Wikipedia talk:Sockpuppet investigations/Scibaby

Discussion from Report date December 20 2009, 11:27 (UTC)
Is it just me or does this look like more fishing? Does anyone seriously believe that one person could be creating all these socks? I'm just sayin', overall there's a HUGE number of socks attributed to this one user. I don't know how accurate and specific the checkuser details are, but I would hope that they are pretty specific to justify so many checks/blocks. What is the rejection rate for these queries over time? How many people are having their privacy invaded via these checks? --GoRight (talk) 23:46, 20 December 2009 (UTC)


 * If you have concerns that checkuser is being abused, the Audit Subcommittee is thataway→. Short Brigade Harvester Boris (talk) 23:54, 20 December 2009 (UTC)

Holy crap, this is what is being put up as evidence of sock puppetry to justify a check user?
 * Putting up a single edit for the POV template on RealClimate?,  This is a signature unique to SciBaby?  No, no one but Scibaby could think that article has a POV problem.

Most of the others seem to be generic AGW skeptics material as well. Are we still condoning the practice of check usering any new account that expresses an even vaguely AGW skeptic POV? What are the signature characteristics in this evidence that suggests that they are made by Scibaby? --GoRight (talk) 00:01, 21 December 2009 (UTC)


 * Scibaby is known to have created at least 500 sockpuppets (see Category:Wikipedia sockpuppets of Scibaby). He's certainly the most prolific sockpuppeteer I've ever come across. Unfortunately there are technical difficulties with blocking him permanently, otherwise it would have been done long ago. In my (limited) experience of Scibaby sock investigations, I'd say that the overall rejection rate is pretty low. You can spot him fairly reliably if you're familiar with certain aspects of his behaviour. I agree that his behaviour is somewhat "unbelievable" but it's indicative of the malicious nuttiness that seems to be unpleasantly frequent on one side of this debate.


 * I'm not going to go into details about "signature characteristics" since that would only help Scibaby evade his blocks (see WP:BEANS). Looking at those diffs you posted, it's clearly the same user - probably Scibaby, perhaps another sockpuppeteer. -- ChrisO (talk) 00:02, 21 December 2009 (UTC)
 * Heh. Like I said, you guys are fishing and invading people's privacy in the process.  --GoRight (talk) 00:11, 21 December 2009 (UTC)
 * No one's privacy is being invaded, because no private details have been provided. The information being examined is only that which is submitted every time you visit any website anywhere on the Internet, and even that limited information is being kept private. MastCell Talk 00:17, 21 December 2009 (UTC)
 * If the FBI or some other police agency looks into someone's phone records without probable cause, is that considered invasion of privacy? I think it is.  This is the same thing.  I'm not worried about Scibaby, it's all those rejections that are having their privacy invaded.  --GoRight (talk) 00:34, 21 December 2009 (UTC)


 * Your concerns are noted. Perhaps you might find a moment to criticise Scibaby's abusive conduct, since concern about that seems to be rather lacking in your comments. As for your comment about "AGW skeptics" being targeted, the unfortunate fact is that there is a disproportionate amount of abusive conduct coming from that side of the debate. I don't normally edit AGW-related articles but since I started monitoring a few for POV and BLP violations I've noticed a considerable amount of abusive conduct - vandalism, sockpuppeting, ranting, BLP violations and so on. Scibaby's persistent extremism is admittedly an outlier, but it's indicative of some wider trends in the way that so-called "skeptics" appear to operate on Wikipedia. You might want to ask yourself why the side that you associate with should stoop to such tactics. -- ChrisO (talk) 00:24, 21 December 2009 (UTC)
 * "the unfortunate fact is that there is a disproportionate amount of abusive conduct coming from that side of the debate" - I've observed the GW articles for more than a year now, and that is completely wrong. I think maybe your POV is clouding your judgement here. ATren (talk) 01:38, 21 December 2009 (UTC)
 * All I can say is that in the relatively short time I've been involved, all the abusive conduct I've seen - sockpuppetry, vandalism etc - has come from one side. Where are the pro-AGW sockpuppeteers? -- ChrisO (talk) 07:49, 21 December 2009 (UTC)
 * "I'd say that the overall rejection rate is pretty low." - Well, just looking at the 17th and 19th from above we find 25 suspects, 8 or which were rejected. (Assuming I counted correctly.)  So you definition of "pretty low" is on the order of 32%?  --GoRight (talk) 00:16, 21 December 2009 (UTC)
 * I would suggest you do some more research. But yes, even 32% is pretty low - that means 2 out of 3 suspects are Scibaby - or, to look at it another way, there are more Scibaby socks than legitimate editors. Also note that "rejected" does not guarantee that the account in question is not a sock - it only indicates that it cannot be shown to be sock of Scibaby. It can still be a better disguised sock, or a sock of another user. Of course, it can also be a harmless new account, but then it's not hurt in any way at all. --Stephan Schulz (talk) 00:33, 21 December 2009 (UTC)
 * I would suggest that you need to be required to provide some actual evidence rather than something as flimsy as a single edit which is a POV template on RealClimate.


 * "It can still be a better disguised sock, or a sock of another user." - Nice use of WP:AGF. Also, if they run a checkuser on someone because of your Scibaby allegation and it turns out to be a sock puppet of someone else, do they just ignore that or does the account still get blocked citing the other account as the sock master?  If the latter then your second option would be vacuous.


 * "Of course, it can also be a harmless new account, but then it's not hurt in any way at all." - Perhaps in Britain you are more used to having your privacy invaded. Here in the US we take such things seriously.  --GoRight (talk) 00:42, 21 December 2009 (UTC)


 * In addition, I think that Wejer might disagree about whether they were harmed, or not, . --GoRight (talk) 01:05, 21 December 2009 (UTC)


 * Or maybe we should ask User:Q Science if they were harmed?   --GoRight (talk) 01:21, 21 December 2009 (UTC)


 * Or maybe we should ask User:GoRight ... oh wait that's me ... whether being the target of such a witch hunt damages your reputation? See   Yes, I can confirm that it does.  The accusation of being a sock master or a meat puppet, even when you are exonerated, never disappears.  Need a case in point?  Just see your comment above.  As for the longevity of such baseless accusations, I refer you to  and .  --GoRight (talk) 01:32, 21 December 2009 (UTC)

I note that the history on this page only goes back to June. If Scibaby has over 500 confirmed socks where are the rest of the history records of the investigations? --GoRight (talk) 01:34, 21 December 2009 (UTC)

Why was User:Elixthecat suspected of being a Scibaby sock puppet? Looking at their contributions I see one edit putting up a POV template on a rather obscure BLP for David Deming (or at least I had never heard of him). Looking through the history there I see no evidence that Scibaby has been a problem on that page, but I may have missed it so I am asking for an explanation. The edit in question was reverted by some IP and ultimately restored by User:Tony Sidaway. So why was this account checked? --GoRight (talk) 01:58, 21 December 2009 (UTC)
 * The correct fora for taking up concerns is WP:AUSC, as Boris pointed out about. --Kim D. Petersen (talk) 07:35, 21 December 2009 (UTC)

This extended discussion might perhaps have been better located on the talk page. This section is for brief comments pertinent to the case. --TS 03:15, 21 December 2009 (UTC)

Okay, this is completely ridiculous. This "sockmaster" was reported by a desysopped admin (who was also topic banned from global warming articles a long time ago) the pattern being making edits about the blog "Real Climate" (where the former admin was one of nine contributers), his friend on facebook, Kim Dabelstein, backs him up, his other friend Stephan Schulz backs him up and his friend, currently going by Short Harvester Boris also backs him up - and all these people have had the same behavior and edits on articles regarding global warming articles for YEARS - instead you go after some guy for making a single edit? Madness.

Finally, you want a good explanation for some of these people? I imagine a lot of them are new users to wikipedia who are being led here by blogs complaining about wikipedia's bias (and there are been several recent articles about this). The reporting admin, in fact (WMC for clarification), has been the subject of many news and blog articles about the bias of wikipedia - it is no surprise they'd show up on the articles he and his friend WP:OWN. TheGoodLocust (talk) 22:43, 24 December 2009 (UTC)


 * If you want to be taken somewhat seriously, you should try to avoid making obviously wrong statements - the Big Lie does not work too well if you cannot control all media and everybody has easy access to the raw data. As obvious from the full page history, WMC has never ever reported anyone as a sockmaster on this page. He also was never topic-banned from any topic. If you believe otherwise, I suggest you show some evidence. --Stephan Schulz (talk) 08:57, 25 December 2009 (UTC)


 * Obviously wrong statements eh? What do you make of | this diff? Connolley accuses him of sockpupptery, blocks him indefinitely, and then YOU come along to confirm it all and link to your complaint. I guess you must've forgotten that this all started with you people.


 * But I suppose I did make one minor mistake, | Connolley was "paroled" from global warming articles, not topic banned, for a period of 6 months, but since all he seems to do is revert content away then that would be a semi-ban I think. Oh yeah, you really got me there Schulz. TheGoodLocust (talk) 02:26, 27 December 2009 (UTC)


 * Oh, and unlike some people I don't watch articles, and certainly not this one, if you want a real response then I'll need another heads up. Please don't state any more provably incorrect statements with such conviction - I probably won't return to correct them. TheGoodLocust (talk) 02:26, 27 December 2009 (UTC)


 * Actually, your diff does not show William reporting anyone anywhere. I did, indeed, report him, and the CU called it a WP:DUCK case. William was not "paroled from global warming articles" - that does not even make sense. The 1RR parole was revoked by ArbCom. --Stephan Schulz (talk) 09:38, 27 December 2009 (UTC)


 * Scibaby's blocklog shows one blocking admin - Connolley. Connolley accused him on being a sock on his talkpage, blocked him, and then you came along and linked your complaint. If the chronology is slightly altered that doesn't change my main point - he indefinitely blocked someone he was "edit warring" with - although, to be fair, scibaby's account didn't last long enough to edit war too much. If I accidentally confuse you, Connolley, Petersen, and Boris, then that's only due to the similarities in your behavior. And William was paroled from global warming articles, my link makes that quite clear and that is even the language that was used - if he managed to talk his way out of it later then I'm not surprised. TheGoodLocust (talk) 20:02, 27 December 2009 (UTC)

BC Builder
This sock took an exceptionally long time to reach checkuser, although once it was listed here the pace was reasonably fast given the sheer weight of work facing the clerks and checkusers. I suppose I could have listed this user myself about a week ago but the temptation to assume that this was good faith if slightly eccentric, slightly sarcastic editor was strong. I have similar suspicions about another recent editor on talk:global warming but am unsure about how I would go about measuring the likelihood that it was a sock to my own satisfaction, in order to avoid a harmful listing of a perfectly innocent editor. --TS 15:07, 22 February 2010 (UTC)


 * BCBuilder was a very, very obvious Scibaby sock. I'm surprised people kept arguing with him. Short Brigade Harvester Boris (talk) 15:33, 22 February 2010 (UTC)

Split out BillMarrsx
Can BillMarrsx be split out ... it was added after the orginal endorsement which is becoming an annoying and recuring practice. Zulu Papa 5 * (talk) 17:27, 3 June 2010 (UTC)

Error rates
Is it possible to count up the True Postives and False Positives from the archives?

I would like to estimate the "Net Benefit" which is (TP - (w * FP))/N Where w is the ratio of (Good / (1 - Good)) ) or Good/Harm (1- Good) / (Good)) or Harm/Good of a false assumption. 

Thanks Zulu Papa 5 * (talk) 15:59, 4 June 2010 (UTC)


 * Sure. Go ahead. Collect counted "more than 20" reported accounts which were not Scibaby in the first 4 months of 2010. I found 133 Scibabys and 12 other confirmed socks which may or may not be Scibaby. My impression is that Collects 20+ include those 12, but I'm not sure. Of course, as far as I can tell most "false positives" were never blocked - I've asked Collect for clarification. And, of course, you need to come up with reasonable costs for false positives and false negatives if you want to apply that model. --Stephan Schulz (talk) 16:38, 4 June 2010 (UTC)

Ok, except after checking again w = (1- Good) / (Good)) or Harm/Good.  It's simple to start with w = 1 where the good = harm.  Them move from there. So as is:

TP = 133 Scibaby's

FP = 20 non-socks

TN = 12 socks but not Scibaby

N = 133 + 20 + 12 = 165 NB1 = (133 - 1*20) /165 = 69.9%

Or, with socks found included in TP,

NB2 = (133 + 12 - 1*20)/165 = 75.75%

Ok, let's set NB=0 for each, and find that w = (TP/FP) at no net benefit, which then implies the amount of harm to offset the good.

W1 = 133/20 = 6.65

W2 = (133+12)/20 = 7.25

So, then W2 - W1 = 0.6 increase in (harm/good) at no net benefits. Seems like a reasonably harmless increase, unless you are the one harmed. Zulu Papa 5 * (talk) 17:20, 4 June 2010 (UTC)


 * I don't remotely understand your computation or the aim of it. Of course, the problem addressed in your links deals with the much more frequent case in medicine, where a given condition is rare, and there are significant costs to both the test and false positives. Here, the condition (sock) is obviously much more likely than the opposite. Which formulas (pages?) on the PDF did you apply? --Stephan Schulz (talk) 17:40, 4 June 2010 (UTC)
 * The formula is a variation from page 41 on the original pdf. It can be found again in the simple form I applied on page 26 of .  This is Decision Curve Analysis, a new technology.  It is simply a risk weighted average of the rates.  Applying this to sock investigations is original research. Zulu Papa 5 * (talk) 18:34, 4 June 2010 (UTC)


 * According to the formula on page 26, the net benefit is $$\frac{TP-wFP}{N}$$, for w=1 it would hence be $$\frac{133-20}{153}=0.74$$ (ignoring the 12 other socks - adding them would make a minor change, but net benefit would in either case be much bigger than 0.5). How do you arrive at "no net benefit"?


 * Sure Ok, I found the Net Benefit to be 0.6969 and 0.7575 when w = 1 .... the No Net Benefit was my asumption that when we set Net Benefit = 0 and then solve to find w, we have a way of normalizing each situation for an equal comparison between the implied w values.  In your example w = (133/20)= 6.65 at the no net benefit point.  The No Net Benefit was an assumption not a conclusion.  My analysis attempted to find inflections points ( w = 1) and the singular values (NB=0) first and then make a few implications from them. My over all conclusion below is that harming many is no fair trade to catch one. Zulu Papa 5 * (talk) 20:58, 4 June 2010 (UTC)
 * So invading Iraq to get Hussein was not only botched, but inherently wrong? You've got your logic wrong. If Scibaby has free reign, many more will be harmed - either by increased workload to get his nonsense out of the encyclopaedia, or, worse, by them getting wrong and misleading information. Our aim is not to "catch" Scibaby, our aim is to keep the encyclopaedia as correct as possible. --Stephan Schulz (talk) 21:13, 4 June 2010 (UTC)
 * Tyrants bring accused criminals to punishment without fair justice. Nation leaders apply their sovereign power. Scibaby wreaks his havoc through controlling your hands now. Zulu Papa 5 * (talk) 19:19, 7 July 2010 (UTC)

The real harm to good ratio
What is easier to understand for harm to good, is that up to 20 editors have been harmed for the benefit of disciplining 1 Scibaby sock. That could be w = 20 and way out of range in the previous New Benefit analysis. I would say Scibaby is continuing to wreak harm on Wikipedia, at the expense of his pursuers hands doing the work for him. The down side of terrorism, is the collateral damage folks face with investigatory disruptions. The Scibaby investigation must be better controlled to prevent False Positives. Zulu Papa 5 * (talk) 18:34, 4 June 2010 (UTC)
 * Well, define "harm". Very few accounts who are not confirmed socks have been blocked. Sure, 1 is "up to 20", but then we have also stopped "up to 10^6 disruptive edits". And how do you arrive at "1 Scibaby sock" when we seem to agree that there are 133 of them? --Stephan Schulz (talk) 19:14, 4 June 2010 (UTC)
 * I guess maybe we can rate the harm in each case, since the number is tractable. In the same vein, we must define benefit of keeping Scibaby out. A strength in the Net Benefit calculation is we only have to define a ratio w, where many factors may cancel out. My analysis concluded the harm would have to be about 6 to 8 times greater harm then the benefit before there is negative net benefit.   You may notice the first paper treats harm more with complexity then the second by subtracting the test cost (i.e. community time).  It we include the community time to track Scibaby, we may conclude is it really worth it?  However, I am suspect the folks who do it receive a benefit for the challenge as much as Scibaby does in being pursued.  Zulu Papa 5 * (talk) 21:08, 4 June 2010 (UTC)
 * Somehow I get the feeling that you wouldn't be investing so much interest in defending a disruptive editor if he was pro-climate science... -- ChrisO (talk) 19:25, 7 July 2010 (UTC)


 * @ZuluPapa5: assuming one accepts your contention that the current approach causes more harm than good, then what concrete suggestions do you have for improving our means of dealing with sockpuppetry on these topics? MastCell Talk 21:51, 7 July 2010 (UTC)

Are 'accused parties' notified?
I noticed the lack of responses under the responses section, even when users were found not to be Scibaby. Looking at the user pages of the current allegations, there doesn't seem to be anything there. How does a suspect get informed of the checkuser request? ‒ Jaymax✍ 06:08, 10 August 2010 (UTC)


 * To be honest, usually they don't get informed. Scibaby's socks are so numerous and so relatively easy to spot that notifications are rather a waste of time. -- ChrisO (talk) 07:22, 10 August 2010 (UTC)


 * But this is really the problem. In this and archived pages, there is an instant and automatic assumption of guilt, despite the negatives that DO sometimes come up on checkuser.  I expect that the same assumption is applied by the checkuser folk, because they are used to finding a positive 9 times out of 10.  The end result of all this is a total breakdown of natural justice.  Sure it's only a WP account, but seems to be a systemic issue, and on top of that, the accused is given no opportunity to respond.  Where is the harm in notifying?  The maximum effect it could have would be to take some of his time off the articles, and when disputed, possibly encourage a more objective outcome.  ‒ Jaymax✍ 22:14, 10 August 2010 (UTC)


 * Are you familiar with checkuser? Scibaby cases go one of two ways:
 * There's a positive and definitive link between the accounts. "Guilty". In this case, nothing they can say in their defense will change anything, since the link is incontrovertible.
 * There's no link whatsoever between the accounts. "Innocent". In this case, nothing they can say in their defense will change anything, since they're innocent.
 * Now, in principle, I agree with you that people should be notified of SPI cases raised against them. But in the specific case of Scibaby (which is what we're talking about here), there is no practical purpose to notifying them. --Deskana (talk) 22:27, 10 August 2010 (UTC)
 * I've read the page a few times, and again just now - plus the Extension:CheckUser page. The precise manner in which a checkuser determination is made is not clear, but a degree of experience being required is explicit, and as I read it, a degree of judgement being required is implicit. Now, I might be wrong - if the scope for a 'false positive' (even when an presumption of guilt is present) is exceedingly low, (or non-existent as you suggest) that would certainly provide reassurance!
 * However, tbh, it perplexes me how a 'full-time' sock master could be using accounts which can be successfully tied by IP addresses over an extended period, and not have been somewhat restricted by - for example - IP blocks. Also, presumably there is by now a long chain, with comparison not being made just to the SciBaby account, but the many others accounts and IPs that have been implicated over time. On one hand, I can understand the necessity of not revealing the patterns that are used to positively identify SciBaby socks at checkuser stage; but on the other, without that info being open, it's very difficult for justice to 'have been seen to be done', in the absence of an accused accounts ability to mount a defence.  ‒ Jaymax✍ 01:31, 13 August 2010 (UTC)
 * It seems the analysis above by Deskana and ChrisO MAY be well removed from the true situation. See .  To me, the commentary at ArbCom provides extremely strong evidence of exactly the problem I was worried about - that an assumption of guilt leads directly to an absence of natural justice.  ‒ Jaymax✍ 11:59, 10 October 2010 (UTC)

Tentative suggestion
While I'm hopeless at spotting socks, seems a bit odd and a second opinion would be appreciated. . . dave souza, talk 06:14, 26 July 2011 (UTC)


 * Highly unlikely to be Scibaby, I'd say. Prolog (talk) 08:03, 26 July 2011 (UTC)
 * Thanks, just seemed a bit odd. . dave souza, talk 08:37, 26 July 2011 (UTC)