Wikipedia:WikiProject Editor Retention/current discussions/innocent prisoner's dilemma discussion

= Innocent prisoner's dilemma =

Block policy essentially requires that an editor admit his guilt, then explain what he will do in the future to make sure it doesn't happen. What about the times when the blocked person is innocent? This can happen when an admin is acting improperly, or is acting properly and just makes a mistake. For this discussion, I would prefer we assume it was a mistake and simply focus on what the solution should be. When it isn't an innocent mistake, then we have two problems, not one, so limiting the discussion to the one will keep it on topic. First reading the article "Innocent prisoner's dilemma" is required to get a full understanding of the problem here. The ideas proposed here are just rough ideas, and I wouldn't suggest running off an proposing them anywhere too soon. The purpose of the discussion is to try to find solutions and work out the bugs before proposing changes, to give a higher chance of reaching a consensus. Keeping arguments on topic and pithy is also helpful

A few facts
Currently, when a person is blocked for policy violation, they only have access to their talk page and to ask for an unblock. WP:UNBLOCK states: ...the blocking administrator should not decline unblock requests from users they have blocked. as well as ''Except in cases of unambiguous error, administrators should avoid unblocking users without first attempting to contact the blocking administrator to discuss the matter. If the blocking administrator is not available, or if the administrators cannot come to an agreement, then a discussion at the administrators' noticeboard is recommended.''

Third party assistance for unblocking isn't covered specifically in policy, and acceptance of third party requests are inconsistent.

While it doesn't specifically state this (and perhaps it should), if an editor makes multiple requests for unblock, a different admin should review the unblock each time. The only exception would be if one of the previous declining admins decides to accept and unblock.

It should be obvious that no "justice system" is perfect, and that means that sometimes, innocent people ARE getting blocked, just as innocent people are convicted of crimes in the judicial system. What is needed is a system to allow review, but won't be used primarily as a source for abuse.

Since sometimes good admins make mistakes in blocking, we have to create an environment that doesn't attack the blocking admin when they are wrong. Accountability is important and should be part of the system, but we have to allow that mistakes will be made in good faith. In order to get admins to freely admit a mistake, you have to not bludgeon them for that occasional mistake, and encourage self-review.

Dennis Brown - 2&cent;    &copy;  18:54, 1 July 2012 (UTC)

Discussion

 * Depends on whose version of the truth you want to believe. What's civil? What's incivil? Depends on your background, beliefs, and culture. Pumpkin Sky  talk  19:19, 1 July 2012 (UTC)
 * This is the first time the word "civil" is used on this page, so at a loss as to what you mean. I'm trying to start conversations in singular, digestible chunks.  Of course, I can't control where the discussion goes, but I try to keep it focused on ideas for solutions, and put different problems in their own full thread, all on the same page.  Dennis Brown -  2&cent;    &copy;  19:57, 1 July 2012 (UTC)
 * I like the idea in theory, but in practice how do you handle those admins who are not amiable to self-review? Admins are human, too, which means there will always be those who do not feel that they make mistakes or bad calls. Intothatdarkness (talk) 16:05, 2 July 2012 (UTC)
 * There is a key to it over on the block pol discussion page, that slab someone put up mentioning the last 500 blocks, now that editor we need, That editor can write, and write well. Sure, people didn't agree with them, makes no difference, they could be lured and forced to do our bidding :) No, but seriously, that editor has the gift of the gab.
 * Now the point was the last 500 blocks, and the analysis thereof, separating the manner in which unblocking is handled by breaking it down into groups is very important. Throw away socks, (oh btw there is a sock investigation request for me right now, :) quite an earnest well-meaning editor opened it, unfortunately the docs are such a complete mess, he/she has found how to request a check with great competence, without the legitimate uses of socks policies being placed in his/her path.) Throw away socks, their blocks, and unblock requests, and the requests of long-time editors in good standing need to be handled differently. There is little investment in explaining a block to a throw away account, because there is little investment in making them, so admins are constrained by math to be brief when encountering those accounts. Investigating the block as part of a review would need the blocking admins input because of the time cost in writing down the reasoning, it couldn't be done n the first place. So the admin has to be involved. For longtime editors who have been blocked, the resource constraints are inverse. The faster the admin makes the block, and the less information that is given at the time, results in more time consumed later by editors over the block. The quicker they are with old editors, the worse the block, and the more time is consumed down the road. One size doesn't fit all, not in the block, and not in the unblock pol.
 * The greater the investment an editor makes in the community the more they expect in return, argue all you want with me there, you'll just be wrong, it's elemental human nature. Editors of long standing expect to be treated fairly and expect to have their say, and expect 'due process' a proper fair show. If there is no show to be seen, and a long time editor is blocked and reviewed the same way as a 5 minute sock you lose ten editors at least for every one editor you block (counting editors who stay, but become a ghost in the shell, disillusioned and no longer motivated, counting them as lost, so you can count 20 half hearted editors as 10 lost). The long time editors have community ties, and cannot leave unnoticed. 100 bad blocks for editors with 10 edits, and nobody is going to care less when they examine them, well, they will care, proportional to the likelyhood that it was a gf newbie, but not nearly as much trouble is caused when they look at the 10 edits and the admin says sock or something like that,
 * For a long time editor, all doubt must be erased from the minds of onlookers, and the block has to be properly explained and open to review, otherwise, it's just a matter of onlookers thinking 'when is it my turn to be (mis)treated that way'
 * For the shortie newbie block, admin required. For the longtime editor block, there is always going to be the question of conflict, there is always going to need to be little or no doubt left, and face the facts, there are always going to be a lot of editors involved with it. So many that the original admin is more of a hindrance to the process than an asset. There is going to be so much input either for or against that the participation of the original admin will be problematic, and the amount of influence the original admin has over the outcome of the review, the more damage is done, and the further apart from the community the admins are pushed. Independent reviews have the opposite effect, the community has more trust in the decision, the more reviewers involved, the greater the trust in the decision. Greater the trust in the original admin also, it is easier to have faith in someone whose mistakes can be corrected, someone who is happy to rely on the community or colleagues to 'fine tune' the block, then a dictator, where the only thing the community is asked to do is trust their dictators whose decisions can't be questioned. So unblock reviews need to be handled differently, how to do it is down the line, but that is the initial analysis. Oh I read that report on people who had left, I haven't ever seen such a report before, that's brilliant. Everything in it came as no surprise except for one thing, and that was the influence of the editor interface, I never really thought about the interface that much, it's one of the smallest things, but it is significant. The remainder of the report was all very simple. Penyulap  ☏  06:08, 3 Jul 2012 (UTC)
 * What report are you referring to? ```Buster Seven   Talk  18:04, 3 July 2012 (UTC)
 * Penyulap is currently blocked, so can't reply. I would expect it's among some of the links below (in the "Further Research" section. Intothatdarkness (talk) 18:28, 3 July 2012 (UTC)


 * Right now we have a situation (I'm not being specific as this is part of a larger, general problem, but I'm sure everyone knows what I'm talking about) where a user has been offered reprieve from a two-week block in exchange for "a credible inclination that his problematic behaviior has stopped". Obviously this isn't going to happen, as the user honestly feels his behavior was not problematic (other than a minor issue to which he has admitted and said he would have redacted if given the chance). So, unless an admin with balls comes along, and says, "whatever the case, this block has served its purpose", this user will remain needlessly blocked, unable to improve the encyclopedia, for the full of two weeks. We have admins with balls; lots of them. It takes quite an incredible amount of testicular tissue to do some of the things I've seen admins do as of late. If only this chutzpah could be used to bring us together. I've seen many an admin display a userbox stating, "this administrator can and will make difficult blocks, if necessary". How about difficult unblocks? Joefromrandb (talk) 06:20, 4 July 2012 (UTC)
 * I've involved myself in a few and reverted a few actions (and i'm barely 2 months into having the mop), but I'm not sure of the best way to go about it all the time. The problem is, if I hung up a shingle that said "I will make difficult unblocks", then I would be busy around the clock researching cases where 99.9% of the time, it was a good block, wasting my time.  I already say on my user page that I will consider 3rd party reviews of admins and get a fair amount of traffic there, but I've not figured out how to streamline the process and be fair to everyone (including the admins), except on a case by case basis.  Dennis Brown -  2&cent;    &copy;  21:50, 8 July 2012 (UTC)
 * The innocent prisoner's dilemma is something I've been coming across regularly in Real Life for over a decade now, because of one of the other (many!) hats I wear. It is a real problem, and I think maybe that one of the biggest problem areas may be editors getting blocked as socks on the duck test.  This is something I remember being talked about a few months ago.  Of course people who live in the same area, maybe went to the same school / college (learning the "house style" for that facility while they were there) and have similar views are going to look very much alike, and it's not at all impossible that they are, in fact, two entirely separate individuals who geolocate to the same area.  But if an entirely innocent duck-tested newbie tries to do anything about it, they're met with a "Stop quacking, we're done here; another sock complaining, do they think we're idiots? ... move on, folks, nothing to see here" response.  Now that's a problem, and the only way I can see around it would be for them to email ArbCom with some personal ID or something, which they're not likely to want to do.  At present, though, there seems to be no way to appeal a duck-test result.  Pesky  (talk ) 05:34, 31 July 2012 (UTC)

Possible solution: reduce block lengths
In most cases, the length of a typical block is just too long, and admins should be re-educated to make shorter blocks. A block should typically be a matter of days, not weeks or months. For instance, a suspected first-time wp:Sockpuppet is often blocked for 1 month, rather than days. Newsflash: No one is so stupid that they need a month to figure out wp:SOCK. Most likely, overly-long blocks are used as a substitute for taking time to talk with people who fail to understand the editing norms, such as the general acceptance of the typical WP:MOS style, or the complex limits of using copyrighted texts. For example, people who write technical documents might recommend a different font style for some articles, and they might be blocked for insinuations that someone is not as experienced as them in technical writing, and therefore get blocked for "2 weeks" for thinking that their knowledge or experience is somehow valuable in Wikipedia as it would be to the real world. Anyway, make blocks shorter, and the blocks would expire before unblocking is needed. -Wikid77 (talk) 12:05, 10 August 2012 (UTC)

Possible solution: have unblock advocates
If there were a group of "unblock advocates" then they could help to analyze the actions of a blocked user and get the unblock faster. I have acted as such an advocate, and it has worked, in several cases, to perform the "talk" needed to get agreement to avoid conflicts and respect established guidelines. However, some people remain difficult, and they require hours of dialogue, whereas other people are less combative and willing to stay within the boundaries without trying to instantly "revolutionize" Wikipedia. -Wikid77 (talk) 12:05, 10 August 2012 (UTC)

Possible solution: Block requirements/reform
A lot of blocks can be justified, others are down right stupid. I've accidentally been blocked for 30 minutes or so simply because one admin clicked the wrong name on a page. I couldn't even notify my blocker BECAUSE of the misplaced block. In that short timespan I learned how terrible our block system is to someone who is completely innocent. Though, the entire system is a little awkward and is in need of reform. I've heard of many of these 'oops wrong person' blocks and frankly, the unblock appeal process is not friendly or human. So here's a few ideas.


 * 1) All blocks should be accompanied by a link to a special page in which the user could object and make a real appeal that can be archived and viewed with something like the RFA toolkit. With a mere glance see how many edits, how long the user is active and specific reason the admin blocked the user. Blocking admins must reference the infraction and why they are blocking.
 * 2) Blocks should be restricted to absolute needs, not convenience, so that a block has a justifiable purpose beyond instituting a 'time out' prior to discussion. Warning templates are one thing, but bringing the matter to WP:WQA or WP:DRN should be noted BEFORE a block is done for issues of minor policy, civil, or content matters. An editor avoiding discussion or attempt to solve said problem while continuing to engage in the activity is a justified purpose for blocks.
 * 3) Socks can be legitimate, most are not. Quite a few are unintentional, a lot are intentional. Socks with >50 edits or are present on the same page as noted by a Checkuser to sway consensus should have to explain the situation or face block. Meatpuppets are not technically socks, but a couple of real life or other friends who like the same topic may seem to be socks, even if they are not. My Little Pony comes to mind. Since we are all about the !votes, meatpuppets and poor arguments shouldn't matter. Suspected socks should be watched and be more transparent, i.e. not need to go to SPI to make claim with next to no one watching or discussing.
 * 4) Vandals: accounts only to vandalize should be faced with blocks and then still have the template to explain. Vandalism must be disruptive and intentional. AGF has limits, people make mistakes. An account with 10 edits and 7 of them being vandalism is one thing an editor with 5000 edits and a rare 'vandalism' is another.
 * 5) Creation of a 'actions' page or something like that should be created. This would be open to all to view. Suspected socks should be noted by a bot or something that transcludes edits from a 'notice' by an admin. If disruptive edits/ignoring of notice continue a block can be imposed. Same with vandals, note: by Clubot cannot be used as justification for a block, an admin must review it personally after final warning. Currently, blocks seem to be issued automatically (I know otherwise) after warnings, whether or not they were justified, leading to one or two edits called vandalism to get an editor blocked. This includes editors innocent of any wrongdoing or making constructive edits to sexuality related topics. All other pending templates should be noted with template, dispute resolution form and a general list of options to view the matter.

What these changes will do is defuse the tense situations of many admins or editors. When a 'final warning' template is generated, any action thereafter usually falls to instant block category with one admin seeing it. Whether or not their edits were actually disruptive or vandalism, AGF should be noticed as editors have a tendency to warn without justification and errors with Huggle and STiki are common. If the actions were not vandalism, wipe the slate clean, close the pending and inform the editor (personally) of the situation. Make appropriate notice to taggers if they acted in bad faith or error in placing the tags leading up to the 'final notice'.

For socks, EVERYONE can watch them and ring the admin-bell for a block if they continue to sock after their discussion and explanation is required. A Checkuser will not be needed as often as the sock may go dead soon after the notice. For socks, their activity monitoring should be long term in case of resurgence via a list which can be updated automatically.

All other actions, no change, just more eyes on it. Like how everyone runs to ANI currently. Except a few additional protections are put in place for the accused. ChrisGualtieri (talk) 14:48, 16 August 2012 (UTC)