited by the fact that user agents can be spoofed, and that proxify.com offers paid subscribers the option of substituting their own user agent with a different valid user agent.
3. The checkusers have investigated a suspicion that the SamiHarris account was set up by Wordbomb to falsely implicate Mantanmoreland in additional (post-Lastexit) sockpuppetry. The basis for this suspicion is that Wordbomb has also occasionally used proxify.com proxies [1] [2]. Due to the nature of proxy editing, it is unlikely that this suspicion can ever be proved or disproved.

I have personally seen no persuasive evidence

Because there has been unseemly and false speculation in some quarters that I know this (or related claims) to be true, and that I have admitted as such in private forums, it is important for me to state what I know and what I don't know.

Claims about Mantanmoreland being author Gary Weiss have been floating around for a long time. Various claims of "proof" have been made, none of which I have found convincing. At times I have believed one way, at times I have believed another way. I have investigated the claims to the best of my ability and I have been unable to find proof one way or the other.

An email I sent to Mantanmoreland and others has been widely quoted as evidence that I supposedly "know" this claim to be true. Such interpretations are malarky, and most of the people making the claims appear to me to be acting in bad faith. What I said, at a point in time, was that I believed it to be true that Mantanmoreland == Gary Weiss. This was specifically in the context of a conversation in which I was trying to get more evidence... a proof, one way or the other. Me believing at a point in time in an investigation that something was true, is not the same thing as an assertion that it is true, nor of an "admission" or anything else.

Mantanmoreland steadfastly denies being Gary Weiss. Ask him yourself if you want to know.

Related allegations that I am protecting a "friend" are nonsense. Mantanmoreland and I do not get along well at all.

Related allegations that I have some vested interest in the underlying content dispute are even worse nonsense. I have no opinion about "naked short selling". I have never sold a stock short in my life. I have no financial interests of any kind in this case. If you read anything otherwise, or hints to that effect, on the overstock.com blog or elsewhere, well, I don't know was else to say but: nonsense. I think such allegations tell more about the people who are making them than anything else.

Regarding the specific claim at issue here, whether Sami Harris and Mantanmoreland are the same user, I can say quite firmly that I do not believe it to be true. I have interacted (argued!) with both users over an extended period of time by private email, and I have not seen any reason to think it true. The offsite "evidence" relating to this comes from a highly questionable source, and furthermore strikes me as completely unpersuasive. For all we know, these are faked screenshots from someone who has engaged in a campaign of harassment and bad behavior (on-wiki and off-wiki) that has been really astounding to witness.

I have reviewed my email archives to look for similarities between the users. I have examined email headers. I have looked for textual similarities, time patterns, etc. I see nothing to lead me to a conclusion that Sami Harris and Mantanmoreland are the same user.

For these reasons, I do not believe it to be true that Mantanmoreland == Samiharris. --Jimbo Wales (talk) 02:19, 15 February 2008 (UTC)[reply]

Sockpuppetry patterns

This is a slight update to evidence presented at the RfC.

It's important to remember that Mantanmoreland has abusively used at least two sockpuppet accounts in the past: Lastexit and Tomstoner. He never admitted to this, nor did he apologize, but the accounts fell into disuse after User:Fred Bauder warned him.[3] With that in mind, consider these facts:

• Samiharris always used a proxy.
• Mantanmoreland and Samiharris shared a POV, and were especially prolific on articles related to Naked short selling, Overstock.com, and Gary Weiss.
• They edited dozens of the same articles at similar times and to similar ends, and !voted on several issues together.
• Mantanmoreland and Samiharris edit at precisely the same times of day. (see chart, details)
  • These accounts have better correlation coefficients with each other's editing patterns than any pair of 21 other accounts studied.
• In spite of this fact, they have only interleaved edits within 30 minutes of each other five times over the last year (details), which is much much less overlap than with two randomly-selected accounts that made a comparable number of edits.
  • The closest edit was within 3 minutes. MM SH copying text (MM wrote this text the previous hour)
  • If their edits were randomly distributed over the year, there would be less than 1.5% chance that not one pair of edits would occur during the same minute. Considering the non-uniform and similar time distribution for both accounts, the true chance that this would occur randomly is much less.
  • Their contributions dovetail. That is: there are long strings of one user editing, followed by another editing, with no intersection.
• Both accounts have used uncommon expressions such as "asked and answered many times," and "lipstick on a pig"—in fact it appears the latter expression has only occurred 20 times by 14 users in talk space, and two of those instances were from Mantanmoreland and Samiharris.
• A lengthy catalog has been compiled of editing traits which these two accounts share. Of particular note are edit summaries using the strings " -- ", "rply", "duplicative", "Talk" (capitalized referring to talk page), and "as per" (instead of simply "per"). These traits are seen in several edit summaries by both of these accounts. Mantanmoreland's previous socks also shared many of these traits. However, these traits are uncommon among editors at large, based on a sample of Wikipedia editors.
  • " -- " (space, hyphen, hyphen, space) is particularly notable because it was also often used by both of Mantanmoraland's previous socks, and because there are many viable alternatives, and it seems less likely that the style would be "learned" like "rply" or "lipstick on a pig. Moreover, it's a A comparison with 14 other established editors showed that few other accounts ever used it in their edit summary, and out of all 18, the four most prolific users of this precise string were Samiharris, Mantanmoreland, and his two socks (see summary table).

Given Mantanmoreland's history of abuse, and given that these accounts shared interests, ideologies, "phraseologies," editing traits, and hours of operation, this is an easy case.

These are sockpuppets.

Some users complain that there's no statistical analysis, but that's never been required in a duck test as far as I'm aware. The comparisons I've made to other user are only to increase confidence that the traits and patterns highlighted really are uncommon, and that their simultaneous reincarnation in a single user is unlikely given (1) the user's history of sockpuppet abuse, (2) both account's shared POV, and (3) Mantanmoreland's motive to shunt edits with the "appearance of COI" off to another account.

I'm willing to do anything to provide more analysis, and I've continued to elaborate on my findings. It seems, however, that no amount of circumstantial evidence will convince some users. If that's this Committee's opinion, then I'll quit wasting my time.

If, however, someone could be persuaded by a test of their choosing, I'll do anything within my means to conduct that test.

Lastly, Jimbo and others claim that the users are stylistically quite different. They refer to prolific email conversations with both users. I think the evidence in this case should be confined to the site and its official appendages, but if Jimbo or others could articulate any specific stylistic differences at all, I'll look into it with all my might, as I've already offered. Cool Hand Luke 05:17, 15 February 2008 (UTC)[reply]

Conflict of interests

Mantanmoreland appears to be a certain financial writer that I will call "W" for convenience.

It has been pointed out that Mantanmoreland edited a fairly obscure place called Varkala in India. Four months later, the writer W announced his vacation and marriage there. Whitstable suggested comparing edits in this vacation period for evidence of decline over the period W was in India. There's a slight decline, but what's really remarkable is a shift in editing patterns. Neither Samiharris nor Mantanmoreland has edited between 8:00 and 10:00 (3 AM EST, to 5 AM EMT) over the last year, but for slightly over a month, Mantanmoreland's time zones radically shifted. This causes me to conclude that user was in fact editing from the other side of the world at the precisely the same time W was visiting.

In conjunction with Mantanmoreland's interest in Varkala, I think it's safe to assume that the user is W. That said, has there been COI abuse? Yes.

• Edits to mainspace Gary Weiss
  • Mantanmoreland, 26 (before May 15, 2007, when Mantanmoreland quit editing this article)
  • Samiharris, 18 (beginning May 14, 2007, when Samiharris started editing the article)
  • Lastexit, 7
• Edits to mainspace Judd Bagley, who W has an adversarial relationship with
  • Samiharris [4] [5] (revert warring with User:Phil Sandifer)
  • Mantanmoreland [6] [7] [8] (continuing revert war in apparent violation of 3RR by using sockpuppets)
  • See also the Judd Bagley deletion debate, where accounts appear to have double-voted.
• Edit to Julian Robertson, where W is the rival in a real-world legal struggle. Here W appears to add his own source in the mix as Lastexit.[9]
• Other diffs where Mantanmoreland or his old socks used Weiss (often Weiss' Overstock.com-focused blog) as a source in the mainspace: Pump and dump Pump and dump II Arthur Levitt Hedge fund Julian Robertson (also a potential personal life COI) Naked short selling I Naked short selling II CUNY Richard Grasso Incidentally, one of Samiharris first edits argued for retaining the naked short sources added by Mantanmoreland.
• See also Relata Refero's conclusions of long-term tendentious editing below.

Respectfully submitted. Cool Hand Luke 18:30, 15 February 2008 (UTC)[reply]

Reply to Smallbones

User:Smallbones complains that comparisons to other users are meaningless due to small sample size and particularly for unannounced methods. I hope to address this imagined shortcoming.

User:Alanyst has compiled the edit history of every user who made between 1000 and 2000 edits in 2007, constituting 3629 accounts. See Alanyst's section below. Building upon Alanyst, who did the hard work, I've begun to use this data to extend my previous comparisons. These are the same methods I've used, and I don't know how they'll compare to the whole group. This is also practically the largest sample size possible, so I think that decisively deals with Smallbone's objections.

Of these 3629 accounts, the editing patterns of Mantanmoreland and Samiharris are each other's 10th and 15th best match, respectively—out of 3629 active accounts, mind you. At this point, I believe we can say with some reasonable certainty that their editing patterns are a very good match. See details. (There's a caveat that it's unclear what good correlations actually mean. I posit that they mean that the underlying editor(s) have similar lifestyles and behavioral patterns that tends to make them edit at similar times of day. This is a therefore a proxy for offline lifestyle.)

Incidentally, all but 1 of 31 other accounts correlating best with these two users has at least one editing collision (that is, an edit which occurred with in the same minute). Recall that these two accounts only have a single edit within three minutes and only five examples of back-and-forth editing occurring within 30 minutes. As Alanyst details, 9.5-16.8% of all accounts with 1000-2000 edits don't have same-minute collisions with Mantanmoreland and Samiharris respectively, but accounts with superior correlation tend to have more collisions (which makes sense—accounts editing when Mantanmoreland is asleep have virtually no chance to collide). So once again, the non-overlap of these two accounts is curious.

Note, I may extend my previous studies to these 3629 accounts to determine (1) the number 30-minute interleaving edits, by the definition that I used for these two accounts here. Once again, this is now an established methodology, and I don't know what the results will be, so could not possibly use selection bias—consciously or unconsciously. If it takes comparing every user on the site to show that Smallbone's assumption of bad faith is misplaced on me, then I'll do it. Cool Hand Luke 07:35, 18 February 2008 (UTC)[reply]

Sockpuppetry and bad faith participation

To show that we're setting the bar for proof of sockpuppetry here much higher than we usually do, I was going to present some diffs from the Suspected Sock Puppet noticeboard. A quick look, however, showed that almost every single incident reported there was accepted and blocks were handed out with much less scrutiny and analysis than is going on here with Mantanmoreland and Samiharris. The question has to be asked, why does Mantanmoreland get such different treatment than others? The fact is, Mantanmoreland has received preferential and troubling treatment throughout the two years of this episode, including help in pushing POV in several articles and retaliating against editors who tried to resolve it.

Mailing list participation with certain, active admins

First, it appears from comments by Jimbo above and JzG and Durova below and elsewhere that Mantamoreland and Samiharris were members of at least one of the private, invitation-only mailing lists to which several of our administrators and perhaps one or two sitting arbitrators belong to [10] and [11]. I believe that list is owned and operated by SlimVirgin. I suspect that here [12] may be where SlimVirgin invited Samiharris to participate in the mailing list.

POV pushing with help at Gary Weiss

Mantanmoreland and Samiharris (and Mantanmoreland sock Lastexit in at least one instance) were very active in protecting the Gary Weiss article from containing any information, no matter how well cited from reliable sources, that might not reflect positively on the article's subject.

• Mantanmoreland: [13] [14]
• Samiharris: [15] [16] [17] [18] (this last is a paragraph that had been agreed to on the talk page) and argues that a NYTimes article can be listed but not actually linked to [19] [20]
• Lastexit: [21]

They were helped in protecting the Weiss article by several administrators.

• SlimVirgin: Protects the article [22] [23] [24] [25]. She also protects the talk page and, while protected, archives discussion [26] and then protects the archive [27] and admin deletes history from the talk page[28].
• JzG: After I initiated a content RfC on the article, he shuts down the discussion and archives it [29]. Later, he removes cited text and then fully protects the article [30]
• Dmcdevit: Appears to admin delete previous versions of the article along with the edit history [31] [32]. These edits were subsequently oversighted. Is this a violation of GFDL?
• Unknown oversight admin: In this diff Mantanmoreland appears to be aware of talk and main page comments being deleted [33]. I request that the ArbCom check the oversighted material from the main page and post the name of the admin who oversighted them.
• Jimbo: Admin deletes the Weiss AfD discussion Wikipedia:Articles for deletion/Gary Weiss. I don't think Jimbo is trying to help push POV in the Gary Weiss article with this deletion, but, I don't think that there's anything in the deletion discussion that should have been hidden. But then, perhaps after being alerted via mailing list (IMO), Jimbo comes to the article talk page after I opened the content RfC which was shut down by JzG and states, "zero tolerance, shoot on sight" [34]

I was going to also show POV pushing at Patrick M. Byrne, Naked short selling, and Overstock.com by Mantanmoreland and Samiharris but others have already presented that evidence below.

Other inappropriate admin actions

• SlimVirgin indef blocks WordBomb about two hours after he filed a request for mediation concerning content in the Gary Weiss article [35] for "libellous edits to an article; attempted to "out" another editor" [36] without any prior warning to stop the behavior [37]
• David Gerard indef blocks Piperdown as a "sockpuppet/meatpuppet for overstock.com" [38]. More discussion on the block here [39].
• David Gerard blocks an entire IP range in Utah [40], stating that it was an open proxy [41], apparently in an effort to keep Judd Bagley (WordBomb) from editing Wikipedia. When asked about it by a Utah newspaper reporter three months later, who confirmed that it wasn't an open proxy [42], he unblocked it [43]

(more to come)

Anti-Bagley/Byrne/Overstock.com vitriol and personal attacks

• Mantanmoreland edits a userbox on his userpage to say "This user has been stalked by Judd Bagley" [44]. Bagley then alters the personal attack to a message advertising his blog [45]. SlimVirgin then reverts it back to the version with the personal attack on Bagley.[46].
• SlimVirgin states that Bagley conducted "some of the worst stalking that Wikipedia has seen" [47].
• JzG calls Bagley an "obsessive troll" [48], a "net.kook", "absolutely not above forgery", and a "vicious, agenda-driven troll" [49], "Bagley's lunacy" [50], "his (Bagley's) vile smear campaigns" [51], "harassment meme inventor" [52], "long history of abuse by Bagley" [53], "paranoid fantasies of banned abusers" [54], "targets of his harassment" [55], "Bagley uses disinformation and harassment against anyone who does not uncritically support his company" [56], "Bagley is a vicious hatemonger" [57], and "Bagley's idiocy" [58], "
• Phil Sandifer says, "Overstock is an unprofitable business run by a lunatic who rants about sith lords, with a sociopathic executive who infects his critics with spyware" and "run by the criminally insane" [59], "Overstock is a money-losing company with a staggering record of despicable actions" [60], (on Wikien) "Bagley is so odious that everybody is unwilling to take him off the blacklist" [61], and "Bagley, or any other nutjob running an attack site" [62]
• Thatcher calls editors with association to Overstock.com "abusers" [63] and appears to try to plant a poison pill that Samiharris might be an abusive sockpuppet of Bagley [64]
• Georgewilliamherbert on WordBomb (and Piperdown), "They're banned because they behave sociopathically and abusively towards editors here, tracking down real names, calling their homes, their employers, their friends, trying to get them fired, urging others to stalk them in real life, threatening violence, etc." [65] and on the Wikien mailing list, "Bagley's claims are false, for reasons unrelated to him being a sociopathic, evil harrasser" [66]
• Morven referring to Bagley's revelation that SlimVirgin had abusively used a sockpuppet "Nor do I think it worthwhile to give stalkers, especially stalkers for pay, satisfaction" [67]
• JoshuaZ (on the Wikien message board), "Bagley is a complete ass" [68], "We should confuse bad faith (i.e. Judd Bagley) with good faith editors" [69], "Bagley, Brandt, Amorrow and others have engaged in real life harassment that has ruined lives" [70]
• David Gerard tells Bagley on the Wikipedia Signpost news suggestion page to "Fuck off" [71] and on the Wikien mailing list, "Bagley isn't your regular corporate spammer, he's actually notable in Reliable Sources™ for his odious stalking behaviour" [72]

(more to come)

Retaliation and personal attacks

• My RfA nomination was within three hours of closing when SlimVirgin requested that it be extended [73] stating, "There was a serious concern a few months ago that Cla68 was either a sockpuppet of banned User:Wordbomb (a very abusive sockpuppet and stalker) or was helping him. My recollection is that there was no technical evidence that Cla was Wordbomb, but he definitely seemed to be helping him, and some of his edits indicate that they're based in the same area."
• SlimVirgin then alleges in the RfA that I was a supporter, if not an outright sockpuppet or meatpuppet of WordBomb and that I posted conspiracy theories about the Gary Weiss article on Wikipedia Review [74]. I've never had an account on Wikipedia Review.
• After repeating the allegations in the RfA, SlimVirgin states, "Also, I forgot to say earlier that, judging by some of Cla's edits, he appears to be based in the same state as Wordbomb." [75]. I have resided in Japan since September 2006 and before that I did not reside in Utah, which is apparently where Bagley resides (see Gerard IP range block above).
• Mantanmoreland also gives a detailed oppose vote in the RfA accusing me of helping WordBomb and calls me a "classic troll" [76].
• A quick succession of oppose votes then follow, including FloNight [77], Crum375 [78], and Jayjg [79], among others.
• During the RfA, discussion turned to Bagley's blog, and I was threatened with a block by Crum375 for even mentioning its name (AntiSocialMedia.net) even though no policy at that or the present time prohibited doing that [80].
• At almost the same time that Crum375 threatened the block, SlimVirgin edited the blocking policy to support the threat [81]
• I subsequently opened an RfC on my involvement in the Weiss matter and the RfA discussion. I posted a notice of the RfC on the AN noticeboard [82]. SlimVirgin deleted the notice six minutes later [83]. When I asked her about that on her talk page, she immediately archived her talk page without responding (times, dates, and content of my post located here [84])
• Crum375 likens unblocking Piperdown to "If a psychopath who violated your mother and your sister, say, wanted to live with you, would you let him, until he violated your wife too?" [85]
• JzG to another editor on the Gary Weiss talkpage, "You're sure doing a lot to give the impression that you prefer your friend Mr. Bagley to my friend Mr. Wales." [86]
• David Gerard states on the Wikien mailing list when discussing my involvement with the Gary Weiss article and a block I received after Jimbo's "shoot on sight" statement, "Precis: Cla68 has been a dick about this for quite some time, knew *precisely* how much of a dick he was being, and thoroughly deserved the block, and probably a longer one. He's not here to write an encyclopedia." [87]
• Mantanmoreland taunts me about my failed RfA and makes clear that he helped the RfA to fail because I opposed him on article content and that I'm a "WordBomb supporter" [88] [89][90]

(more to come)

Naked Short Selling has seen long-term tendentious editing

As I stated here, i had not previously been involved in this issue, under this or any previous username, nor have I to my recollection ever been involved in discussion with either Mantanmoreland, Samiharris, or indeed anyone from this dispute. Nor do I have a view on the real-world dispute this reflects. Accordingly, I thought it would be helpful if I had a look at this article and the related talkpage.

• My first impression was that the article had major problems, which I summarised in the linked diff, although there were no tags to indicate that. I did not at that time consider carefully what had happened on the talkpage, merely analysing the general approach.
• I then considered, purely on the basis of the current version of the talkpage, what the mechanism was by which this unfortunate state was arrived at in a fairly high-traffic article. In each of the below cases I link to my summary of the talkpage section immediately above it.
  • Here I summarise a section of the talkpage in which User:Samiharris argues with an anonymous IP about the wording of a section on legal action by companies that believe themselves defrauded by manipulative short selling of their stock. He says "Not one has succeeded. I have added this information to the encyclopedia. Please do not remove." However, as the IP correctly points out, "not succeeding" is a mischaracterisation of the then status of those and related lawsuits. Either way, relevant legal studies were available; instead a DTCC press release was aggressively defended as the sole useful source of information.
  • This is damning as far as I'm concerned. A completely impartial observer from the good people at the Business and Economics Wikiproject drops by the page as part of their regular assessment patrol. He says the style is too journalistic and over-dependent on such quotes; that there is too much focus on the media controversy and on specific cases within that controversy; and there is absolutely no discussion of the economics behind it. User:Samiharris replies (hilarious, if these allegations are true) that "this article was written by committee" and that there is only a smattering of news articles to cite, as no independent reliable sources exist. As demonstrated at the end of the talkpage, and as anyone with ten seconds to spare can determine at SSRN and EconWeb, multiple surveys exist. Why then this emphasis on journalism, quotes and controversy, in the face of impartial external input, to the point of making what on WP we can call at most an inexplicable exaggeration?
  • [91] User:Mantanmoreland and User:Samiharris agree with each other, with no other participants in the conversation, that an article on Regulation SHO, set up by the SEC to curb naked short selling, should be replaced with a redirect to Naked Short Selling. It is called a 'blatant POVfork' in spite of the fact that an article about regulation X set up to regulate Y might more than conceivably have considerably different scope and content than an article on Y. The examples I provide in the above diff are of the obvious parallel, insider trading and the various regulations adopted to deal with that issue. I see no discernible reason why this should be different, especially since the regulation itself was highly notable in terms of dedicated coverage.
  • Samiharris removes a tag placed by another uninvolved user. The tag's removal is completely uncalled-for and again involves some creative overlooking of fairly obvious sources.
  • Finally, on the matter of a particular financial journalist's views. These come up again and again on just that section of the talkpage I see. The particular theory he espouses has received little or no attention from reliable sources, and seems to be the province of a particular kind of online bulletin board, and a few opinion columns in the financial press. My summary of a survey relevant literature is in this diff; I end with the statement "I'd like to point out this appears to be a relatively fringe-y theory, or at best something that is considerably less important than a dozen points that are not in here, and any sort of frequent pushing of it should have set off some alarm bells."

• To sum up, I have looked at a snapshot of one particular article among many involved in this dispute.
  • That article shows all the signs to me of long-term tendentious editing, something I am tragically familiar with identifying.
  • Judging by talkpage discussion, that recent tendentiousness seems to originate from User:Samiharris and User:Mantanmoreland.
  • That tendentiousness takes a particular side in what appears to be a real-world dispute over regulation policy.
  • This is not the first, nor the worst article in which real-world disputes and fringe POVs have spilled over into a Wikipedia article. This is, in my memory, however, the first where a long-term effort by the article's owners has succeeding in creating the impression of an article that is stable and non-disputed even though it appears to have received considerable attention. (This is particularly pernicious, in my opinion. When an article is problematic, that state should be telegraphed to the reader.)
  • Whether this is caused by false consensus, or by the fact that they received no opposition, or the possibility that all opposition was blocked as a presumed proxy for a particular editor who disagrees with this views in RL, is something on which I shall not express a view.

• Given this, I additionally submit that all evidence or proposals that begin with the assumption that "nowhere has it been demonstrated these accounts have injured the encyclopaedia" or that there is no evidence of editing that in and of itself has violated policy is flawed to start off with, and should be viewed with concern.

Have a nice Arbitration,

Relata refero (talk) 14:07, 15 February 2008 (UTC)[reply]

Responses

To User:Smallbones

Smallbones below makes the claim that the "statistical" evidence is rubbish. I have some considerable experience with statistical and econometric methods, and I don't think its rubbish. The size of the dataset is extremely small; a model is unspecified; and there are sundry other problems. (As someone has pointed out somewhere, what does Pearson's coefficient strictly mean here? Does a correlation in the probability of either one of these accounts editing at a particular time - because that is what it means here - actually imply something real? What is the underlying distribution of probabilities that would permit us to test the likelihood of similarities? Yes, all that is inexact.)

However, there are some basic points that are well-taken. If we assume a uniform distribution, then there is a very, very small probability that edits from two such accounts would not overlap. If we assume that the two accounts are from the same timezone, that probability is even lower. Combined with the textual analysis - phrases limited on talkspace here to fourteen out of thousands of editors, and edit summary patterns - things are more interesting.

Further, as has also been pointed out, statistics is essentially a matter of good faith in some cases. We trust SF and CHL to, when analysing MM' edit summaries, to pick out every relevant tic and compare it to SH, and vice versa. We assume they weren't data mining.

If all that is true, then I submit that Smallbones' objections go away. (Otherwise, I invite him to prove that MM and I are sockpuppets.) Relata refero (talk) 23:17, 16 February 2008 (UTC)[reply]

Further response: I note that there is no response to my point that, if we assume a very simple model of independent times of editing, uniformly distributed, then someone calculated that the probability over a year that no such edit periods overlap is around a single percentage point; obviously, if the individuals are in the same timezone, that probability is even lower. The strength of this artefact of the data is strong enough that other concerns about the degrees of freedom etc. go away.

Also, I absolutely agree that there is a tendency to not report negative results - its particularly bad in economic data work. However, in this case, I believe someone can confirm that at least CHL and SF both have said that they particularly did not find any 'tics' in one account's editing that were not duplicated to a degree in the other account's editing. Relata refero (talk) 14:11, 17 February 2008 (UTC)[reply]

To User:Mantanmoreland

I do not believe, on reflection, that User:Mantanmoreland has as yet raised any points that actually require a response. For the interested, here are the rejoinders that I had originally made, and have deleted as unnecessary. Relata refero (talk) 14:41, 17 February 2008 (UTC)[reply]

For the avoidance of doubt, and at the risk of lapsing into A-O-Hell, my interactions with Mantanmoreland have been similar to Jimbo's, in similar circumstances and in some of the same groups, as with Jimbo I can't say I get along especially well or especially badly with Mantanmoreland, I too have investigated the supposed evidence and found it wanting, I too have compared emails (I have hundreds each from both Mantanmoreland and Samiharris). I can believe that Samiharris might be Weiss, I accept Mantanmoreland's often repeated statement that he is not Gary Weiss, I do not believe that Samiharris is Mantanmoreland.

As an aside, it is my personal belief that most of the problem here is caused by one individual whose perspective is simply wrong. We know from copious external evidence that Judd Bagley has a vendetta against Gary Weiss. It is clear that Weiss and the issue of naked short selling amount to an obsession for Bagley, but I've seen no suggestion from anyone worth listening to that the reciprocal is true. This seems simply to be one of a number of issues that interest Weiss, a financial journalist of some reputation.

It is also clear that in as much as a problem exists, the problem is caused by Bagley's frustration (by the looks of it rage would be a better word) at being rebuffed from his attempts to abuse Wikipedia to promote his agenda.

By this I do not mean that sockpuppetry by Mantanmoreland or Samiharris would not be a problem, and use of open proxies is definitely a no-no, what I mean is that nobody has presented evidence of edits by Mantanmoreland which in and of themselves represent a problem per Wikipedia core policies. His edits, or such of them as I have reviewed, have been verifiable, neutrally stated and reasonable. His comments in debate, again per a sample review, are I think unproblematic: he has been civil and has usually focused on policy and content, avoiding ad-hominem despite acute provocation.

I think that SirFozzie did the right thing by digging through this. I think the result is inconclusive. In Scottish law, we would probably have a verdict of "Not proven", which in Wikipedia terms amounts to not guilty.

Assertions have been made by several people that the Wikipedia community has tried to bury this or protect Mantanmoreland, due to supposedly influential friends. These influential friends have not been identified. I don't know who they might be. I do know that the issue has been discussed well beyond what even the most jaded onlooker might consider suppression, given that each new round of discussion amounted to exactly the same thing: unproven assertions. In short: the community has investigated this, the community has found no compelling evidence to support the claims, the community cannot therefore take action, and hopefully the arbitration committee will, in its capacity of arbiter of community action, endorse this finding.

Cla68 asserts that the bar is set much higher here than for other cases of suspected sockpuppetry. I cannot recall another case of suspected sockpuppetry where the purported puppeteer has such a long history of edits, has asserted so strongly that the accounts are unrelated, and where checkuser evidence has completely failed to support the claim of sockpuppetry. It is possible (I don't do the SSP board so I don't know) that the bar is set lower where the purported puppeteer is a banned user with a long history of puppetry, but that is not the case here. Most of the socks I've blocked recently have been those of user:Jon Awbrey, and any which have been checkusered have been confirmed. In any case, most of them reference offsite solicitation to disruption by Awbrey, so would qualify as disruptive meatpuppets anyway. Perhaps Cla68 can identify a case of a long-term user in good standing who denies the puppetry but has been banned anyway, since that seems to be what he's calling for here.

If the arbitrators want to wade through about six hundred emails for purposes of textual analysis, I will be happy to forward them.But it's not off-wiki"evidence" as such, merely the basis for my opinion (and the same with Jimbo, I guess). Guy (Help!) 15:16, 15 February 2008 (UTC)[reply]

Notes to Cla68:

1. See [92] for an example of the strong (I would say incontrovertible) evidence that Bagley is a vile agenda-driven troll who would be completely unwelcome here even if Gary Weiss were the sole editor of the overstock.com article. Bagley's problem is simple: he attacks anybody who suggests that overstock's poor stock performance is down to poor management and consistent failure to return a profit. Yes, his job is to defend overstock, but his methods are completely out of line, and this should be obvious to everybody. There is no crusade against Bagley, merely an understandable unwillingness to let this exceptionally difficult individual abuse Wikipedia for his own ends. Ironically, Byrne comes across as a much less problematic character.
2. Even if SlimVirgin did own the mailing list (which I can't remember), so what? SlimVirgin is an admin in good standing, and the lists were set up to discuss a problem which was experienced by a number of editors, albeit you and Dan Tobias feel the need to pretend this problem does not exist. Jimbo thinks it does, I think it does, the victims of harassment think it does, and really there is nothing at all wrong with wanting to discuss how best to handle that. Your "evidence" on this point is therefore moot.
3. As for the rest, I recommend a career with these folks, your talent for picking cherries is undeniable. Guy (Help!) 09:42, 18 February 2008 (UTC)[reply]

Linkholder to my original investigation

[93] This is the state of my Investigation page at the beginning of this Arbitration. (the core is my work, additional evidence by many others (including Durova, Alanyst, and others is also present.) There is also a Sandbox page for everyone to work on that is linked off my investigation.

Evidence found compelling by the community

On the related Request For Comment Wikipedia:Administrators' noticeboard/Mantanmoreland/RfC, the overwhelming view of the community was that the evidence provided linked the two accounts to each other. At the time I write this, the outside view of User:Cool Hand Luke which summarized the evidence he gathered, with the conclusion that these two accounts were related, had 34 endorses.

On WP:DUCK's, Geese, and other waterfowl

I regret that such an imperfect standard has to be the core test for this evidence. Most of us value certainty, the knowledge that something IS or IS NOT the case. Unfortunately, due to one user using open proxies, that certainty is not possible.

Instead, what I tried to do is look for reasons why they could be linked, as well as why they may not. What I found was a lot of the first, and not a lot of the second. I stand by my investigation, the investigation of other users, and my personal conclusion that these two accounts are linked.

I can understand why any action based off personal conclusions and not facts is problematic. That is why we're here in front of ArbCom after all. However, in previous ArbCom cases, the standard is clear:

It is rarely possible to determine with complete certainty whether several editors from the same geographic area are sockpuppets, meat puppets, or acquaintances who happen to edit Wikipedia. In such cases, remedies may be fashioned which are based on the behavior of the user rather than their identity. Editors who edit with the same agenda and make the same types of disruptive edits may be treated as a single editor.

(see:Wikipedia:Requests_for_arbitration/Starwood#Who.27s_who)

Even if we cannot prove these two accounts are a certain-Real Life identity, that is secondary to my goal once the investigation started, and that's to determine if these accounts are linked.

On-Wikipedia vs Off-Wikipedia regarding evidence

One thing I note with concern, is that one "side"'s off-Wikipedia evidence is routinely dismissed and castigated, but now a lot of off-Wikipedia evidence is being offered by the other "side" in this discussion.

When this whole thing started, myself and Durova limited ourselves to on-Wikipedia evidence only specifically because of the contentiousness of some specific off-Wikipedia evidence, and the tactics used to gather said evidence. Durova said it best in an essay:

If something's wrong and it's not getting fixed, please be patient and keep working on fixing it the right way. If you let your own standards drop because you get frustrated, people will go ewwww and walk away. Then it'll take even longer to get your problem solved. That's not a happy place to be.

I have tried my hardest not to slip into off-Wikipedia evidence, to avoid that "eewwwwww". I went so far as to redact a comment that I had made that contradicted the way someone viewed this case, because that comment to me was based in a personal email that I was asked to keep confidential.

However, while I am perfectly cognizant of the private nature of the information being offered to ArbCom to support the viewpoint that these two accounts are not linked, I cannot help but be disappointed. We have attempted to be as open as possible, to get everything we know, and somethings we suspect on the table. In otherwords, we've been open and honest, and hope that Sunshine, indeed is the best disenfectant.

In poker terms, we've set our hand on the table, and we have two pair (not an unbeatable hand, but a pretty good one). Our opponent is refusing to tell us what he actually has, but is telling us his hand beats two pair.

(this is just my thoughts, I speak for no-one but myself on this)

Mantanmoreland, Lastexit, Tomstoner

• 11 December 2005: Tomstoner's account created. logs
• 28 January 2006: Mantanmoreland's account created. log Mantanmoreland's first edit, and 16 of his first 20, are to Naked short selling.
• 8 February 2006: Tomstoner's contributions begin. The first seven edits are to articles about Indian cities or people.
• 11 February 2006: Tomstoner's eighth and ninth edits are to Naked short selling. His tenth edit is to support Mantanmoreland on the talk page.
• 19 February 2006: Tomstoner's eleventh edit is to nominate Patrick M. Byrne for AFD. The majority of the remaining edits are to pages related to the Weiss/Naked short selling/Bagley/Overstock.com dispute.
• 13 March 2006: Mantanmoreland adds massively revises in the first person a post in which Tomstoner is explaining Tomstoner's actions.
• 22 April 2006: Lastexit's account created. logs Lastexit's contributions begin. Within 2 hours, contributing to Short (finance). For the remainder of his editing history, the majority of edits are to pages related to the to the Weiss/Naked short selling/Bagley/Overstock.com dispute.
• 13 May 2006: Lastexit makes an AFD nomination. Less than 3 hours later, Mantanmoreland votes delete in the same AFD. Mantanmoreland had not edited (admin only) the article, participated in no other AFDs on the 13 May 2006 log, and had previously only participated in one AFD.
• 7 July 2006: Diff of abusive use of Lastexit that Fred included in the warning on 23 July.
• 22 July 2006: Tomstoner's last edit. [94]
• 23 July 2006: Fred Bauer's warning to Mantammoreland about the abuse of Lastexit.
• 23 July 2006 1.5 hours later, Mantanmoreland archives all threads on his talk page, including Fred's warning. Subsequent edits to his talk page and talk archive make it clear that Mantanmoreland is aware of the warning. Conversation continues on Mantanmoreland's talk page; though it appears that part of the history has been oversighted due to claims about Mantanmoreland's real world identity that were included. Fred reiterates the warning. Fred has later said that this was checkuser confirmed sockpuppetry.
• 24-25 July 2006: IP editor(s) and Lastexit edit war over using Lastexit's edit page to label as a sockpuppet of Mantanmoreland.
• 29 July 2006: Last edit by Lastexit. [95]
• My review of the history of these three accounts indicates to me that they clearly meet the duck test; even absent Fred's warning.

Impact on Wikipedia and NPOV

I see that Relata Refero addressed in detail the same issue I looked into about what abuse or damage has occurred. There are several other articles that I think should be considered, however, while I frankly find it hard to believe anyone who has read either Overstock.com or Patrick M. Byrne has not seen the problem. In any event, here are a few examples from those and elsewhere that should raise concerns. [96][97][98][99][100][101][102][103][104]

Note that in the first link, Samiharris adds Overstock as an example to Smear campaigns (the same was done in Astroturfing, a similar issue[105]). The next link in particular seems to pursue disputes outside the scope of Wikipedia. Each of the others seems to suffer from the same types of problems. A fair hearing is certainly important here, as anywhere, but I think these should be considered along with the rest of the evidence.

Other implications

Finally, since people are just now beginning to discuss it, I think the community will need to reexamine the impact of the way WordBomb has been greeted on Wikipedia, first by Mantanmoreland and a now confirmed sock, Lastexit.[106] To say the evidence here changes nothing, or that WordBomb was the only editor with an agenda, or that his actions were uniquely inappropriate, are all at this point unsupported, as are the more recent claims that allegedly serving a corporation somehow places WordBomb beyond the pale (with respect, it seems he’s alternatively obsessed or just doing his job). Clearly both sides are interested in an off-Wikipedia dispute, both sides have pursued their position on Wikipedia with questionable means, and really we shouldn't be surprised by either. For that matter, I have no problem with saying the editors are too involved to edit this portion of Wikipedia, or with applying blocks for specific reasons. However, I believe a reevaluation either here or in the near future of the hard positions that have formed on this larger issue is clearly necessary to restore some sensibility to the matter. Mackan79 (talk) 22:56, 15 February 2008 (UTC)[reply]

Background of WordBomb and Mantanmoreland's dispute on Wikipedia

Based on some of the current proposals implicating the larger context here (particularly as regard mitigation) I think it is necessary to inquire further into the background of WordBomb and Mantanmoreland’s dispute as evident on Wikipedia.

I believe we all know the long-standing conventional wisdom on this, that WordBomb arrived on Wikipedia with an agenda, attempted to damage a BLP and harass an editor, persisted in doing so upon warnings, and was then appropriately banned before proceeding to justify the ban in numerous other ways. Presumably this would be seen as mitigating evidence for Mantanmoreland.

However, a number of problems with this can now be noted. First, by the time WordBomb arrived in July of 2006, Mantanmoreland and Lastexit had been editing Gary Weiss,[107] Lastexit and Tomstoner had been editing Patrick M. Byrne,[108] Lastexit had been editing Overstock.com,[109] and Mantanmoreland, Tomstoner and Lastexit had all been editing Naked short selling.[110] The edits on Patrick Byrne and Overstock were decidedly negative, with the opposite true for Gary Weiss. This appears to be the issue that WordBomb intended to raise.

When he did so, WordBomb appears to have added the allegation to article space. He was immediately greeted by a vandalism warning from Mantanmoreland. A more appropriate response may have been to inform WordBomb of the proper way to raise his concerns, but clearly WordBomb’s approach was incorrect, so fair enough. Unfortunately, Lastexit then arrived to reinforce Mantanmoreland’s warning.[111] Nevertheless, it appears that WordBomb agreed to hold his concerns as of 20:28 7 July 2006 for a proper forum.[112]

One thing I do not fully understand is why SlimVirgin then arrived two hours later and blocked WordBomb’s account, “indefinitely.”[113] Of course one possibility is that she did not see WordBomb state that he intended to wait for a proper forum (although his edit summary did clearly state “concession”[114]); even so, an “indefinite block” without discussion or warning does not seem to have been necessary. By SlimVirgin’s most recent explanation, I see she states it was following this warning and because WordBomb repeated his allegation that she then decided he was editing in bad faith and protected his talk page.[115] This raises two concerns, however. First, SlimVirgin’s statement in originally blocking WordBomb indefinitely asked him if there “is an explanation for your edits.”[116] I am not sure what this was expected to elicit, but I don’t think it is entirely surprising that WordBomb would then repeat his allegation. Second, I see that in then protecting his talk page, SlimVirgin specifically stated in her edit summary that his claims were incorrect.[117]

We know then at this point that WordBomb was blocked, after being warned by what he recognized as two accounts of the same person, by a third who had specifically stated that his claim was incorrect. We know WordBomb believed, correctly, that there were other sockpuppets involved. Upon being asked to email his evidence on to SlimVirgin, then, we know he included an image, by his explanation, to determine whether the evidence was read. Incidentally, I don’t believe there is any evidence that WordBomb had a specific qualm at that point with SV or knew anything about her. Rather, after seeing a number of socks succeed in getting him blocked and silenced (rather than as he expected, the person who was socking), he used his own methods to figure out what was going on.

This is by no means the entire story. I believe this has been the source of a great number of other problems over the last two years, however. Moreover, to the extent WordBomb has been pilloried on Wikipedia over this time, while Mantanmoreland and Samiharris appear to have continued using multiple accounts to keep many of these related articles in a one-sided state (as well as often contributing to the pillorying), I think they are points that should be addressed here. Mackan79 (talk) 20:06, 16 February 2008 (UTC)[reply]

Mantanmoreland has learned from previous sockpuppetry mistakes and is thus more difficult to detect

Mantanmoreland has previously engaged in the abusive use of alternative accounts, which is uncontestable having been warned for it. Per the evidence presented by SirFozzie and CoolHandLuke (here, here and here) there is the very real likelihood that Mantanmoreland has also abusively used other accounts, or which he received no warnings but which were quietly retired[118] upon noting of suspicious behaviour. These previous instances of sockpuppeting are over two years old. I contend that having been caught out previously that anyone seeking to use the same method of supporting their point of view would have learned to be more circumspect if not to arouse further suspicion.

Also uncontestable is that Samiharris edited out of an untracable open proxy. Without the availability of two tracable ip addresses there has had to be a painstaking collecton and review of the contribution styles and mannerisms, dates and times, article interests and similarity of points of views of the two accounts. In those many thousands of edits there are relatively very few instances where it can be clearly demonstrated that there are aspects which the two accounts share and not by others. I contend that this would be expected from an individual who was both determined to evade the raising of suspicion, and who is experienced and sophisticated in the processes of Wikipedia.

While there is a good deal (subject to the above comments) of circumstantial evidence which may point toward abuse of alternative accounts there is no evidence of on Wiki action that effectively counters this material; there is nothing that a single editor may have concocted to support the illusion that two accounts were being operated by the one individual. If it is contested that there is no smoking gun to be found, I would strongly suggest that there has been found no White Hat either. LessHeard vanU (talk) 23:08, 15 February 2008 (UTC)[reply]

The reliance of off-Wiki emails to "prove" that the accounts were different individuals

In the current absence of permission to post (some) content of the emails sent by Mantanmoreland and Samiharris, as mentioned specifically by Jimbo and Guy, I have made a request that the timestamps for the email correspondence be presented so they may be checked against the same criteria as have the contributions of the two accounts. Again this would be looking for the instance of disparity (the White Hat) which would argue for the accounts being different people. I would however caution that emails are potentially very different from edits, as they are often literally composed and reviewed before sending. The likelihood of there being a smoking gun is considerably reduced, but also so is the White Hat scenario. Two posts may be created at different times but sent (and thus received) together. I therefore contend that evidence suggested by the interpretation of the content of emails is far less compelling than that offered by the review of edits made in "real (cyber) time". LessHeard vanU (talk) 23:08, 15 February 2008 (UTC)[reply]

Samiharris has left the building

Samiharris had previously been vocal in the condemnation of alleged socking by certain individuals ([119], [120]) or on the pages of certain articles ([121]), and has raised these allegations themselves. Upon notification that such an allegation against themselves was being investigated Samiharris posted a couple of messages commenting on off-site harassment and removed themselves from Wikipedia. There has been no indication of outrage or disbelief on what would be a slur against their good name, and there has been no indication of support to a fellow editor (with whom they had a good working relationship over a series of articles relating to a shared interest) who has been tarred by the same accusation. I find this very unfortunate, but also reminiscent of past accounts with the (strong) taint of sockpuppet. LessHeard vanU (talk) 23:08, 15 February 2008 (UTC)[reply]

Comments and response to Evidence by Mantanmoreland

Firstly, there is no hate campaign against Samiharris evident in the investigation of the allegations of sockpuppetry between the Mantanmoreland and Samiharris accounts. What campaigns there may be off-Wiki, and there is no specific evidence offered, are likely not of very recent instigation or either to have increased recently outside of these allegations. Insofar that those who are inclined toward accepting there is substance in the allegations the party receiving any indication of bad faith is Mantanmoreland, and there is very little of that outside of the serious accusation of violating a core Wikipedia policy. Mantanmoreland appears to me to be projecting any feelings of percieved ill-will toward him onto the personae of Samiharris.

Further, in his sub-section Response to Cla68 diffs Mantanmoreland appears to speak for Samiharris regarding a number of edits made by the latter, without supporting diffs or other evidence that this is exactly Samiharris' position. It is clearly Mantanmorelands point of view, but the points are made on behalf of Samiharris in a manner that indicates that Mantanmoreland is privy to their inner thoughts, and there is no dissimilarity between his and Samiharris' understanding.

This is of course opinion, but also evidence that there seems to be more than just an overlap in the perceptions of two accounts that have some common interests in editing a group of articles. It continues to point to the possibility that such commonality between the two accounts is not the product of coincidence or chance. LessHeard vanU (talk) 22:31, 17 February 2008 (UTC)[reply]

Overstock.com appears to be a matter of some interest to Gary Weiss

Above, User:JzG states "It is clear that Weiss and the issue of naked short selling amount to an obsession for Bagley, but I've seen no suggestion from anyone worth listening to that the reciprocal is true." While I won't take it on myself to engage in amateur psychoanalysis to determine if "obsession" is the correct word to describe this circumstance, I did a little original research on Mr. Weiss' blog, and it appears that he does indeed spend a fair amount of time discussing the Overstock.com/naked short selling/Patrick Byrne/Judd Bagley issue. (Note: I don't know much about the underlying disputes here, although some brief perusal of Google News demonstrates that Mr. Weiss is certainly not the only financial journalist with a low opinion of Overstock.com). A brief analysis of the last two months follows - more can be presented if helpful, but I believe it is sufficient to demonstrate that Mr. Weiss is indeed a relatively frequent participant in this dispute.

I was considering providing some representative samples of each post, but much of it would likely violate our biographies of living persons policy, so I will instead simply link to them. Again, let me be clear that I am not taking sides in the underlying dispute (my knowledge of financial markets is mostly limited to my 401(k)!), and that Mr. Weiss may well be correct in all of his characterizations.

February 2008

Mr. Weiss has made eleven blog posts thus far in the month of February. Of these, five are directly related to Overstock.com, Patrick Byrne, or Judd Bagley: [122], [123], [124], [125], [126], and another discusses short selling in general with a brief mention of Byrne - [127]. Thus far in February, Mr. Weiss has devoted 54.54% of his blog posts to this particular issue.

January 2008

In January, Mr. Weiss made twenty-eight blog posts. Of these, twelve are directly related to Overstock.com, Patrick Byrne, or Judd Bagley: [128], [129], [130], [131], [132], [133], [134], [135], [136], [137], [138], [139], and two discuss short selling in general with brief mentions of Byrne - [140], [141]. In January, Mr. Weiss devoted 50% of his blog posts to this particular issue.

The cyberstalking list

It's time to set the record straight about this list. Fortunately, Internet harassment is not a problem for the vast majority of Wikipedians. Some people do deal with it and the problems can be serious. The list came into existence because these are issues that really aren't feasible to discuss onsite. I am not at liberty to discuss other people's problems, but after careful consideration I've decided to disclose two of my own. The risk in doing this is that people who have a malicious intent could mine the information to intensify the harassment.

One early example I did address openly, and which was later mined for ammunition against me, was the community ban on Arkhamite. This was someone who posted a series of explicit sexual fantasies about me both onsite and offsite, and who had a self-disclosed arrest record and an inpatient psychiatric history. See Wikipedia:Community_sanction_noticeboard/Archive6#Proposed_community_ban_of_Arkhamite_and_68.84.17.112. His blog no longer contains the objectionable material and only Wikipedia sysops can read the deleted version of his user page, but the rest of the links I tested should still be good. This was a very straightforward ban that has never been contested. Nonetheless, a website known to the Wikipedia community selected that example and claimed that I had singled this editor out for no good reason at all, and driven him off the site because he had borrowed his username from Arkham Asylum of the Batman series. Actually my tastes run more toward Jane Austen than Marvel Comics and it was some time after the ban was implemented that I learned about this particular literary allusion. Other content at the same page left no doubt as to the malice with which this was chosen. And obviously I wasn't going to touch it.

For half a year, until quite recently, that same site hosted a photograph of my seventy-four-year-old uncle along with a not very subtle threat to harass him in real life if I don't stop editing Wikipedia. The caption correctly identified him as a World Trade Center survivor. Let there be no mistake: I've said this onsite before. He wasn't sipping cappuccino thirty blocks away--he was on the seventy-second floor of the north tower and watched his boss die before his eyes.

Those are two of the reasons I joined that cyberstalking list, along with other harassments I won't mention. A bunker mentality developed, in part because the people who called themselves critics of Wikipedia failed to uphold a fundamental baseline of ethics. Anyone could have removed that photograph from that website; it runs on wiki software. How much common sense does it take to see that going after someone's family is foul play?

Now I'm not going to voice an opinion on the merits of the circumstantial evidence here. I'm of two minds, frankly. On the one hand I want to believe the best and I have serious misgivings about my decisions that led the matter here. On the other hand, I think Mantanmoreland could have lied to me and played upon my trust in a place where I came to protect my elderly uncle, and that prospect makes me very very angry.

Per my presentation at SirFozzie's userspace, there is reason to accept this as a minimal baseline conclusion: Mantanmoreland and Samiharris have acted in concert to bypass normal consensus procedures at financial topic articles. Durova^Charge! 10:24, 16 February 2008 (UTC)[reply]

Samiharris has edited controversial articles without any participation from Mantanmoreland

I have edited extensively with Samiharris on financial topics. He appears to be very knowledgable in finance. Controversial articles that we have edited together include George Soros, Martha Stewart, Unrestricted Warfare. I've also seen him at "technical analysis" and asked him (as a person knowledgeable in the technical aspects of finance) to do a Good Article review for Option (finance). A couple of times we seemed to be operating at cross-purposes (rather than really disagreeing), but mostly we've shared similar views. Soros, especially, has been a very controversial article. Mantanmoreland never once (as far as I can tell) poked his head into any of the controversies there. The recent Martha Stewart controvery was much shorter, but Mantanmoreland did not show up. In short, I'm very familiar with Samiharris's editing in certain controversial articles, and Mantanmoreland did not act as a sockpuppet for Samiharris in any of these articles.

Most of the "statistical analysis" above is essentially meaningless

Extraordinary claims require extraordinary evidence. Since there is no ip evidence showing that Samiharris is a sockpuppet of Mantanmoreland, people have started presenting "statistical evidence." Anybody familar with proper statistical procedures will tell you that the "statistical" methods used are just horrendous. Given enough data and enough time, I could "prove" almost anything using those methods.

Response to Relata Refero -- RR disputes my interpretation of this analysis above, but says "The size of the dataset is extremely small; a model is unspecified; and there are sundry other problems." Actually, I think that says it all, particularly the unspecified model part.

She also states that the use of stats implies trust - that we have to trust the analysts not to do data mining (this is correct). So perhaps (my reading) I'm violating WP:AGF by not taking the analysis at face value. I think that trust and AGF should also apply to Samiharris.

To be clear, I'm not accusing anybody of intentionally mining the data, but I think there is a tendancy - even among some professionals - to just keep on looking until they "find something," without reporting all the negative results (what else did they think of looking for and did not report not finding?) The idea that this interleaving pattern that was found a) means something, and b) was the only pattern they looked for I find incredible. Smallbones (talk) 13:53, 17 February 2008 (UTC)[reply]

Samiharris is not Gary Weiss

As I understand it, the claim is that Sami = Mantanmoreland = GW. I read GW's book on short selling (at the same time I was editing here with Sami). While I don't claim any extensive comparisons of text, time stamps, etc., it is obvious that they have completely different styles. GW is a journalist, Sami has a more academic style.

Smallbones (talk) 16:17, 16 February 2008 (UTC)[reply]

Further COI evidence, linking Gary Weiss to Lastexit and Tomstoner, hence to User:Mantanmoreland

It has been established that the Lastexit and Tomstoner accounts were operated by Mantanmoreland; Fred Bauder got CU confirmation on Lastexit, and everyone has seen the diff where Mantan extensively revises a Tomstoner post and even writes a fresh paragraph from scratch, replete with first-person references to Tomstoner's contribution history. And the Mantanmoreland account itself has been compellingly linked to Gary Weiss by the Varkala edit, preceding as it does the dramatic time-zone shift in Mantan's editing schedule when Weiss left for India, where he was married in Varkala.

I believe I have now found Weiss' fingerprints in the contribution histories of both the Lastexit and Tomstoner accounts; fingerprints which, together with the Varkala edit, make the case for Weiss' COI problem virtually bulletproof. Please bear with me if the presentation is a little novelistic. There are real toads in this imaginary garden.

Tomstoner set up shop on February 9, 2006. His first edit was to Mohandas Karamchand Gandhi, and focussed on antisemitism in India. (His next six edits likewise focussed on India, with the emphasis now on tourism.) Weiss's interest in India is by now well-known, and he refers to it often enough on his blog; what is less known however is that he has written specifically on "India's Jews" in an article for Forbes magazine. The Forbes piece was published more than a year after Tomstoner's edit, so it isn't possible that Tomstoner read it first. As with the Varkala edit, the Wikipedian's expressed interest precedes the journalist's.

Tomstoner's choice of user name becomes more interesting when User:Lastexit arrives on the scene on April 22, sets himself up as a Southwesterner, and makes his first edit [142] to Tombstone, Arizona, a ghostly homage to his predecessor and matching sock. Lastexit then makes edits to Josephine Earp, who made her last exit in Tombstone, and then to Earp Vendetta Ride, a three-week bloodbath set in motion by the assassination of a U.S. Deputy Marshall in Tombstone, Arizona.

Right about this time, April of 2006, Gary Weiss has the Southwest on his mind. Perhaps he travelled there that spring; at any rate three days before Lastexit makes his first entrance Weiss writes the following on his blog:

If you're ever in Albuquerque, be sure to visit the Ernie Pyle home, now a municipal branch library. It's beautifully preserved, has a memorial to Ernie on the grounds, and contains a large amount of memorabilia.

Now turn to Lastexit's edits – three weeks or so later – to Albuquerque, where he "add(s) Ernie Pyle library to list of historic places", and to Ernie Pyle: [143], [144]. He next creates an article stub for Ernie Pyle House/Library: [145] [146].

Now hold those croaking toads while we stroll through what remains of this imaginary garden. Not mine but Weiss's, which if I'm correct lies west of Brooklyn, all the way over in the Arizona territories, where the aforementioned Earp federal posse avenged the killing of one of their own, Morgan Earp. Just about a hundred years later, Don Bolles, an investigative journalist for the Arizona Republic looking into connections between organized crime and stock fraud, was lured by a mafia "source" and killed by car bomb in downtown Phoenix. A "posse" of his fellow IRE journalists rallied and organized "The Arizona Project," which

descended on Arizona for a massive investigation. They set out to find not Bolles' killer, but the sources of corruption so deep that a reporter could be killed in broad daylight in the middle of town. They were out to show organized crime leaders that killing a journalist would not stop reportage about them; it would increase it 100-fold. [147]

Weiss has written in Forbes about the Arizona project, and how it inspired his own "Project Klebnikov," founded in 2005 after some thirteen reporters – including the American journalist Paul Klebnikov – were killed in Moscow:

The parallels between Klebnikov's slaying and the murder of Don Bolles, an Arizona journalist slain in 1976, are becoming increasingly apparent. Bolles was killed for probing the mobsters and land-fraud schemes that plagued the Southwest in the mid-1970s.

The Bolles murder resulted in the creation of the Arizona Project, a consortium of journalists that was created to continue Bolles' work. Scott Armstrong and I, along with Richard Behar and others, are members of Project Klebnikov, which has similar aims in continuing Paul's legacy."Murder in Russia," by Gary Weiss, Forbes magazine, January 30, 2007

A few hours after Lastexit debuts with his Tombstone and Earp vendetta ride edits, he turns to Gary Weiss, highlighting Project Klebnikov (previously added by Mantanmoreland) by giving it its own section. A week later Lastexit adds another section to the same article, this one focusing on how Weiss has been described as "an old-time gumshoe, with a soupçon of little-guy champion Jimmy Breslin and a dash of 1950s bad-boy comic Lenny Bruce," and how this courage and intrepidness "has provoked a vituperative response, including threats." These perils of being a straight-shooter are the imaginative link between Weiss and Paul Klebnikov and Don Bolles and Morgan Earp.

The two-year edit war now drawing to a close is Weiss's vendetta ride, this time not against shadowy killers targeting journalists abroad but rather against Overstock, Byrne, and other anti-naked-shorting activists allegedly polluting our domestic markets and financial discourse, "threatening" old-time gumshoes like Weiss, and – most recently – disseminating "Bagley memes"; a vendetta ride, as we've seen, with Mantan Moreland in the starring role, and Mantan's uncle – a proud resident of Albuquerque, New Mexico – and Tombstone Tomstoner riding side-saddle (at least until they're picked off towards the end of the first act). Mantan's sockpuppetry has always seemed so crude to me, but in its initial conception it was noble, comically baroque in a Pynchonian way, and even kind of beautiful. Tombstone is its luminous touchstone. Earp's posse rounded up in Tombstone in 1882 is the mythical precursor to the IRE posse rounded up for the "Arizona project" in 1976, which in turn is the precursor to Weiss's posse rounded up for "Project Klebnikov" in 2005; which in turn, finally, is the precursor for the posse of sockpuppets – each with its homage to Tombstone – that Mantanmoreland began to round up on the semi-lawless frontier of Wikipedia in the spring of 2006.--G-Dett (talk) 20:54, 16 February 2008 (UTC)[reply]

Postscript: A fellow Wikipedian has pointed out to me that Josephine Earp, Wyatt’s wife and the focus of User:Lastexit’s first edits after his Tombstone debut, was Jewish. As it happens, there was a significant Jewish population in the Tombstone of the 19th century; indeed the town's very name was given to it by a Jewish pioneer who opened its first general store. [148] In light of the striking nexus of interests here – Jewish history, the Earp vendetta ride, historical/tourist sites in the Southwest, and the legends of Tom(b)stone(r) – I regarded this as an interesting lead. I understand that what I’ve found and am presenting is 'evidence' of a peculiar sort.

Tombstone’s most famous tourist site is the Boothill Graveyard, where many of its legendary gunslingers and historical personalities are interred. Boothill has within it a Jewish section, which went unnoticed for over 100 years; a memorial was added in 1984. [149] [150] The small Jewish burial ground has no remaining headstones, and only one grave – that of a child. [151] He died in 1889, when he was one year and four days old. There is still a small stone marker for the child in the burial ground today, next to the memorial. His name was Sam Harris.

I would like to be able to say that User:Samiharris was created one year and four days after Mantanmoreland was created, but that would be a few hours off. He was created – for what it's worth – one year, three days and ~three hours after Mantanmoreland. There is a touch of the poet (as well as the gumshoe) in old Weiss. I am moved, almost to clemency, just reflecting on it.--G-Dett (talk) 01:05, 18 February 2008 (UTC)[reply]

Sockpuppeting

I can't respond to the statistical evidence that has been introduced because, as Smallbones correctly points out, it is "horrendous." In fact, all of the pages associated with this case are horrendous, not the least by personal attacks of every kind, including one left on my user page by an administrator[152]. This case reads like some kind of appendix to Wikipedia Review or Overstock.com's antisocialmedia.net website. Most of the contents of this page have been on the ASM website for well over a year. If Judd Bagley, Overstock's spokesman and operator of ASM, were not coordinating this, I am sure he would have a case for copyright infringement.

The editors who know Sami and I best, and are not necessarily friends of either of us, believe quite firmly that we are different people based on writing style and the positions we took, and didn't take, in 600-odd emails.

The articles that seem to be most at issue here (Overstock.com, Gary Weiss, Patrick M. Byrne and naked short selling) are stable. They are not a "battleground." They have had very little activity except for periodic assaults by sockpuppets of WordBomb (that is, Bagley). No one is claiming that Sami and I edit collusively in those articles. Indeed, the main objective of all the "statistics" is to prove that we don't edit collusively. Despite all the hysterical calls for these articles to be deleted and salted and exorcised, the articles are in good shape.

In what I assure you was an unsolicited opinion on these articles generally (or possibly just naked short selling), Jimbo said:

There is a fringe conspiracy theory here, one which is not reflected in reliable sources, and refusing to allow paid corporate POV pushers to control an article because they have a willingness to engage in "dirty tricks" campaigns is always going to be the right thing to do. The article, when I have looked at it (not recently) seemed to be quite good to me, whereas the version proposed by the other side was not even remotely close to ok.

[153]

We don't behave like sockpuppets. A good example is when Samiharris became involved in a major content dispute in Gary Weiss in December. Here is the edit history for the relevant period. [154] I did not participate in that discussion. It was civil, it was confined to the talk page, and it ended amicably, despite provocations. Samiharris could have used help from another editor, but I did not want to get involved. Even though I promised SlimVirgin in November 2006 that I would avoid the Weiss article because of all the off-wiki harassment, I had not pledged to not participate in talk page discussions. A sockpuppet would have chimed right in because Sami was outnumbered.

Samiharris additionally became involved, without my participation, in enforcing policy in Depository Trust & Clearing Corporation. The article was being subjected to extensive POV edits, WEIGHT violations and gross misuse of sources. I did not become involved in that dispute and have never edited that article or commented on its talk page.[155] An administrator intervened at Samiharris' request, because of the issues he raised, and put the article under protection.

Response to Relata Refero

It is ludicrous to claim that there was "tendentious editing" in naked short selling. All the edits have been well within policy. You may not agree with them, but that does not make them "tendentious." There was no edit warring over anything you cite. It is correct that there are few academic studies cited within the article. This is not evidence of anything. No one attempted to add academic studies, only to be rebuffed. Naked shorting is a subject that has received widespread news coverage, so naturally there are news articles cited. For most substantive points, the article relies on the SEC website. You also disregard the fact that the article has been edited by quite a few editors, and was rewritten by a previously uninvolved editor, Marty Markowitz, late last year.

(Further response, citing specific diffs, will follow.)

Response to Cla68 diffs

Cla68 is mainly upset with the way he is treated by administrators, but he does cite several diffs relating to my editing and Samiharris', all of which warrant response:

In this [156] he wished to make an entire section out of a ten-year-old libel suit that was withdrawn. It was removed per WP:BLP and WP:WEIGHT. Cla68 himself conceded the latter point.[157]

He objects to this [158] edit by myself properly removing disruptive "fact" tags after every paragraph in the article, and clarifying that the sources were the websites noted.

He objects to this [159] perfectly proper edit by Samiharris, cutting the length of an overlong paragraph on the antisocialmedia.net smear campaign, correcting an inaccuracy that attributed an allegation to the wrong source.

For some reason he objects to this edit by Samiharris [160] that removed an allegation by Weiss that Patrick M. Byrne was "in the middle of a meltdown," turning that into the far less inflammatory "has made critical comments about him in articles and in media interviews." This edit protects Byrne, not Weiss. That same edit also removed a notable journalist's comment:

'Bloomberg.com columnist Susan Antilla writes that the website attack on Weiss, "Is but the latest example of the public relations path Overstock and Bagley have taken to wage their bizarre battle against naked shorts."

Removing a quote favorable to Weiss is not "POV pushing" on behalf of Weiss.

This Samiharris edit[161]simply trimmed an overlong external links section, removing a New York Post article that does not mention Weiss, an article by Susan Antilla highly complimentary of Weiss, and a Times column that is uncomplimentary to Weiss and a critic. This is a routine cleanup and hardly POV pushing.

This Samiharris edit [162] is misrepresented as "(this last is a paragraph that had been agreed to on the talk page)." In fact, as can be seen from the relevent portion of the talk page [163], Samiharris removed it in consultation with an administrator under WP:NPF. ("That point was clarified by an administrator uninvolved in the editing of this article.") The administrator's post on that point is here.[164]

I have no idea what he is trying to prove with this Samiharris edit [165], which is a reasonable talk page post, as is this one [166].

The rest of his "evidence" relates to the raw deal he supposedly has gotten through much of his wiki-life, and has nothing to do with this arbitration.

Response to Noroton

I think the effort to find "uncommon phrasing" in edit summaries has reached ridiculous proportions as an example of confirmation bias, and this is a good example of it. VERBing is used frequently on the pages of Wikipedia, particularly by the bots that crawl over Wiki articles. Here is one[167], with a 100% VERBing rate. Here is another [168]. I am not a sockpuppet of VolkovBot or DumZiBoT or Alison[169]

Response to LessHeard vanU

Samiharris has left the building because of precisely the hate campaign that you see reflected in these pages and off-wiki. He said months ago on a private email list that he would leave if something like this happened. It did, and he has left.

Study of editing collision patterns in 2007

I have conducted some research that may be pertinent to the question of non-colliding edits by Mantanmoreland and Samiharris. The fundamental question I sought to answer is: How unusual is it for editors with edit counts at the level of MM and SH to avoid editing at the same time?

I will detail my methodology at User:Alanyst/Edit collision research. But here are the raw numbers so far (all numbers derived from the early January 2008 data dump):

• I'm looking at 3629 editors (including anon IPs), all of whom had between 1000 and 2000 edits during 2007. MM and SH are included; MM had 1680 and SH had 1201 edits during that time.
• There were 343 (roughly 9.5%) of those editors who never edited during the same minute as Mantanmoreland.
• There were 610 (roughly 16.8%) of those editors who never edited during the same minute as Samiharris.

Caveats:

• I have not analyzed how these editors' edits interleave with MM or SH's edits
• I have not analyzed which of these editors' active times are inverted with respect to MM and SH's active times
• My methodology has introduced certain biases that might have skewed the numbers slightly, though I believe they still give a pretty good picture
• I have only looked at collisions with other editors for MM and SH; I have not yet examined collisions for other editors drawn from the same group
• Some IPs represent multiple individuals editing at various times, which might have an impact on the overall pattern

Discussion:

• This seems to challenge somewhat the notion that a lack of simultaneous edits for editors with edit counts similar to those of MM and SH is indicative of a coordinated pattern. ("Simultaneous" here means "during the same minute".) There's a low but still reasonable chance that two independent editors will never produce near-simultaneous edits.
• On the other hand, it may be that most of the non-colliding editors have substantially different times of activity than MM and SH, which would improve the odds of never colliding. This needs to be investigated further.
• It's not clear what an appropriate definition of simultaneity would be, as far as edits are concerned.
• Wikipedia is waaay too big to analyze easily, and I'm not good at cool graphs; sorry.

I wish this could be more solid, but perhaps it can aid the discussion. I will try to write up the methology but it will be at least a day before that happens. alanyst ^/talk/ 09:17, 17 February 2008 (UTC)[reply]

Use of "VERB-ing" as first word in edit summaries

It would be very hard for any single similarity in the editing styles of these four to conclusively prove that they're the same person. You need as many similarities as possible, and preferably each similarity would be very rare in Wikipedia. I doubt this particular similarity is extremely rare, but when I look for it in the edit summaries of other editors, I just don't find it. So I think it's worth adding to the pile.

The contributions pages for each of these four accounts have an unusually high number of edit summaries using the same kind of "VERB-ing" construction as the first word on the line.

Here's a sample of a cluster of them from Mantanmoreland, who uses them less than the other three accounts, but still more than anyone else I can find (I've put the initial verbs in boldface):

13:25, April 18, 2007 (hist) (diff) Wikipedia talk:Attack sites‎ (→How Hanlon's Razor isn't going to help us: expanding comment)

13:15, April 18, 2007 (hist) (diff) m Wikipedia talk:Attack sites‎ (→How Hanlon's Razor isn't going to help us: putting in correct timeorder)

13:13, April 18, 2007 (hist) (diff) Wikipedia talk:Attack sites‎ (→How Hanlon's Razor isn't going to help us)

13:04, April 18, 2007 (hist) (diff) Wikipedia talk:Attack sites‎ (→How Hanlon's Razor isn't going to help us)

12:35, April 18, 2007 (hist) (diff) Wikipedia talk:Attack sites‎ (→How Hanlon's Razor isn't going to help us: clarifying)

12:20, April 18, 2007 (hist) (diff) Wikipedia talk:Attack sites‎ (→How Hanlon's Razor isn't going to help us)

12:15, April 18, 2007 (hist) (diff) Wikipedia talk:Attack sites‎ (→How Hanlon's Razor isn't going to help us: clarifying)

12:14, April 18, 2007 (hist) (diff) m Wikipedia talk:Attack sites‎ (→How Hanlon's Razor isn't going to help us)

12:13, April 18, 2007 (hist) (diff) m Wikipedia talk:Attack sites‎ (→How Hanlon's Razor isn't going to help us: fixing indent)

There was no particular reason to use the -ing form of the verbs. Many editors use either the past tense ("fixed") or present tense ("fix") if they use a verb in the edit summaries.

Look at the contributions pages for each of these four accounts and you'll find more:

• Mantanmoreland (talk · contribs · deleted contribs · logs · filter log · block user · block log) -- uses them less than the rest,
• Samiharris (talk · contribs · deleted contribs · logs · filter log · block user · block log) -- especially the earliest page
• Lastexit (talk · contribs · deleted contribs · logs · filter log · block user · block log) -- especially the earliest page
• Tomstoner (talk · contribs · deleted contribs · logs · filter log · block user · block log) -- especially the most recent page

This evidence would be stronger if it were compared to those of other editors, but I'm unsure about what set of editors to compare it to (perhaps editors who contribute to other finance-related pages; preferably not the ones that these accounts contributed to?), and I don't have the tools to give a count. The most striking way to look at it is simply to go look at it on the contributions pages.

It's been noted that inevitably, given a big enough sample, you'll find purely coincidental groupings of the same characteristic. And you'll find them even more if there are underlying similarities (editors from the same place, with the same interests). But each time you add a strand to the rope you make it less likely that we're dealing with coincidence. Consider this characterisic one more strand.

Use of space-hyphen-hyphen-space in edit summaries

Cool Hand Luke mentions the space-hyphen-hyphen-space construction (" -- ") as a distinctive feature of the edit summaries of Mantanmoreland, Samiharris, etc. To the extent that the construction is rare in edit summaries, Luke has a point. But that was a standard way to represent a dash in the Before Computers Era, so we can expect other editors to use it, possibly in edit summaries. The Associated Press Stylebook (1998 edition) states:

WITH SPACES: Put a space on both sides of a dash in all uses except the start of a paragraph and sports agate summaries.

LOCATION ON KEYBOARD: On most manual typewriters, the dash must be indicated by striking the hyphen key twice. On most video display terminals, however, there is a separate key that should be used to provide the unique dash symbol with one keystroke.

-- from the "Punctuation marks and how to use them" section under "dash"

Noroton (talk) 20:11, 17 February 2008 (UTC)[reply]

Response to Mantanmoreland

I did notice a bot that used "VERB-ing" as a first edit-summary word, and there seem to be more, meaning editors who create bots have put that in the program. What would be dispositive is to find a fair sample with a good number of human editors who do it. I'll continue to look. Noroton (talk) 21:12, 17 February 2008 (UTC)[reply]

On "confirmation bias": I've been waiting to learn more and see more of your defense before coming to a conclusion, and I don't need to confirm anything. I've shown that I'm not interested in overselling the case against you, which at this point seems strong. Noroton (talk) 22:24, 17 February 2008 (UTC)[reply]

This will be your finest hour

Hi. I am just checking in on this after a long day on the road. I have another tomorrow, but within a day or two will post something that I think most will find decisive. I fear it will be disappeared down the Orwellian memory hole, but at this point I think that would be noted. In any case, I will also post it on my blog DeepCapture.com so that cannot happen. For what it is worth, I know that this must be distressing to many good people of the Wikipedia community. I have an idea of where this is going to lead (short of a Battista-style counter-attack), but assure you, in retrospect it is going to stand as your finest hour. - Patrick PatrickByrne (talk) 01:40, 18 February 2008 (UTC)[reply]

before using the last evidence template, please make a copy for the next person

{Write your assertion here}

Place argument and diffs which support your assertion; for example, your first assertion might be "So-and-so engages in edit warring", which should be the title of this section. Here you would show specific edits to specific articles which show So-and-so engaging in edit warring.

{Write your assertion here}

Place argument and diffs which support the second assertion; for example, your second assertion might be "So-and-so makes personal attacks", which should be the title of this section. Here you would show specific edits where So-and-so made personal attacks.