Wikipedia talk:Requests for adminship/RfB bar

From Wikipedia, the free encyclopedia

Contents

[edit] Question to crats

In order to change the 90% threshold, do we need to have a consensus on a new % or just consensus that it ought to be lower? If lots of people advocate for 75% and 80-85%, does this create a new discretionary range or demonstrate lack of agreement on what the minimum should be? Avruch T 18:33, 7 March 2008 (UTC)

I've been pondering that question. There are many variables here. Someone who supports 85% doesn't necessarily support 90% or 80%. Maybe people should list their names under any number they support. Some people said 80% - 85%. Some said 80% only. That might help determine what the will of the editors is. Or it might create a mess. Kingturtle (talk) 18:38, 7 March 2008 (UTC)
Yeah it sounds like a recipe for a mess. I think we can first start with the premise that there is general support for lowering the minimum, and then take the most popular minimums and incorporate them into a discretionary range - i.e. if most people say 85%, and second most say 80%, make 80-85% discretionary. Does that make sense? Avruch T 19:11, 7 March 2008 (UTC)
This could be made a separate question: If you support lowering the limit to N%, would you oppose lowering it from 90% to M%, where 90 > M > N ? I can only think of one good reason someone might so oppose - that lowering the limit a little would reduce the pressure to lower it to their preferred N%. People will, of course, come up with bad reasons, too.
I'd think that unless the outcome of the question I posed above was not overwhelmingly yes, that we can assume that everyone supporting 75% also supports 80% and 85%, and determine consensus based on those totals. Right now, based on that reasoning, there is 79.2% support for 85%, 66.2% for 83.333%, 64.9% for 80%, and a minority for lower than 80%. (Vote totals based on 77 votes.) To me, that looks like a consensus for 85%, and probably not a consensus for anything lower. (However, vote 9 in 90% should probably be moved to 85%, which would make it >80% support for lowering to 85%.) Argyriou (talk) 21:39, 7 March 2008 (UTC)
Just a note but my opinion (vote 9) should stay where it is - I want 90% as the guideline level but as has been the case before there has been discretion in the 85-90 range with promotions below 90 as I do not think it should just be a vote. I do not support an almost automatic promotion at 85.1% , 86% etc. Davewild (talk) 21:49, 7 March 2008 (UTC)
"we can assume that everyone supporting 75% also supports 80% and 85%"...that's what I was thinking originally, but some people supporting 75% will say 85% and even 80% is too high. So we cannot make that assumption. Kingturtle (talk) 21:51, 7 March 2008 (UTC)
I think we could, as a reasonable simplifying assumption. People opposing 85% because they want 75% are not really consensus building, they're kind of gaming the system. Obviously alot of people are going to have to compromise; otherwise, it's a textbook case of the perfect being the enemy of the good. --barneca (talk) 21:55, 7 March 2008 (UTC)
Each person is going to have their high bar and their low bar. X is the lowest I will go; Y is the highest I will go. For some X and Y will be the same: "80% is as high and as low as I will go. 85% is too high; 75% is too low." We cannot assume every 75% supporter will also support 80%. Kingturtle (talk) 22:04, 7 March 2008 (UTC)
That would be precisely correct if we were trying to come up with a breakpoint from scratch. However, I think it breaks down in this case because we already have a breakpoint in the high 80%'s to 90%. If consensus to change isn't reached, it will remain at a number that most people think is too high. Therefore, for the sake of having any hope of building consensus, I think people's default Y has to be considered 90%. I doesn't make sense to have consensus that it should be lowered, but not lower it because we can't get consensus on what exact ratio we consider a reasonable breakpoint. I think the only reasonable way to interpret this is: lower the breakpoint until you no longer have a consensus to lower it further. --barneca (talk) 22:11, 7 March 2008 (UTC)

The problem is we're adapting answers to one question to a second question, when if you want a new answer you really need to ask a second question. If the proposal was "Do you support 85%" then you could say that there is x% support for changing it to 85%, but the proposal is "A percentage below 90%" which has 60:16 support (~79%). Its handy that most people specified a preferred number, which I'd asked for, but the spread shouldn't be construed as showing conclusive support for a particular percentage. Whether this means that it can't be changed because the community couldn't decide on a particular number, I don't know. I don't think we can assume that supports for 75% are an oppose for 80%, though - more reasonable to assume the opposite, because the proposal specifically asks for less than 90% and most of these comments are prefaced with Support. Avruch T 21:58, 7 March 2008 (UTC)

Just brainstorming here...another idea I was bouncing around is to average the responses...

  • Minimum of 90% - 18 editors
  • Minimum of 85% - 14 editors
  • Minimum of 83.33% - 1 editor
  • Minimum of 80% - 38 editors
  • Minimum of 75% - 17 editors

...would average to 81.9%. However, I imagine such an idea would have to be told to editors prior to the poll, not during or after the fact. But it is an interesting idea to play with. Kingturtle (talk) 19:20, 8 March 2008 (UTC)

Well, I'm not sure an average figure is quite what we should aim for - ideally I think there should be a particular guideline proposed that can then be endorsed by the community in subsequent discussion. I also don't think chaining ourselves to a number is the best approach - the figure would presumably be a guideline but there seems to be support for bureaucrats weighing arguments as usual. I would suggest something like:
The community expects that requests for bureaucratship supported by more than 85% of participants would generally be successful and that those with less than this would not be. The closing bureaucrat(s) nonethless have a traditional discretion to assess the discussion and make appropriate decisions in cases that fall either side of this marker based on the strength of argument.
Thoughts? WjBscribe 19:26, 8 March 2008 (UTC)
Or how about:
The community expects that a request for bureaucratship will succeed when there is no significant opposition. The closing bureaucrat(s) will have discretion to assess the discussion and make appropriate decisions based on the strength of argument. The closing bureaucrat(s) will also assess the request and decide whether it meets the no significant opposition threshold. Bureaucrats are expected to confer when closing requests where the significance of the opposition is not readily apparent.
RxS (talk) 19:36, 8 March 2008 (UTC)
That's the sort of thing, but I'm getting the feeling that overall there isn't much support for the "no significant opposition" standard and that something lower is wanted? WjBscribe 19:38, 8 March 2008 (UTC)
Maybe but there's so much it's not a vote talk going around maybe it's time to put our money where our mouth is? I'd be happy with you guys making these calls...otherwise it'll always be a vote, and the argument will just shift to wherever the threshold is at the moment. Surely no one is saying a bureaucrat should be promoted when there is significant opposition are they? RxS (talk) 19:45, 8 March 2008 (UTC)
We're trying to find a consensus as to what constitutes significant opposition. Kingturtle (talk) 19:47, 8 March 2008 (UTC)
Wouldn't my wording would put the trust in you to evaluate the opposition to find it's significance. Otherwise like I said, what constitutes significant opposition will be always be a numerical threshold. You guys are in a position of trust, each RFB has it's own uniqueness that goes beyond numbers...why couldn't we trust you to make those calls? RxS (talk) 19:52, 8 March 2008 (UTC)
That is an idea being supported in this poll under the heading Trust the bureaucrats. Kingturtle (talk) 19:57, 8 March 2008 (UTC)
Yes it is and added my support yesterday. I think some specific wording is a useful thing to think about. RxS (talk) 20:02, 8 March 2008 (UTC)
I think it's unfair by not counting those who voiced at WT:RFA but didn't post in here. OhanaUnitedTalk page 22:25, 8 March 2008 (UTC)

[edit] Reconfirmation of bureaucrats

I think that the arguments for reconfirmation of bureaucrats are probably stronger than for admins. The arguments raised at Wikipedia:Adminship renewal and WP:PEREN, that reconfirming admins would create an unacceptably high workload, would be less important in the case of crats. If we reconfirm crats after a probationary period, it might be acceptable to have a lower bar to get in. Any thoughts? Obuibo Mbstpo (talk) 22:24, 7 March 2008 (UTC)

We were talking about this somewhere else just now... I think on the discussion page for the 'crats discussion of the Riana RfB? I guess thats sort of an out of the way place to expect anyone not already involved to comment ;-) Avruch T 23:20, 7 March 2008 (UTC)
I think this is an idea whose time has come, especially with the potential of a significantly larger group promoted. Many of my misgivings about lowering standards would be eased with a realistic recall/reconfirm process in place. RxS (talk) 05:59, 8 March 2008 (UTC)

[edit] Pay no attention to the numbers

You've always had the power to have a sensible method for evaluating RfBs. Just close your eyes and tap your heels together three times. And think to yourself, "RfB's not a vote, RFB's not a vote." --bainer (talk) 05:35, 8 March 2008 (UTC)

I confess to being irked at the way this was set up. My previous comments have been refactored into a hierarchy with which I disagree and are identified as supporting a threshold for which I never indicated any support. I made comments about the impropriety of having a discussion to shift RFB standards in the middle of active RFBs. Frankly, I feel that my comments were quite on the mark, as really only a tiny portion of the discussion has been about why we should increase the number of bureaucrats, and the majority of the discussion (that I've seen) has been about the specific case of Riana. The two issues are, at this point of this discussion, inextricable. I supported Riana and have been vocally critical of many of the editors who opposed her, and I'm still finding it extremely difficult to separate my feelings about this particular case from feelings of what really makes sense for Requests for Bureaucratship in general. I do not agree with the way my comments have been refactored, as I was trying to have a discussion (silly to attempt, I can see that now) and was never voting for any particular numerical threshold. --JayHenry (talk) 09:42, 9 March 2008 (UTC)
Jay, in that case, please move your comments to where you believe they belong. Kingturtle is just as human as we are . -- Avi (talk) 15:32, 9 March 2008 (UTC)
Jay, I apologize that by creating this page I adversely affected your comments. In the old format it had become extremely difficult to discern the support for the various positions. I did my best to carry over each editor's position into the new format. Please adjust my mistakes accordingly. Kingturtle (talk) 17:11, 9 March 2008 (UTC)

[edit] RFB needing to be harder than RfA

Hi, Slim. Firstly, at around 70% the grey area for admins begins, but even assuming a hard rule of 75%, 90% for RfA's is more than three times as hard. It is three times greater mathematically (9:1 vs. 3:1) and the inherent higher standards (just see my own recently failed RfB) of RfB means that the same percentage is implicitly harder to achieve at RfB than RfA, although I am uncertain as to how one can quantify that. -- Avi (talk) 06:32, 9 March 2008 (UTC)

The point of the high bar is to make sure that it's very tough to pass. We don't need many bureaucrats, and we want them to have the complete confidence of large sections of the community, which means we need very few opposes. SlimVirgin (talk)(contribs) 06:34, 9 March 2008 (UTC)
What it seems to boil down to is whether or not the complete confidence of 4 out of every 5 responding editors is as telling as 9 out of every 10 responding editors, with the inherently higher standards factored in—notwithstanding that the origin of the 90% mark is rather murky. Well, the community itself is in disagreement about that; thanks for taking the time to make yourself heard! -- Avi (talk) 06:40, 9 March 2008 (UTC)
I think even looking at it in terms of percentages is wrong-headed. I think anything over 20 opposes starts to be a problem if you're looking for someone with full community confidence. Most people don't vote to oppose, even if they do oppose. They rely on others to do it, and they hope it's enough because they want to avoid bad blood. So if you're looking at 20 actual opposes, you have to assume there are quite a few unspoken ones too. SlimVirgin (talk)(contribs) 06:56, 9 March 2008 (UTC)
But with the growth of the community, 20 opposes represents less than it does even a few years ago, does it not? -- Avi (talk) 07:11, 9 March 2008 (UTC)
Yes, it does. But if you look at the last few successful RfBs, the opposes have been very low. The Rambling Man, 3. WJBscribe, 3. Andrevan, 12. Deskana, 12. Cecropia's reconfirmation, 5. Redux, 1. Taxman, 3.
These are the kinds of figures we're looking for if we want full community confidence.
It means that only a particular kind of contributor can stand. No one who has been a very active admin is likely to get it, because they'll have made enemies. Ditto if you've edited contentious articles. But given that bureaucrats are supposed to be above the fray, it does mean the number of opposes have to be minimal, almost regardless of what percentage they are of the supports. SlimVirgin (talk)(contribs) 07:30, 9 March 2008 (UTC)
If you want to look for someone that is supposed to be above the fray, it should be steward and not bureaucrat. There are some stewards who got elected in 2007 and received more opposes than the examples you gave (like DerHexer - 35 opposes, Nick1915 - 25 opposes) They became steward because the amount of votes reached the plateau on the graph created by someone on WT:RFA. OhanaUnitedTalk page 07:53, 9 March 2008 (UTC)
The Steward analogy is profoundly flawed. Stewards are on a different and self-selecting project. Their purview is strictly limited to projects which are not their home project, one steward can reverse another steward, they must be identified to the foundation, and they must stand for reconfirmation every year. There's no value, in my opinion, of comparing these two positions. Like comparing apples and uranium, as they say. --JayHenry (talk) 09:20, 9 March 2008 (UTC)
  • FWIW, I fully agree with Kylu's and EdJohnston's comments: People, on average, are much more careful before supporting an RfB anyway, so what's the merit in making it doubly hard by additionally having a ridiculous and arbitrary 90 percent bar? Dorftrottel (taunt) 10:12, March 9, 2008

[edit] A quick summary

As of now, there are 30 editors who believe 90% is fine, and there are 84 who believe it is too high. I would suggest this is a fairly sound endorsement for the bar being lowered, and I would be surprised if this were to change significantly.

As to what it should be lowered to, go with the majority choice (80%). Nothing is final and if this causes problems we can always readdress the "pass percentage" a few months down the line. Neıl ☎ 11:55, 11 March 2008 (UTC)

It is very important to discuss how we're going to interpret the results. The poll ends in four days. Kingturtle (talk) 12:08, 11 March 2008 (UTC)
It appears that even of those giving numbers, such as myself, the concept of b-crat discretion is still paramount. I beleive there is a strong case for removing the de jureness, as it were, from the de facto 90%. As to a specific number, I think that is less clear, and less important. Becoming a 'crat should remain significantly more difficult than becoming an admin, but much of that is already taken into account by the respondents' own comments. So if we see one potential crat promoted at 85% and another not at 87%, if the decision of the 'crats is justified and explained based on support/oppose comment quality as well as quantity, (be it on a subpage or in the heading) I think that most of us will be happy. I also could be dead wrong about everyone else but myself :) -- Avi (talk) 12:30, 11 March 2008 (UTC)
  • According to the current (20:52, 13 March 2008 (UTC)) numbers:
  • Minimum of 90% - 35 editors
  • Minimum of 85% - 20 editors
  • Minimum of 83.33% - 1 editor
  • Minimum of 80% - 63 editors
  • Minimum of 75% - 16 editors
(omitting Dweller's non-specific "No fixed figure (but <90%)") the current average value is roughly 82,8%, which happens to be in the middle of the bureaucrat discretion range of 80%-85% suggested by several users. Dorftrottel (talk) 20:52, March 13, 2008

I believe that might be more an artifact that the midpoint of the high and low ends is 82.5 than anything else (low = 75, high = 90, median = 82.5). -- Avi (talk) 00:21, 14 March 2008 (UTC)

You're probably right. Didn't occur to me. Dorftrottel (warn) 01:02, March 14, 2008

[edit] Next step

This poll is closed. We have over 170 editors voice their opinions on this issue.

Depending on how you interpret the data, 75.5% support lower the bar; 65% of the supports support "80%" as the new bar. The average of the opinions is 82.9%. "80%" is the median.

I am interpreting this poll as an indication that Bureaucrats should use 80% as the new bar.

Once a new policy is agreed on, I think we should re-assess the policy in one year. Kingturtle (talk) 19:02, 16 March 2008 (UTC)

This seems to be a reasonable interpretation. Keilana|Parlez ici 19:08, 16 March 2008 (UTC)
Unless you looking at this as a straight vote I don't see a consensus. If you consider the discussion, many of the objections haven't been answered, and a large amount of the supports fall along the lines of "too high" or "80 is about right" or "is decent" without really saying why. I don't expect this to mean anything, but voting isn't how we are supposed to judge consensus and for a change as potentially disruptive we should be looking for a clearer consensus. RxS (talk) 19:15, 16 March 2008 (UTC)

I would have preferred a bit more discussion and a little less polling. Reading the discussion as a whole, I don't think its a simple case of suggesting 80% as the new bar - there seems to be a very significant input from people who want to see more discretion based on the strength of opposition argument. Ideally I would look to trying to factor in as much of the opinions expressed into our decision making - not just the one that received most support, or the average figure - even within each section, people's response are quite individual an nuanced. I would propose a fairly loose formulation as a response to this:

"Whilst RfB is not a vote, it is generally expected that RfBs with more than 90% will be successful, whereas those with less than 80% will not be. Bureaucrats should assess the level of consensus bearing in mind the high levels of community trust expected for appointment."

I wonder if that is a good way to have an approach that factors in as much of the positions expressed as possible? Effectively I'm using what I interpret as the consensus on RfA closing, but acknowledging that the community expectations seem to be about 10% higher. WjBscribe 19:17, 16 March 2008 (UTC)

In the course of this poll and discussion, and the "crat chat" in the middle, its been made clear that some bureaucrats are committed to the idea of 90% as a threshold. The italicized statement above seems to represent a license for this opinion to continue to be enforced, which would be a pretty disappointing result in my view. In response to RxS' comments above, few of the supports were addressed in the opposes, and the support side of the question certainly presented a far more robust argument (with mathematical analysis and graphing, no less...). Many of the opposes in favor of the 90% minimum basically said "We don't need more 'crats!" which is, in my view, beside the point. Avruch T 19:33, 16 March 2008 (UTC)
That's a mischaracterization of the opposes here. If you boil away the raw vote and focus on the debate there's no consensus for this change. And even if you do treat it like a vote (which we don't do) the number is barely above the discretionary range used in an RFA. I think we need more unity for a change like this. RxS (talk) 19:46, 16 March 2008 (UTC)
13 used that rationale in voting for 90%. At least a few others went for 90% because RfBs were in progress. As for barely being enough to pass an RfA... Well, as many have said, different discussion - different measurement of consensus? What is the threshold for a policy change? Avruch T 20:58, 16 March 2008 (UTC)
I'd be ok with WJBscribe wording, I'd prefer to change 80% to 85% in a perfect world though. RxS (talk) 19:49, 16 March 2008 (UTC)
That wording sounds fine with me. Kingturtle (talk) 20:08, 16 March 2008 (UTC)
I Concur with WJBScribe's wording as well. I think very few of the respondents believe that 80% should be considered a sure pass; rather, that it should no longer be the sure fail that it was. -- Avi (talk) 20:52, 16 March 2008 (UTC)
I think the poll is good evidence that 80-85% or 80-90% should be the new discretionary range, not that circa 80% is a sure pass into 'crathood. But when in doubt, don't mix up a working system! Deacon of Pndapetzim (Talk) 21:16, 16 March 2008 (UTC)
One thing I noticed is that the poll should have had a section that said "80-85%". There were quite a few "votes" in both sections which said they preferred 85, but 80 was ok for Bureaucratic discretion. ("I demand a recount" - rofl) - Joking aside, it would seem to me that the "safest" (read as lesser disruptive) would be to "set" 85 instead of 90 as the "rough new precedent", and suggest that discretion even to 80% shouldn't be considered "disruptive", or requiring a bureaucrat chat, or whatever else passes for drama in what some "vote counters" may consider a controversial closure. - jc37 22:23, 16 March 2008 (UTC)
I think there is a consensus that what we have now is not what we want and I agree that WJBScribe's wording and/or Jc37's characterization is much better than what we have now (though I continue to despise the entire idea that consensus can ever be reduced to a percentage - even if we come to a consensus that it is so).--Doug.(talk • contribs) 23:10, 19 March 2008 (UTC)
That seems the best description so far, actually. There does not seem to be agreement as to what the "discretionary" threshold should be, but there does seem to be consensus that the current situation is not the one we want either. And, as said before, there is a very real chance of votestacking by aggreived parties in cases like this, which can and should be at least potentially allowed for, particularly if they know which set percentages need to be met. WJBScribe and Jc37 seem to have made fair assessments of the discussion in my eyes anyway. John Carter (talk) 23:19, 19 March 2008 (UTC)
There's no strong consensus for any single one of the options, but that was bound to happen. However, by my reckoning, there's not even a firm consensus, using the standard applied in the past to other polls of this type (80%), for lowering the "bar" at all. I'm counting the "no change" lobby as 40 people, and the "change, please" lobby as 110 people (22+1+71+15+1), which means 73% of those participating favor some amount of change. I'm not sure where to put the last two groups, so I've left them aside.
However, this fact doesn't worry me terribly much. It's clear enough that a lot of folks want to see us judge RFBs more leniently, so I will begin to do so, and I hope the other bureaucrats will do the same. 85% sounds like a fair place to start. This should not be taken as a permanent policy, nor as a guarantee that the number will grow lower in time: I make no long-term commitments. I would rather ask the community again in a while. (For especially diligent readers, I should mention that this comment is largely a copy of the comment I just made here.) — Dan | talk 03:04, 20 March 2008 (UTC)
85% is fine (I think I voted for that...I forget...) and I'm glad to see a 'crat take the community's wish on board. dihydrogen monoxide (H2O) 04:00, 22 March 2008 (UTC)
"However, by my reckoning, there's not even a firm consensus, using the standard applied in the past to other polls of this type (80%), for lowering the "bar" at all." — Please define "firm consensus" with respect to the poll. What would have been firm consensus? Does that mean nothing will ever change because a supermajority of only 73% favours change? Dorftrottel (canvass) 06:47, March 22, 2008

I would like to add "Whilst RfB is not a vote, it is generally expected that RfBs with more than 90% will be successful, whereas those with less than 80% will not be. Bureaucrats should assess the level of consensus bearing in mind the high levels of community trust expected for appointment" to Wikipedia:Bureaucrats#Promotions. Are there reasons why not to? Kingturtle (talk) 13:11, 8 April 2008 (UTC)

I'm uncomfortable with 80% for the lower figure. Dan points out that leaving aside the last two groups, 73% of participants favored thresholds of less than 90%—a supermajority, but low enough that he thinks it is not a firm consensus to lower the bar at all, using past standards. By similar reckoning, only 57% favored thresholds of 80% or less. I would proceed cautiously and await a stronger mandate before dropping it that far. Tim Smith (talk) 18:12, 8 April 2008 (UTC)
I think that is acceptable - my own preference would be for 75%, but 80% seems to net the larger share of the support from the poll. I'll also point out that its almost irrelevant - current bureaucrats will promote based on their own opinion, as they have expressed it, and the supermajority in support of a particular metric (whether reflected in the policy or not) is unlikely to change that. Those 'crats that are likely to be influenced by the outcome have been already. Avruch T 18:40, 8 April 2008 (UTC)

[edit] post poll commentary

  • 80%...I've made my position known on this. For those not familiar, it's basically the same as the above people. Nousernamesleftcopper, not wood 23:33, 19 March 2008 (UTC)