Wikipedia talk:Reflections on RfX/Archive 3
![]() | This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page. |
Archive 1 | Archive 2 | Archive 3 |
Crat assumptions
I was bold and rewrote this section since I had a similar idea, but the previous version wasn't fully fleshed out.
An interesting part of the Crat Chat for Liz's RfA is what the starting assumptions of the crats should be when determining consensus. (See here). Namely, is consensus assumed and the weight of the opposes need to show that there isn't consensus, or is no consensus assumed and supports need to show that there is consensus?
Though set up similarly, this is not a support/oppose/neutral poll. Support of any shows that you support a discussion about crat assumptions. The divisions are simply to help figure out how to frame that discussion after this brainstorming round.
Consensus is Assumed
Because a candidate must be incompetent or malicious to cause problems with the tools, consensus should be assumed and opposers need to show that the candidate cannot be trusted or will be incompetent with the bit.
Support
Oppose
- The claim above is simplistic and unsupportable, and the record of those officially desysopped does not support it as a blanket claim. "Cannot be trusted" is a gross exaggeration, and "incompetent" is also a straw man here. The issue is generally "will they act fully in accord with all policies and standards of Wikipedia from our direct ascertainable knowledge." If our direct ascertainable knowledge had major holes in it, then we can certainly oppose. Collect (talk) 12:10, 9 August 2015 (UTC)
- I don't see it this way. It's a discussion, and there really isn't a starting consensus, because consensus should emerge as the process goes along. There can be an assumption of WP:AGF, but not an assumption about the outcome. --Tryptofish (talk) 21:33, 9 August 2015 (UTC)
- If it's assumed than anyone with X years of experience and Y number of edits should automatically get the bit. That may not be bad, but it's not the assumption. Quite the opposite. We assume they need a certain level of approval. --DHeyward (talk) 21:17, 17 August 2015 (UTC)
- Ah, no. Per Collect. GregJackP Boomer! 18:43, 25 August 2015 (UTC)
Consensus is not Assumed
Because of the powerful tools involved and potential for problems, no consensus should be assumed and supporters need to show that the candidate can be trusted with the bit.
Support
- Sort-of Support First let me explain that the way I understand it is that the discussion/RfX must demonstrate consensus. Unless there is a consensus of respondents, the candidate does not receive access to the toolkit. This means that the status quo, the default assumption, is that the candidate continues as they are.
- That being said, I think that both these sections are a bit of a generalization of the discussion I had with Andrevan. I don't think any of us believe that the very fact that someone posts their candidacy means that they are auto-assumed to be an admin unless actively refused. Rather, in the case of Liz, to me it seems (and I may well be wrong) that Andrevan looked at the supports first, and felt them to be indicative of a consensus, and then looked at the opposes to see if that consensus was materially weakened or damaged, whereas I looked at both the supports and opposes in one block and was trying to determine whether a consensus arose from the combined discussion. That is what I believe he meant by "…this is a hair-close case that is on the edge of being an uncontroversial promote, we only need to demonstrate the minimum level of counter-weight to the oppose arguments to find a consensus here." Not that there was an assumption of passing ab initio but that the 200+ supports created the consensus and now it was the opposes "job" to counteract that. I think that since the candidate starts with a tabula rasa, and since the purpose of the entire RfX is to determine if the candidate has gained the project's trust, that the entire discussion must be looked at as one unit. I think that is the core of our disagreement as linked to above, and I think that is a more subtle, yet important question on which I for one would like to see the project's collective opinion. Thank you. -- Avi (talk) 06:44, 11 August 2015 (UTC)
- I have a feeling that that discussion between you and Andrevan was a significant part of the reason why we are having the discussions now at this page, so it's worth going into it some more. I saw that discussion as it occurred, and I actually was puzzled about why the two of you were even discussing it. I agree with you that candidates always start with a sort of a tabula rasa, to the extent that whoever closes the RfA should not work from any sort of assumption about promote or not-promote being the presumed default. That is part of why I'm not enthusiastic about the proposals about assumptions here. But I also think that the closer might sometimes see either the supports or the opposes as having made a particular case in the discussion, and then the closer might quite reasonably examine whether that case had or had not been adequately rebutted. Consequently, I understood Andrevan as having actually approached it with the same tabula rasa as you – but having felt that there was a case made by the supporters, he was examining whether or not the opposers had rebutted it. I think that the two of you might have misunderstood each other, over the issue of a "baseline assumption" that, in fact, neither of you had made. Maybe I'm wrong, but that's what it looked like to me. --Tryptofish (talk) 23:37, 11 August 2015 (UTC)
- Yes, Tryptofish. I think the difference lies in that Andrevan made his decision in two stages: 1) about the supports absent the opposes and then 2) do the opposes counteract the supports. I made my decision in one stage: Do the combined supports and opposes show consensus? Perhaps my analogy about priors was not only rough, but misleading -- Avi (talk) 17:30, 12 August 2015 (UTC)
- I think I better let Andrevan speak for himself, but I can imagine various ways to think about "counteract". It could mean giving greater weight to the supports than to the opposes, which is undesirable, but it doesn't have to mean that. --Tryptofish (talk) 21:28, 12 August 2015 (UTC)
- The basis of support comes from having the requisite supermajority. It's not a baseline assumption of consensus. The baseline is having ~75% support. Andrevan@ 14:40, 13 August 2015 (UTC)
- Hi, Andrevan. So do you approach each of these by deciding quantitatively prior to qualitatively? I ask this as you state that the "baseline is having ~75% support" and it looks as if that checks a box off in your mind prior to your analysis of the nature of the support (as not every 75% is the same). -- Avi (talk) 15:56, 13 August 2015 (UTC)
- I'd also like to ask what, precisely, "the requisite supermajority" means. It's something I'm not comfortable with, on the face of it, per WP:VOTE. --Tryptofish (talk) 17:12, 13 August 2015 (UTC)
- On WP:VOTE it specifically mentions the 70% rule of thumb. Other ways of stating it in the past were "about 2/3", etc. But yes, so I think there's a quantitative spot test before the qualitative weighing. To quote my own RFB from 2007: What about the numbers, the percentages, all that pseudo-voting mumbo jumbo? Well, the numbers are confusing to some, but basically they represent the quantitative aspect of the promotion process. They are a rule of thumb to judge the level of agreement. 80% is just a jumping-off point to demonstrate the level of acceptance needed, but consensus is not contingent necessarily on numeric percentage. A bot can't say, "77% of editors expressed their support, but everyone generally agrees that so-and-so is qualified, and oppose comments were fairly noncommittal, so I will promote." The human element is needed for this sort of analysis. This is the job of a bureaucrat. Andrevan@ 11:39, 14 August 2015 (UTC)
- So, if I understand you correctly, Andrevan, our discussion during the cratchat was actually a misunderstanding, and where we differed is simply that I did not think the 280 responses as a whole showed consensus and you did? -- Avi (talk) 15:57, 14 August 2015 (UTC)
- I'm not sure I would say our whole discussion was a misunderstanding. However, you stated you felt I had assumed consensus existed before I evaluated the oppose arguments, which I at the time tried to explain was not quite right. While there is certainly a lesser burden on support arguments since they can largely pile on, they also have a much greater burden in the sense that they must find a substantial though not unanimous supermajority and not simply a 50%+1. So in that aspect I think your talk of Bayesian priors was off the mark. We did, though, have a substantive discussion on the nature of the oppose arguments which I think was definitely helpful and not a misunderstanding. Andrevan@ 05:11, 15 August 2015 (UTC)
- So, if I understand you correctly, Andrevan, our discussion during the cratchat was actually a misunderstanding, and where we differed is simply that I did not think the 280 responses as a whole showed consensus and you did? -- Avi (talk) 15:57, 14 August 2015 (UTC)
- On WP:VOTE it specifically mentions the 70% rule of thumb. Other ways of stating it in the past were "about 2/3", etc. But yes, so I think there's a quantitative spot test before the qualitative weighing. To quote my own RFB from 2007: What about the numbers, the percentages, all that pseudo-voting mumbo jumbo? Well, the numbers are confusing to some, but basically they represent the quantitative aspect of the promotion process. They are a rule of thumb to judge the level of agreement. 80% is just a jumping-off point to demonstrate the level of acceptance needed, but consensus is not contingent necessarily on numeric percentage. A bot can't say, "77% of editors expressed their support, but everyone generally agrees that so-and-so is qualified, and oppose comments were fairly noncommittal, so I will promote." The human element is needed for this sort of analysis. This is the job of a bureaucrat. Andrevan@ 11:39, 14 August 2015 (UTC)
- I'd also like to ask what, precisely, "the requisite supermajority" means. It's something I'm not comfortable with, on the face of it, per WP:VOTE. --Tryptofish (talk) 17:12, 13 August 2015 (UTC)
- Hi, Andrevan. So do you approach each of these by deciding quantitatively prior to qualitatively? I ask this as you state that the "baseline is having ~75% support" and it looks as if that checks a box off in your mind prior to your analysis of the nature of the support (as not every 75% is the same). -- Avi (talk) 15:56, 13 August 2015 (UTC)
- The basis of support comes from having the requisite supermajority. It's not a baseline assumption of consensus. The baseline is having ~75% support. Andrevan@ 14:40, 13 August 2015 (UTC)
- I think I better let Andrevan speak for himself, but I can imagine various ways to think about "counteract". It could mean giving greater weight to the supports than to the opposes, which is undesirable, but it doesn't have to mean that. --Tryptofish (talk) 21:28, 12 August 2015 (UTC)
- Yes, Tryptofish. I think the difference lies in that Andrevan made his decision in two stages: 1) about the supports absent the opposes and then 2) do the opposes counteract the supports. I made my decision in one stage: Do the combined supports and opposes show consensus? Perhaps my analogy about priors was not only rough, but misleading -- Avi (talk) 17:30, 12 August 2015 (UTC)
- I have a feeling that that discussion between you and Andrevan was a significant part of the reason why we are having the discussions now at this page, so it's worth going into it some more. I saw that discussion as it occurred, and I actually was puzzled about why the two of you were even discussing it. I agree with you that candidates always start with a sort of a tabula rasa, to the extent that whoever closes the RfA should not work from any sort of assumption about promote or not-promote being the presumed default. That is part of why I'm not enthusiastic about the proposals about assumptions here. But I also think that the closer might sometimes see either the supports or the opposes as having made a particular case in the discussion, and then the closer might quite reasonably examine whether that case had or had not been adequately rebutted. Consequently, I understood Andrevan as having actually approached it with the same tabula rasa as you – but having felt that there was a case made by the supporters, he was examining whether or not the opposers had rebutted it. I think that the two of you might have misunderstood each other, over the issue of a "baseline assumption" that, in fact, neither of you had made. Maybe I'm wrong, but that's what it looked like to me. --Tryptofish (talk) 23:37, 11 August 2015 (UTC)
- That being said, I think that both these sections are a bit of a generalization of the discussion I had with Andrevan. I don't think any of us believe that the very fact that someone posts their candidacy means that they are auto-assumed to be an admin unless actively refused. Rather, in the case of Liz, to me it seems (and I may well be wrong) that Andrevan looked at the supports first, and felt them to be indicative of a consensus, and then looked at the opposes to see if that consensus was materially weakened or damaged, whereas I looked at both the supports and opposes in one block and was trying to determine whether a consensus arose from the combined discussion. That is what I believe he meant by "…this is a hair-close case that is on the edge of being an uncontroversial promote, we only need to demonstrate the minimum level of counter-weight to the oppose arguments to find a consensus here." Not that there was an assumption of passing ab initio but that the 200+ supports created the consensus and now it was the opposes "job" to counteract that. I think that since the candidate starts with a tabula rasa, and since the purpose of the entire RfX is to determine if the candidate has gained the project's trust, that the entire discussion must be looked at as one unit. I think that is the core of our disagreement as linked to above, and I think that is a more subtle, yet important question on which I for one would like to see the project's collective opinion. Thank you. -- Avi (talk) 06:44, 11 August 2015 (UTC)
- Support This is why A) the consensus threshold is high and B) a large number of supports doesn't need justification (i.e. 80% is enough that nothing more than "Support" is necessary. Below 70% and there must be extraordinary support justifications. --DHeyward (talk) 21:19, 17 August 2015 (UTC)
- Support. So long as people are going to harass those who oppose a candidate, those who support must show justification for that support. GregJackP Boomer! 18:45, 25 August 2015 (UTC)
Oppose
- Demur - this is the other half of a straw man position - it is up to a consensus to determine whether the community has confidence that the person will properly abide by all policies and standards - not to require "proof" of any sort. A significant number of folks have been desysopped, and the record is that they in general were neither found to be "incompetent" nor "untrustworthy" but that they did not abide by Wikipedia policies and standards. Collect (talk) 12:14, 9 August 2015 (UTC)
Unsure, but this is a discussion the community should have
As the assumptions made can have an affect on the outcome of a crat chat and RfA, the community should discuss how crats approach consensus finding in RfAs
Support
- (as nom) Wugapodes (talk) 04:10, 9 August 2015 (UTC)
Oppose
- Bureaucrats (Wikipedia usage) have been deemed by the community to be able to exercise reasonable judgment in assessing consensus. As with all humans, each may have a different view - but when they form a consensus considering their differing backgrounds, it is exceedingly unlikely that adding a new layer of "rules" will improve the result. Collect (talk) 12:17, 9 August 2015 (UTC)
- I will second what Collect has written here in opposition. My sentiment exactly! Fylbecatulous talk 21:17, 9 August 2015 (UTC)
- I'm having trouble seeing what good any of this would do. I know that there was that discussion at this one crat chat, but it probably should not be blown out of proportion. Better to leave all of this out. --Tryptofish (talk) 21:36, 9 August 2015 (UTC)
A Support without a rationale is an endorsement of the nomination, but an oppose requires a rationale
Votes need rationales because candidates need to know what they need to do to pass next time, and it is hard to have a discussion if people aren't saying why they have voted or what they have checked.
- Supports
- Though once there is a rationale for oppose other opposers can oppose per thatϢereSpielChequers 07:43, 7 August 2015 (UTC)
- Kraxler (talk) 12:12, 7 August 2015 (UTC)
- With the same caveat as WereSpielChequers. Kharkiv07 (T) 15:00, 7 August 2015 (UTC)
- Also per WereSpielChequers. At the same time, I'll say that supports without rationales, or with canned rationales, come across as lazy, and we really should encourage all RfA editors to put some research and thought into their comments, and demonstrate that they have done so. --Tryptofish (talk) 18:20, 7 August 2015 (UTC)
- Opabinia regalis (talk) 18:37, 7 August 2015 (UTC)
- Kudpung กุดผึ้ง (talk) 10:29, 10 August 2015 (UTC)
- Per WereSpielChequers. Callanecc (talk • contribs • logs) 02:34, 11 August 2015 (UTC)
- Since time immemorial, a "Support" or "Oppose" without a rationale tends to mean per nom or per all of or some of the above. Not just on RfA but everywhere. Gizza (t)(c) 08:02, 15 August 2015 (UTC)
- Oppose
- I think this opens up some problems with canvassing and puts an unfair burden on opposers. Any unexplained vote is a weak vote, but this is like assuming the person requesting admin is automatically an admin unless there is enough detailed explanation showing why they shouldn't. I don't think that is a level field to judge consensus on. Personally, I don't like pure "per Bob" votes either, and think that if you feel strong enough to vote, you should be expected to type one to three sentences explaining why. It seems ridiculous to consider such a small thing a "burden". Dennis Brown - 2¢ 20:23, 7 August 2015 (UTC)
- Sorry, Dennis but I have to partly disagree with you there. Votes such as this need to be indented, struck, and the user warned, blocked, or banned - in this user's 'case' for purely persistent disruption of the RfA process for years while contributing very little else to Wikipedia. Kudpung กุดผึ้ง (talk) 10:29, 10 August 2015 (UTC)
- I'm kind of agreeing with you Kudpung, saying that all votes should have a reason. Whether it is struck in the public eye, or only by the Crat, I'm open to either, but a blank support or oppose isn't helpful to a discussion, hence my note here. I agree with the second line of this proposal, but the first line is completely at odds. The first gives preferential treatment to supports, the second line says all votes are equal. A poorly framed proposal, actually. Dennis Brown - 2¢ 19:05, 10 August 2015 (UTC)
- Sorry, Dennis but I have to partly disagree with you there. Votes such as this need to be indented, struck, and the user warned, blocked, or banned - in this user's 'case' for purely persistent disruption of the RfA process for years while contributing very little else to Wikipedia. Kudpung กุดผึ้ง (talk) 10:29, 10 August 2015 (UTC)
- I do also prefer each !vote to have some kind of rationale attached; in "discretionary" cases, this can be immensely helpful at assessing the situation. Jo-Jo Eumerus (talk, contributions) 21:09, 7 August 2015 (UTC)
- Ideally, all !votes should show some thought as to why they are made. Especially since there have been cases where canvassing has resulted in a possible "false consensus" in support. Collect (talk) 12:26, 9 August 2015 (UTC)
- Unfair burden per Dennis Brown. You can either show a positive reason for support or have your opinion discounted. GregJackP Boomer! 18:53, 25 August 2015 (UTC)
- Discussion
One alternative is not to simply prohibit opposition without a rationale but to clarify that when the 'crats assess consensus, an opposition without a rationale will be viewed as weaker than a simple statement of opposition. We might also consider that opposition with a rationale and a diff is worth more than an opposition with just a rationale.--S Philbrick(Talk) 15:40, 8 August 2015 (UTC)
- I agree with that. --Tryptofish (talk) 20:50, 8 August 2015 (UTC)
- A contentless !vote still contributes to the tally and numbers matter. But I am inclined to say that it carries zero argumentative weight at all. Jo-Jo Eumerus (talk, contributions) 13:15, 9 August 2015 (UTC)
It should be acceptable for a candidate to respond to opposes
Support
- Kharkiv07 (T) 15:02, 7 August 2015 (UTC)
- And for others to respond as well, but doing so effectively requires skill. ϢereSpielChequers 18:16, 7 August 2015 (UTC)
- Broadly speaking yes, and it certainly should not be automatically unacceptable. But at the same time, it is very much a matter of how the candidate goes about it. --Tryptofish (talk) 18:26, 7 August 2015 (UTC)
- Of course it should, otherwise we will have more voters, including and especially rogue admins, telling blatant, vindictive lies, and going unchallenged by the voting community. If the candidate goes about it badly they would obviously not doing themselves a favour, and that's their problem. There should however be a cap on the length of such threads. --Kudpung กุดผึ้ง (talk) 05:11, 8 August 2015 (UTC)
- As is current practice. Extended colloquy, however, should be moved. I disagree with the use of a third party to question the validity of any oppose, the nominee should address any legitimate points raised - in practice it tends not to change anyone's !vote. Collect (talk) 12:42, 9 August 2015 (UTC)
- Yes, but I don't think it's feasible to tell people that they can't use that as evidence of behaviour (which might be that the candidate doesn't know when to stop talking and ignore). Callanecc (talk • contribs • logs) 02:55, 11 August 2015 (UTC)
- Often it is not particularly wise, but I already see it as acceptable. Best to do a single reply to a common concern, but the question posed here isn't one of style, it is of acceptability. Dennis Brown - 2¢ 13:44, 13 August 2015 (UTC)
Oppose
- Blatant harassment by supporters or the candidate is not appropriate. Of course, they never see it as harassment. GregJackP Boomer! 19:15, 25 August 2015 (UTC)