Wikipedia:Arbitration/Requests/Case/Cold fusion 2/Workshop/Analysis

Place here items of evidence (with diffs) and detailed analysis

Abd's Evidence for consensus

In his present surface-level evidence section [1] Abd states that WMC edited a protected page, "ignoring an expressed consensus." The link to diffs yields a similar claim, "WMC edited Cold fusion while under protection, contrary to full consensus, as shown in [6] and [7], for the version of May 31."

The diffs given to support this "full consensus" link to two different polls, each offering two different versions of the page for editors to vote on. The first link (as you can check if you follow the diffs above) takes you to this poll which gives voters a choice between two versions, the version of 2:54 June 1, 2009, which has two supporters, Abd (10 points) and Krelikraver (10 points), and the version of 3:51 June 1 which was given 0 points by the same two editors, Abd and Krelikraver. There are no other votes for or against either of these two versions. Below this poll is a section labeled "Deprecated proposals, withdrawn by Abd" which lists two versions, the version of 19:01 May 21 which Abd has given 6 points, and the version of 16:51 May 31 which Abd has given 8 points. There are no other votes for or against either of these deprecated and withdrawn versions in this section.

The second link leads to this poll which gives voters a choice between two different versions, which you'll notice are the same two listed under "Deprecated proposals, withdrawn by Abd" in the poll above, the version of 19:01 May 21 which was rated "Acceptable" by Hippocrite and Verbal, and the version of 16:51 May 31 also rated "Acceptable" by Hippocrite and Verbal. There are no other votes for or against either of these two versions, or any other versions, in this poll.

Summarizing the data from these two polls, cited in the diffs given by Abd to show that there was a "full consensus" for the May 31 version: two versions, May 21 and May 31, were each supported by three people (the same three people): Verbal, Hippocrite and Abd, V & H in one poll and Abd in the other, although both of these "consensus" versions were marked "deprecated" and "withdrawn" by Abd in his poll. Sixteen people edited the talk page during the two days of the poll wars; 3/4 of them did not vote in either poll. To assert that these three people represent a "full consensus" that WMC should have honored when reverting the page is not reasonable, even if we ignore the fact that the three people supported two different versions. True, Abd gave the May 31 version two more points than the May 21 version, but we can't count the points because we don't know how many points Hippocrite and Verbal might have assigned the two versions, had Hippocrite chosen to ask people to vote on a point system rather than to designate whether the version was "Acceptable" or "Not Acceptable." If anyone cares, the version that I voted for and the version that Abd moved my vote to are not listed in either of these diffs.

At any rate, the many objections to the way the polling was conducted and the furor that surrounded the polling (illustrated with a few examples in my evidence section)) belie the claim that consensus existed on the page at the time of WMC's edit, which is being portrayed as an edit in defiance of consensus. There was no consensus. Woonpton (talk) 08:38, 28 July 2009 (UTC)[reply]

Comment by Arbitrators:

Comment by parties:

Comment by Abd. Woonpton has made a few mistakes. Bottom line: WMC expected his rollback to May 14 to upset editors per his edit summary. There was no consensus for that version. Set aside the confusion about the polls, created perhaps by my clumsiness at one point, and Hipocrite's disruptive second poll. At the time of WMC's revert there were, on the Talk page, four editors who had accepted the May 21 and May 31 versions. The May 31 version was before the edit warring of Jun 1. In addition, one more editor had accepted Jun 1, which would indicate substantial acceptance (but not preference) for May 21/31, as later confirmed. There was no standing objection to either of the May versions. (And, in my case, only a mild expression of preference for Jun 1 over May 31.) Substantially WMC ignored an expressed consensus. It may easily be claimed that this was not a conclusive consensus; however, poll activity continued for a few more days after the May 14 rollback, and it confirmed the already-visible consensus, plus, in addition, showed dissent for May 14. The rollback to May 14 is not a crucial issue in this case, for it was an improvement over the protected Jun 1 version with Hipocrite's mangled lede. However, it's of some import because I disputed WMC's decision, and was thus in a content or procedural dispute with him when he banned me. Again, overall, minor compared to what happened later. The poll incident, though, shows the background of ready assumptions of bad faith, as well as Hipocrite's disruption; further, the use of range polls for rapid estimation of consensus is worthy of consideration for other situations, it's a small step from the approval polls we already use. --Abd (talk) 12:16, 5 August 2009 (UTC)[reply]

Original reply by Abd, long but detailed, see below for newer version [2] (Collapse box replaced with diff link ☺Coppertwig (talk) 17:26, 8 August 2009 (UTC))[reply]

There is no way that consensus could be determined by looking at those polls, and much less a clear consensus. Most people in the talk page not participating, editors complaining loudly about problems with polls and about Abd handling them. Other editors have explained better the multiple problems with the polls here and in the evidence page. Also, I find that they were being used as a replacement for discussion, with hard counting of votes (not !votes) so WP:NOTDEMOCRACY would enter here. --Enric Naval (talk) 00:20, 4 August 2009 (UTC)[reply]

Enric has confused a process of collecting a picture of current opinion or consensus with the process of making a decision. The poll would be used -- by any or all of us -- to assess if continued discussion is needed. The actual decision would be made by a single admin, presumably upon request, who would review the evidence presented,, including polls, arguments, or anything else the admin chooses to look at. I would not have requested an edit without an obvious consensus, it would be rude to an uninvolved admin to inflict that on him or her. Now, as I show below, there was unanimity on May 21 or May 31, excepting Woonpton, who withdrew her vote, leaving that version with no support, or, if we wanted to pull in the fact that her version was suggested by Kirk shanahan, and count her withdrawn vote, two supports. Compared to four for May 21 or May 31, or five, as we can easily derive from examination of Krellkraver's vote in context. There was no "hard counting of of !votes," there wasn't counting of votes at all, at that point. We count them now as a measure of consensus at the time, but simply looking at the expressed opinions leads us easily to unanimity for May 21 or May 31, it matters little which one. I would still have waited more time to allow more expression before making a specific proposal. So Enric Naval is wikilawyering, avoiding the substance here for a technical argument about democracy.

As a substitute for discussion, may be a reference to my request that we not discuss the versions in the "Reserved section" for the poll; but a place was provided for discussion, and nothing was in the way of anyone starting up discussions of the differences elsewhere, and a final proposal to an admin would certainly not repress discussion. My priority was for rapid decision on a version to accept, temporarily, to deal with the Wrong Version protection, and in a normal collaborative environment, this would have easily been done with what I proposed. See below, my response to Bilby, for a detailed response. --Abd (talk) 23:49, 4 August 2009 (UTC)[reply]

I'm not confusing anything. The poll was flawed and confusing, many editors hadn't participated and others had directly rejected it, you handled the whole thing very badly, and you didn't address any of the problems that were pointed out to you. There is no way that an admin could take that poll as an evidence of consensus in the talk page.

And, by the way, dude, you have some nerve to say that the poll reflected consensus because it had more votes a few days later. You should consider that it was partly because, as pointed out by Bilby, you had copied from other poll the votes of two editors who said that they didn't want to participate in your poll. --Enric Naval (talk) 10:03, 5 August 2009 (UTC)[reply]

The votes days later confirmed what was already visible, which also confirmed the state of the article May 31 (not in immediate controversy). Those editors didn't "participate in my poll," sure. But the purpose of "my" poll was to collect information about consensus. They provided that information in "their" poll. I cited it in "my" poll so that all information provided on the topic could be seen in one place. Any admin making a decision, if it were necessary to review those polls (it might not have been if the specific question asked based on the preliminary results of the polls), would surely have lookedd at both polls and all the provided evidence, and any deception in my reporting would have been noticed. I did not "copy" votes, I reported them. Explicitly as such. I was being advised that I should include diffs, and, had I not been banned, I would have. Is Enric really claiming that it's not allowed to report accurately and neutrally what other users have expressed?

It can be argued that by reporting the approvals in the other poll, I translated them to "10," i.e., max rating, and that this was arbitrary. Any other translation would have weakened their "virtual representation" in the range poll. 10 gives maximum strength to their expressed opinion, and is the least synthetic translation. I could have simply reported them as "Approved" or "Accepted." In the description of how one might vote in the range poll, it was suggested that one vote 10 as an equivalent of "approved" or "accepted," with the option of using lower ratings to indicate relative preference. All this was designed, had it simply been accepted, to rapidly estimate consensus, and the disruption of the second poll did not actually make it fail in this purpose, it merely made it more complicated to assess.

Bilby correctly points out, below, that ratings in a range poll may shift depending on the options. That's true, but moot in this case. Hipocrite and Verbal clearly preferred May 21 and May 31 to any of the versions that were in the Range poll at the time of the creation of the second poll. Each would have had a favorite (the differences are minor and noncontroversial). One would not lower the rating of the favorite because of the presence of other options, unless they were preferred over the favorite. The only relevant version in this respect was June 1, and Hipocrite had edit warred to keep part of that out. At most what would have happened would that *one* of the two, May 21/31, would have received a somewhat lower rating. The implications of this in terms of overall understanding of consensus: minor to none. Essentially it could have affected the choice between May 21 or 31 only. Interpreting range polls can be tricky, but, in fact, normally it is not. For example, Range polling, used to make a decision, does not satisfy the Majority criterion, which offends some people. However, it is only in unusual circumstances, where the wisdom of accepting the majority preference is questionable, that Range polling violates the criterion; if the majority preference doesn't win in a range poll, it is because the expressed preference of the majority is weak and that of the minority is strong. (Which can happen, as an example, if the minority is expert compared to the majority.) In any case, my own position is that, if possible, range polling should not be used definitively to make a final decision, it should be used to efficiently propose a solution likely to gain consensus on the first ratification poll, approved by a maximized majority as an answer to a Yes/No/Abstain question, and, on Wikipedia, the ultimate decision is made by a single individual (closing or acting admin or other editor), and this is only untrue before ArbComm, which makes decisions by majority vote. Where the result of the range poll is obvious, however, the second poll may not be needed, and, I submit, that was true here and remained true. It was already visible, when WMC reverted to May 14, that he could have reverted to either May 21 or 31 and it would have been non-controversial, which is what is expected of edits under protection. But picked a version that he, accurately, expected would arouse objection.

Reporting a vote means saying "X has voted for this option in a different poll" in different format from the other votes, it doesn't mean formatting their vote as if they had voted themselves, assigning an arbitrary weighted value to the their support and then writing at the side that you had reported their vote from other poll[3][4]. And most certainly you should have mentioned in the reporting itself that they didn't want to participate in the poll. And it doesn't mean you saying that your poll reflected consensus when it's partly reflecting what you thought that other editors would have voted if only they didn't think that the poll should be closed with prejudice. And don't mention that they had voted in the other poll (which was basically competing with your own poll) while at the same time refusing to vote in your poll which was in the same page. And the "they supported this version but not the other one" is quite weak when there were options on one poll that didn't appear in the other one. And all of this after the incident with moving Woonptoon's vote and the complaints about you moving people's votes in their behald. And you have been told this several times but you still cite your poll as if it had no flaws and anyone could take it at face value. (and I think that you were also told that it was not wise to assign them a "10", which would translate as the strongest support, when they had originally voted under a section that only said "Aceptable"? Well, you have been told now.) --Enric Naval (talk) 22:32, 5 August 2009 (UTC)[reply]

(Abd has replied here, in last paragraph)

As above, your process was flawed, the poll itself was flawed, your handling of the poll was flawed, you counted votes of editors who didn't want to participate, and you did it in a way that might have mis-measured their support. No, Abd, the purpose of measuring consensus was not achieved, and it's not going to be achieved by asserting it all over again and then blaming all the problems with the polls on Hipocrite. --Enric Naval (talk) 22:46, 7 August 2009 (UTC)[reply]

Comment by others:

Brief response to Abd. I can't read this whole thing, but I'll reply to just one or two things my glance caught as I quickly scanned down the vast length of this post:

my analysis is strictly of the evidence Abd provided in the form of two diffs to support the claim that WMC had edited "contrary to consensus." My description and analysis of those two diffs is accurate, as people can see by looking at the diffs themselves), and shows that the consensus claimed is not supported by the diffs. Abd's response to my analysis is at least twice as long as the analysis itself, and on a quick scan it seems that it doesn't confine itself to a rebuttal about those two diffs, but covers a great many things, including the insinsuation, apparently intended to cast doubt on my analysis, that I'm a "skeptic" (I'm not sure what that's supposed to mean, but I am a scientist by nature and training; I have a solid science background; I am a retired statistician. The science of drawing conclusions from evidence, of teaching people not to draw conclusions that aren't supported by evidence, has been my life's work). If that qualifies me for the cabal, then I'm proud to be a member. But that has nothing to do with the fact that the diffs Abd provided as evidence to justify the claim that there was a consensus, do not support that claim. If Abd has better diffs to support the claim, he should provide them. I'm done here.Woonpton (talk) 17:55, 29 July 2009 (UTC)[reply]

Comment by Bilby - I don't think that the Arbs will have much trouble forming their own opinions after they view the polls. However, in the interests of clarity:

Abd commented "I did not oppose Hipocrite's poll" - I'm afraid he did. He stated, immediately after Hipocrite started a new poll, "I object to this new poll as splitting and confusing work on this issue."
You cannot translate someone's support in one poll to a numerical value in the other. Because a) one poll had more options than the other, and b) support in a non-ranged poll does not equal full support in a ranged poll.

In all honesty, I still can't see how any consensus can be drawn out of the mess that those two polls made. - Bilby (talk) 18:30, 29 July 2009 (UTC)[reply]

Original response by Abd to Bilby, see later response below to his more detailed objection [5] (Collapse box replaced with diff link ☺Coppertwig (talk) 17:39, 8 August 2009 (UTC))[reply]

I'm a tad surprised that you seem to continue to think it is ok to translate votes from a smaller approval poll to a larger range vote. In case it helps, my stance is:

I can't see how you can translate votes from a smaller poll to a larger one. As a quick and dirty example: if I give you a choice between chocolate and icecream, and you pick chocolate, I can't assume that you would make the same choice when given a choice between chocolate, icecream and marshmallows. (I'd pick chocolate in the first, and marshmallows in the second). In this case, Hipocrite's poll had only two options. Yours had seven. It is incorrect to assume that someone's choice on Hipocrite's poll would remain the same if they had chosen to vote on yours.
Approval polls cannot be reasonably translated to range polls. You assumed that if Hipocrite and Verbal had voted in your poll, they would have given a full 10 points to each of the items they voted for in Hipocrite's poll. You cannot make this assumption. The other direction may be possible (10 points in a ranged poll = support in an approval poll).
The big one is simple - you can't vote for other people, even if believe that it is how they would have voted. Hipocrite and Verbal did not vote in your poll. You placed votes for them, based on how you believed they would have voted, using evidence from an alternative, but not identical, poll. There was no need to do this, and it was, to me at least, wrong. I'd be very surprised if ArbCom saw this differently, but they might, like me, want to assume that you had good intentions.

This is an important issue, because you argued that WMC edited against consensus. I would argue that he didn't edit, but reverted to a pre-edit war state, as per policy, and that there was no consensus to edit against. Nor any probablity of finding one as things were going. In addition, when WMC reverted only two editors had voted in each poll - to few to be meaningful. - Bilby (talk) 00:50, 31 July 2009 (UTC)[reply]

In context, consensus was apparent, and also WMC's neglect of consensus. I believe I've already stated why. Yes, in general, the problem with comparisons as described in the example by Bilby exists. It doesn't apply to this situation. Here's what can be done to show that consensus was apparent: Look at my poll. The votes standing as of WMC's revert to May 14 showed, versions in reverse order of date:

June 1 (as protected, after edits by Hipocrite

Abd 0

Krellkraver 0

Jun 1 (mid war, justified as a possibility there):

Abd 10

Krellkraver 10.

Then I had added the versions from Hipocrite's poll to mine, so that I and others could express Range opinions of them.

May 31

Abd 8

May 21, the prior protected version.

Abd 6

At this point, Krellkraver had not returned. He voted the first time two days before Hipocrite had created his poll. So what can we derive from this poll? It's clear that Abd prefers both May 21 and May 31 to the version as protected, prefers the midwar version where Hipocrite had appeared to accept one of the changes (he later denied this, but, obviously, he didn't consider the last change important enough to try to modify it to make it acceptable, preferring instead to make an edit to the lede that he absolutely knew would not be acceptable, and which would not stand unless protected.), and is inclined to accept the version of May 21.

Setting aside Krellkraver, who had not voted on the versions Hipocrite proposed, we can see two versions supported by Abd.

Then we look at Hipocrite's poll. He has proposed two versions, May 21, the prior protected version, and May 31, and Verbal also approved both of these.

Then V had commented in Hipocrite's poll, "I don't object to either of the two earlier versions."

Likewise, in the discussion of Hipocrite's poll, Woonpton objects to the removal of the version she had transiently voted for, but if her vote had been considered, even though she withdrew it, it would not have altered any reasonable conclusion at that point.

Setting aside Krellkraver, we have Abd clearly supporting May 31 and also showing what looks like acceptance of May 21. Clearly Abd prefers May 21 to the version as protected. We have Hipocrite and Verbal approving May 21 and May 31, with V accepting them as well. Nobody (except perhaps Woonpton) has rejected either of May 21 or May 31. That looks like a preliminary consensus to me. 4 approving, no !votes standing opposed.

Now what about Krellkraver? I set up the matrix so that any two versions could be compared quickly. Krellkraver preferred the midwar version, June 1. The diffs:

As can be seen, the differences are small, confined to the most controversial sections, which went to mediation, in fact. Remember, Hipocrite and Verbal had accepted both May 21 and May 31. Edit warring began over the differences between May 31 and Jun 1; and my basis for proposing Jun 1 as a possibility was that Hipocrite had back to the Jun 1 version. (He later claimed that this did not indicate acceptance, it was merely reversing a slip of the mouse). However, he's the one who requested protection, and his revert to the Jun 1 version almost certainly represented a decision that he was at risk of violating 3RR; had he not requested protection, he'd have had to let this content stand. If he had not reverted himself, he'd not have a revert "in reserve," in case he needed it before the protection hit. Instead, he gamed the system. Adding new content, no matter how outrageous, isn't going to be considered edit warring, is it? Even though he clearly knew that it wouldn't be accepted. So he made a major edit to the lede, clearly without consensus. He also made it difficult to notice in the edit summary, but put enough there to make this deniable. Sure enough, GetLinkPrimitiveParams reverted one of Hipocrites earlier removals; but Hipocrite didn't spend his revert on it, instead he added "balancing" original research, actually untrue. (Sure, the patent doesn't mention "cold fusion." It mentions generation of energy by electrolyis of heavy water with palladium electrodes. Is that "cold fusion"? Consensus is that we call this "cold fusion" on Wikipedia, whereas in the field it's called "low energy nuclear reactions.")

The point is that Krellkraver, if he or she preferred Jun 1 to the other versions, would surely accept May 31, knowing that this would maximize consensus, and probably May 21 as well.

For reference, the difference between May 21 and May 31.

Truly minor differences. The changes that remained from the edit warring on May 21 appeared to be stable, and could tentatively be presumed to be accepted by consensus. When the article was protected on May 21, I praised WMC for the protection, and that was sincere. Nobody objected to that protection. Yet it wasn't my "preferred version," most of my work May 21 had been reverted out. Essentially, there was one section that had been accepted by Hipocrite, and he'd added reasonable balancing sourced material, to which I had no objection.

So: we had no editor objecting to May 21 or May 31. We had four editors explicitly accepting May 31, or May 21, or five editors, if we infer Krellkraver, pretty much a slam dunk. (And later confirmed).

I call that "full consensus," without any implication that it couldn't change with broader participation. It didn't change. It just got clearer.

So why did WMC revert to May 14? We can speculate, given the cabal affiliations. He knew, we can assume, that Hipocrite had edit warred to keep out the changes of May 21 at first, and he would know that he generally supported Hipocrite's content positions, they were aligned with ScienceApologist (except SA is a physicist, Hipocrite showed no sign of understanding the content at Cold fusion, and, long-term, had no real interest in the article; Hipocrite was merely supporting the anti-pseudoscience agenda of SA, with which WMC is allied). He might have looked at the diffs and saw that hydrino theory had been accepted May 21 for inclusion. Hydrino theory is commonly considered pseudoscience or even fraud; but it's notable, with positive and negative reliable source. Still, the cabal routinely acts to exclude pseudoscience on the argument of undue weight, WMC's instinct would be to exclude it.

GoRight had come along and looked at the edit history and saw that May 14 had remained for a while, thus seemed more stable. So he proposed it for discussion. Clear demonstration that if the Global warming cabal, or SA cabal, or whatever we name it, isn't organized (which is generally true), mine is even less organized. Hipocrite, of course, jumped for it (as did Verbal later). So WMC had some cover for May 14. He could take GoRight's proposal of May 14 as "support" for it, which caused him great amusement, because of his long-term conflict with GoRight. So he had two editors. Was that consensus? Compared to four or five? Again, WMC grabbed the opportunity, but, as usual, he was charmingly frank. He knew what he was doing was controversial, and so his edit summary was: (Lets wind everyone up). He edited the article under protection, expecting it to be controversial. All the criticism of the polls is fluff, compared to this.

There was some legitimate objection to my poll as confusing, but in a normal environment that would have been addressed easily. It was not a normal environment. The confusion in interpretation of preliminary results was created by Hipocrite, more than two days after I'd started my poll, setting up a new one on the same topic, with restricted and new choices. He could have added his choices to my poll, that had been suggested, and he could have expressed exactly the same opinions, fully approving his two proposed versions, and being silent on the others, or voting in any other way. But Hipocrite was there for the very specific purpose of creating disruption, continuing what he'd done when I first encountered him, I'll put evidence on this up. I was his goal or target, not Cold fusion.

However, if it wasn't just a pure disruptive intent, why would Hipocrite not want to vote in my poll, besides the cabal knee-jerk opposition of what I do? Well, the poll included the version as protected June 1. His version. It would be glaringly obvious if he had avoided voting for it, but voted for other versions in my poll. Was this his reason? It's not like the idea of rating alternatives from 0 to 10 is a difficult or uncommon one! (And range voting has, indeed, been used for a major decision on Wikipedia.) In our traditional discussions, anyone can propose a new alternative at any time, it happens frequently, so the objections to additional options is a red herring. And setting aside -- not deleting, just putting in collapse, inviting anyone who wanted to do so to bring them back out -- versions with no expressed support was just trying to make it all less confusing, editors were already complaining about how big the matrix was getting.

It would have been up to whoever made the final decision how to interpret all this, and I certainly wasn't trying to change that. At some point, I expected, the consensus would be obvious and we would have requested review at RfPP. It was already obvious, from the existing votes, with some ambiguity around May 21 v. May 31. I can assure all here that I wouldn't have objected to May 21, even if, as my !vote shows, I had a slight preference for May 31. However, I would have allowed a little more time for editors to show up. At the time I was banned, it was even more obvious.

No, this mess was created by the cabal, through a very disruptive supporter, Hipocrite, and it then gave WMC cover to do what he'd long wanted to do. Begin the process of banning Abd. --Abd (talk) 23:32, 4 August 2009 (UTC)[reply]

Based on what you said above:

You pointed out that WMC reverted when the polls were in this state. At that stage your poll had been running for three days and had only received 2 voters (if we discount the removed votes by Woonpton). Hipocrite's poll had been running for one day, and had only two voters as well, along with one general approval. This is insufficient time and involvement to be used as any sort of guide to consensus - even if someone tries to combine the results.
Reverting to a pre-edit war version, as suggested by GoRight, is explicitly permitted in the policy, for exactly the situation you describe - so as not to reward one of the edit warriors. If what you say is correct, (and I think it is), Hipocrite had the page protected in his preferred version. WMC fixed that by reverting to an earlier version.
You still don't seem to accept the core point - you can't vote for someone else based on how they voted in a non-identical poll. You placed votes for other people in your poll, based on how you thought they would have voted (noting that this was done after WMC reverted, but before you were banned). I accept you believed you were doing the right thing, and that you justify it - I just don't believe that it was right or that the justification holds. I note that GoRight also commented on how this was thin ice, and that both Verbal and Hipocrite (whose votes you placed in your poll), had explicitly stated that they did not want to be involved in your poll.

That said, we're probably just arguing in circles. ArbCom has the evidence and the arguments, and I suspect they're able to come to their own conclusion on how relevant this is and what happened. - Bilby (talk) 00:45, 5 August 2009 (UTC)[reply]

insufficient time. Sure. However, the unanimous agreement of all editors who have expressed a decipherable opinion routinely passes for consensus. Further, the real point here is that there was such an agreement, with nothing contradicting it that was of any clarity, and that it was ignored by WMC in favor of doing something provocative. It's blatant, Bilby.

reverting to a pre-edit war version. Sure. That, in fact, is what May 31 was (likewise May 21). There was no edit warring over what had been accepted May 21. May 14, picked by WMC based on the suggestion of GoRight -- he later said that he was just proposing it as a place to start discussing, not because he considered the version itself -- represented undoing consensus toward what was the result of sustained revert warring by Hipocrite, it merely had not been confronted, until May 21. June 1, I made one edit to reflect some level of compromise from discussion of the remaining changes that had been proposed May 21, with additional sourcing. That's what Hipocrite edit warred to keep out (not I). The "pre edit war version" would have been May 31. Nothing in the period May 21-May 31 was particularly controversial.

you can't vote for someone else. That's correct. And I didn't. For purposes of analysis, I wrote exactly what I was doing, it was explicit. I didn't "vote for someone else," I reported how others had voted in a parallel poll. Neutrally and according to a stated translation (which I can defend, easily, as the most reasonable choice). It was a running analysis. If somebody didn't like that, they could revert it, change it, or disregard it. This was just a poll, Bilby, not a decision-making process. The decision was to be made by an RfPP admin on request, and, my assumption, the request would have been endorsed (or opposed), entirely independently, as a separate question. I'm glad that you accept my good faith in this, and I fully assure you and ArbComm that my intention was to discover true consensus, I'd be shooting myself in the foot if I did anything else there, it would surely backfire. The issue for me wasn't "my poll" or Hipocrite's poll, but what the editors were willing to approve. A range poll was used because it's more efficient, better at predicting what voters will approve if asked specifically about a reduced set. It's not perfect, merely better than Approval or no worse than Approval. If we want to know editorial consensus, we look at what the editors are expressing. The poll technology is a detail.

Arguing in circles. If we continued, yes, and that may have started. Thanks for bringing up the issues, if you were thinking these things, so might be someone else. You are right, ArbComm will decide. As I've said, it's a relatively minor point here, the elephant in the living room is WMC's insistence on his right to stand judge over my editorial behavior, even now, after all this. I've bothered defending this point because

Range polling should be considered more often, it's more efficient in discovering or maximizing consensus, and that's what I propose it for, not decision-making.
The creation of the second poll more than two days later was deliberatively disruptive, designed to confuse the issues and thereofore to delay unprotection, because Hipocrite preferred the article as protected even if he wasn't willing to be explicit about that. Verbal, as would be expected, played along within minutes.
I was charged with bad faith for how I handled the poll, upon first excuse, "Shenanigans." --Abd (talk) 01:51, 5 August 2009 (UTC)[reply]

WMC's involvement and support for his rollback to a FA version

About WMC was an involved editor on Cold Fusion in January 2006..

Short summary: WMC was simply implementing a clear consensus of stablished editors against edits that hadn't consensus and which were made by advocates on the topic. There was also consensus that the edits of those advocates had worsened the article and were the reason for the rollback. Most of his edits in the talk page where about the same matter. After that matter was solved WMC didn't edit the article or its talk page any more until banning Abd.

In reverting to a 15 months old version, he was reverting back to the FA version. As his announcement in talk page says, the still-ongoing FA review had several editors saying that the current version was so bad that it should be rollbacked (like WMC later did): Taxman [6] FrancisTyers [7] Henry Flower [8][9], Causa Sui [10] Just a Tag [11][12], Anville [13], Pjacobi [14]. Geogre said that the "revert or not" question belong to somewhere else [15]. WMC himself expressed his support to the rollback[16] and announced it there [17]

Many cited concerns that at least one CF advocate had been pushing his POV so strongly that it was impossible to get the article back to a neutral state, much less to FA status, and that there were no skeptics working on it [18]][19][20][21] (this was back where User:JedRothwell was still editing actively the article, and User:Pcarbonn hadn't arrived yet) Joke137 made what I find to be an accurate summary of the problem: lack of any high-quality source and presence of two motivated CF advocates (JedRothwell and Obsidian) [22]

One editor asked what version should they rollback to [23], other said that the "revert" comments wee unproductive[24] he clarified his question [25] and he was told that they meant that it would only be kept as FA if it was reverted back and that they were useful because they showed consensus that the FA version was good and the current one wasn't [26]

There were even concerns that CF advocates would just revert it back [27].

There were editors opposed to the FA removal, we can assume that they also opposed the rollback [28][29], and an editor saying it didn't matter becasue it failed to show that it was fraud [30] Those editors included JedRothwell[31] and a few IPs and editors with very few edits [32][33][34][35][36][37][38][39] (it would seem that the FA removal was mentioned at some mailing list or something)

By the time of WMC's revert of the article (20:07, 3 January 2006) there was clear consensus that the current version was very bad, and enough consensus to go ahead and revert directly. WMC's revert was undone, and it was restored by Taxman "this is the consensus (FA) version for now. There is a strong consensus that changes since this have damaged the article" [40]

Other edits by WMC were removals of the {{totallydisputed}} tag that was being repeatedly added by User:ObsidianOrder, the editor who had tried to undo the rollback to the FA version. [41][42][43][44]. Other editors also removed that tag[45], added the tag [46][47][48][49][50][51]

WMC's revert to FA was heavily discussed in the talk page:

replies to his announcement here. Hum... I'll have to make a value judgement here: Obsidian was wrong about the support in the FA reciew, and the rest was arguing by Jed
Jed arguing that the previous version was better here
justifications of the totallydisputed tag here, continuing here
RfC on the removal here, no real outside participation

Other edit was adding two Nature references that he had found [52].

A few days after the rollback WMC was proposed for adminship, it was even announced in the talk page by an IP[53]. He passed it and only two opposing votes mentioned the rollback in cold fusion.

Comment by Arbitrators:

Comment by parties:

Proposed. For supporting William M. Connolley an uninvolved administrator with respect to Cold fusion --Enric Naval (talk) 17:44, 8 August 2009 (UTC)[reply]

Wikipedia:Arbitration/Requests/Case/Cold fusion 2/Workshop/Analysis

Abd's Evidence for consensus

WMC's involvement and support for his rollback to a FA version

Template

Template

Template