Wikipedia talk:Criteria for speedy deletion/Proposal/Test run

This is the talk page for the Test Run proposal amendment to the Criteria for Speedy Deletion.

Comments on Test Run

I need some clarification here. I was about to post the following in the Oppose section:

"This is the right idea wrongly conceived. A quorum of 3 is too few; many proposals will collect far more than 3 votes and the system could be gamed — indefinitely. There is also no specification of a further discussion or amendment but just a straightforward revote." — can anyone clarify what, exactly, would happen? -Splash July 5, 2005 00:22 (UTC)

My first thought was the same as yours, that even if a proposal passed by a supermajority of 200 to 3, those 3 oppose voters could cause a re-vote after a month. However, I'm not sure about "indefinitely". The wording is ambiguous, but one way that it could be read is as implying that there would only be the one re-vote. Uncle G 5 July 2005 02:05 (UTC)
- That's how I read it. However, I am more worried about there being discussion and a chance for change. It doesn't seem to specify. --Dmcdevit 5 July 2005 02:18 (UTC)
- I think that interpretation is probably correct. Radiant maks a good point about good faith in response to User:Dragons flight, but I prefer something along the lines of User:Cryptic's suggestion (under oppose). I'm reluctant to oppose this proposal, but I'm reluctant to let something too loose through. -Splash 5 July 2005 13:27 (UTC)

I think the suggestion to only get a revote if a number of 'yes' voters change their opinion, would be useful.
As an alternative, for a revote we could assume that everyone who voted one way on the original proposal would do so again, unless they explicitly change their opinion (and of course they should all be contacted).
Whichever way, a revote of any kind should be preceded by some discussion (for instance, it may be suggested to slightly reword one of the passed items after a month).
One revote should be enough, WP:NOT a bureaucracy.
Finally, I'm sure we're all mature enough to not make a WP:POINT and, say, ask for a revote of some hypothetical proposal that got 90% support, unless there is a very good reason (no sense in wasting everybody's time if it's merely to reassert something).
The way the Test Run proposal is worded, it doesn't specify how the test run would be, so I'm sure we can work out something that makes sense, do what it's supposed to do, and not be abused by a hypothetical troll and two sockpuppets.
R adiant _>|< July 6, 2005 15:03 (UTC)

How to judge?

Having a recall election is fine and good, but, as an average user, how do I tell how the criteria is being applied? I don't believe that the contents of deleted articles are visible to normal users (but could easily be wrong). So how can I judge if the criteria are being applied appropriately or not? Pburka 6 July 2005 12:29 (UTC)

Special:Log allows you at least to view which articles are deleted, and for which reason. R adiant _>|< July 6, 2005 15:03 (UTC)
This is something of a structural problem with VfU, IMHO. However, if VfU is suddenly overflowing with cries of "oh no you can't" if these proposals pass, it'll be clear enough that something has gone wrong. Admins should be encouraged to say in their deletion message, specifically which new criterion they were using to make this easier to determine.
- Side comment: I know VfU is about process not content, but sometimes, especially with CSD, the content is key to determining procedural validity. I think there should be a subpage in each VfU containing the article under consideration, but locked against all edits (including admins, if possible). That way, it'd be out of article space, out of view of Google etc. but within reach of those with a particular interest. -Splash 6 July 2005 15:40 (UTC)
Who will judge? only those who care? or non-partisan admin? Peregrine _AY 9 July 2005 09:33 (UTC)

Test Run Proposal B

In response to the objections and weak support votes for the test run, I suggest we introduce a modified test run proposal:

Test Run B.1: New criteria stays enforced for up to a minimum time of 1 week up to a maximum time of 10 weeks dependent on the leftover popular percentage over 70%.

To figure the popularity factor:
1. Subtract 70 from the final percentage in favor. ex. for an 80% in favor vote, take 80-70=10
2. Divide remainder by 30 and get percentage. ex. 10/30 = 33%.
3. Round percentage to the nearest week. ex. 33% of the maximum of 10 weeks is 3 weeks. The popularity factor for an 80% approval forces a 3 week test run before a recall can be initiated.

With 75% of the vote, there is a 16% additional popularity factor, new criteria has a 2 week testing time.
With 80% of the vote, there is a 33% additional popularity factor, new criteria has a three week testing time.
With 85% of the vote, there is a 50% additional popularity factor, new criteria has a five week testing time.
With 90% of the vote, there is a 66% additional popularity factor, new criteria has a seven week testing time.
With 95% of the vote, there is an 83% additional popularity factor, new criteria has an eight week testing time.

I think this is much too complicated. DES 17:47, 12 July 2005 (UTC)[reply]

Test Run B.2: A revote can only be called for when the test run is complete, and the number of requests match or exceed 10% of the number of vote originally tallied.
Example1: a criteria passed and 100 total people voted. 10 people would be needed to force a revote after the test run.

Example2: a criteria passed and 250 total people voted. 25 people would be needed to force a revote after the test run.

Test Run B.3: A maximum number of 2 revotes on the inclusion of specific criteria may be held in its lifetime, after which the criteria becomes canon, and should be considered as permanent as any other criteria for speedy deletion. Proposed modifactions to current criteria for speedy deletion are excepted of course, and are encouraged after the test run has completed.

What do you guys think? Can this be added to the list of proposals?

-Inigmatus 16:45, July 12, 2005 (UTC)

Sounds like a more lucid proposal. --Sn0wflake 17:14, 12 July 2005 (UTC)[reply]
This is better. Whether it is worth adding such a proposal at this date is another matter. As long as you are going to propose a change in the test run idea, make it explicit that there shall be at most one revote for any given proposal. Also make it explicit whether any sufferage restriction will apply to a) those who request the revote, and b) those who vote in the revote. I could support this with a single fixed length of test run for all proposals, but not otherwise. DES 17:47, 12 July 2005 (UTC)[reply]
Good point, I forgot to add that. It's in the draft here now as B.3Inigmatus 17:52, July 12, 2005 (UTC)
- So far as I am concerned one test run and one re-vote (for a ttoal of two votes) is all that is needed, and all that I can support. if this proposal calls for more than one re-vote, or if it calls for different lengths of test run based on the vote results, i will oppose it.DES 17:58, 12 July 2005 (UTC)[reply]
  - The different lengths are to ensure that new popular criteria is not revoted on sooner than not-as popular new criteria. Popular criteria should be rewarded with a longer stabilization period. Barely-passed criteria should be giving the minimum amount of time before a new vote can be held. It's logical.Inigmatus 18:20, July 12, 2005 (UTC)
The current test run proposal has been criticised for not saying whether there could or should be discussion of modification of the proposal in light of the experience gainsed during the test run, and whether a changed proposal could be approved by the revote. You should propably be explicit on this, one way or the other. DES 17:58, 12 July 2005 (UTC)[reply]
- Makes sense. I've reworded B.3 to make it clear that the revote limitations are explicit to inclusion of only that specific criteria, and not any new proposals that may modify it.Inigmatus 18:02, July 12, 2005 (UTC)

I do share DES's concern over the use of holding such a poll at this point in time, since any proposal which is accepted would have to be held back until voting on this one was closed. However, it is clear that voting on this proposal would benefit the proccess overall, as it would allow users to expose their point of view regarding the proposal itself, not some ridiculously low threshold. So I would gladly support the inclusion of it, regardless of the consequences it might have. --Sn0wflake 18:42, 12 July 2005 (UTC)[reply]

Should we wait for a fourth person's opinion on this before we put it in the vote?Inigmatus 19:29, July 12, 2005 (UTC)

This is not a vote, but I think it would an error to add it to the vote now. It would not make proper sense: the current polls close in about a week and this would need the usual two weeks. By the time this one passed, we'd already have a set of new policies (or not) along with a test-run (or not). To add yet another option so late in proceedings will only cause the kind of upset there was at the start. Also, as others have pointed out on the voting page, any user can force a review of any policy at any time so this proposal (and your replacement) are largely academic anyway. I suspect that's not the 4th opinion you were waiting for... -Splash 00:43, 13 July 2005 (UTC)[reply]

If 70% is needed to pass Test Run as it stands now, I don't see it doing so. I believe the other proposal modifications were put up because it looked like their originals were going down in defeat. But anways, if I get a fourth person here to support Test Run Proposal B, I'll see about getting it added to the vote.Inigmatus 14:21, July 13, 2005 (UTC)

No, the original proposals were modified to correct errors of substance in the wording. It is not possible to suggest that they were heading for defeat in their first 24 hours! Other proposals were added a couple of days in by others who disagreed entirely with the proposals as they stood; again not a question of imminent defeat. Note that those (IMO quite good) new proposals look like failing largely because they were introduced late; I expect that a new test run proposal would be similarly defeated. Do bear in mind that any user can challenge anything at any time, so replacing one test proposal with another is slightly academic. -Splash 14:47, 13 July 2005 (UTC)[reply]

For example I added 3-C because i opposed both 3-A & 3-B, but thought that something along the same general lines would be a good idea. Ideally such veraient proposals would have emerged before the voting started, but that didn't fully happen in this case. DES 14:51, 13 July 2005 (UTC)[reply]

Please do not count me as the third person supporting. I would support something like your B.2, but not in conjunction with your B.1 or your B.3 as currently posted. If this proposal were put up for a vote as it currently stands, all three parts for one vote, I would vote to oppose. DES 14:27, 13 July 2005 (UTC)[reply]

This should not be voted on. It is improper to attempt to "armor" your policy proposals against future change. If any proposals that are adopted aren't well accepted in practice, there will be a further debate, and if consensus is achieved, they will be changed when it is reached, regardless of all this Wikilawyering and Wikibureaucracy-building put in place to defend them. That is the way of the Wiki. Such proposals only serve as instruction creep, which is to be avoided. Building and maintaining real consensus is the only real option to maintain good policies. Un focused 14:55, 13 July 2005 (UTC)[reply]

Obviously setting the bar for holding a "revote" higher than the bar for starting a new proposal to amend or repeal a newly added policy has very little point. Whether setting the bar so low as to make a revote almost automatic (as the current test run proposal does) has much point might also be debated. I am also not clear, under the curent proposal would a revote again require 70% approval, or would it now be a vote to repeal, and so require 70% the other way to change? This whole buisness of test runs and revotes needs to be thought out better. A test run is IMO a good idea in principle. The details of how it should work should be better worked out. And IMO it would be best to have a standard format, not redebate the mechanics for every policy change in future. It is too late to do that for this proposal IMO, but it would be a good idea in the near future, I think. DES 15:36, 13 July 2005 (UTC)[reply]

Hear, hear. I think the test run proposal being voted on is one of the weakest proposals up for grabs at the moment. I only supported it because it is, in principle largely academic and does little harm. If 100-4 was the consensus to pass and those 4 users decide to go crazy, we can rely on the other 100 to tell them where to go: in the meantime the passed proposal continues to have effect (I presume). However, as I said, to add a new proposal into the mix now would be a mistake because of the hoo-ha there was at first; it would do much more harm than good (and there is the question of the suffrage of the nominator). Anyway, I'd certainly support a discussion of the principles of test-runs (or not) in the near future. Though we'd have to start by accepting that no policy can be 'armour-plated' as Unfocused points out.-Splash 15:53, 13 July 2005 (UTC)[reply]

Of course no proposal can ever be 'armour-plated' -- any policy can always be changed if a new consensus develops. To my mind a proper test run is intended to make a further change easier, not harder. In the immedated wake of the inital use of a new policy, people should discuss that policy, possibly modify it furhter, and come to a more settled consensus on what the relevant policy should be. A vote may be involed to demonstrate that consenssus, but, I would hope, only after a chance for discussion. Whether the policy uinder test would need to pass a second vote to remain policy, or whether the re-vote would be considered in effect a new proposal to change the recently adopted policy, I am not sure. DES 16:11, 13 July 2005 (UTC)[reply]

I guess to answer all these questions, I should mention that Proposal B is only meant as a policy for keeping new criteria yes or no. Yes to keep it, no to remove it entirely. Proposal B does not deal with any proposals for modification of that criteria. It only deals with whether the current criteria should be kept or removed entirely. Seriously, I don't think Test Run is needed (which is why I voted against it) if we just have a reasonable way to constantly modify by consensus current criteria. Test runs serve no purpose, I think, except to provide a chance for people to review new criteria and how its working. Proposal B just makes that process variable; it doesn't give a set time (1 month), nor does it make it easy for just any criteria to be called into question (the 3 or 4 user dissatisfaction). If Test Run is to be considered as a needed element in CSD, then I propose to make these limits variable according to the situation so as to minimize abuse (the 3 persons calling any criteria into question) and maximize efficiency (a one week test run limit for not-so-popular passed criteria vs a standard 1 month which really is unfair). And no, I don't agree on padding passed criteria with a safety net so they aren't called into question. That is why I propose that if Proposal B goes to the voting block, that each part be seperately voted on. To be honest, I believe if any criteria that is passed by a 70 to 30 vote should had a mandatory review not in a month, but at most a week; whereas criteria that passes with a 95 to 5 vote is privilaged to not face a review for at least 8 weeks. Seemed more fair to me. And B also ups the recall ante to 10% of the number of voters who had participated in a criteria's passage instead of letting 3 dissatisfied people call the entire community back to the issue. Anyways, that's my two cents. As with all things wiki, I leave this to the development of a consensus. Inigmatus 23:09, July 13, 2005 (UTC)

Well, my points, in summary are these: Voting is by far the least preferable method of finding consensus as it is evil (See WP:NOT, and m:Polls are evil). Polling is second most evil (See m:don't vote on everything). However, any new proposals that do find consensus, regardless of how it is found, should be adopted until such time as the bitching about them spurs further change. There's no need to formalize everything. Don't clog the system with rules and thresholds and further selection trees and acceptance levels that really do very little other than encourage more voting rather than civil discussion, give and take, negotiation, and assent. Rather than work on building ever more complex voting schemes, consider the result if the same efforts are put into building better diplomacy skills. We're here for a common goal, but voting (by it's very nature of "select from these choices") divides the forces. Each of the original proposal group could have been proposed as a consensus as a result of discussion, and put into place immediately absent major objections. This would have been far preferable, and how I suggest anything following this vote be handled, including the length of time before these are reviewed again. Voting is a last resort, not a first step. Un focused 00:51, 14 July 2005 (UTC)[reply]