3

In another post, suggestions for improving the quality of PTIJ material were discussed. In the highest voted answer there, one suggestion was to impose a limit of how many PTIJ questions each user may post in one PTIJ season.

The post also included another detail, which was included in at least one of the suggestions for updating the policy. However, the detail of user-caps was not included in any of the suggestions (as of the posting of this question).

Is there community interest in having a cap on PTIJ questions per user per season? If so, what should those caps be?

Please only include one suggestion per answer, for voting purposes.

  • Some related chat: chat.stackexchange.com/transcript/message/28555529#28555529 et seqq. – msh210 Mar 28 '16 at 19:36
  • Sorry for not reading everything, but would someone please summarize for me why it is deemed so necessary to improve the quality of Purim Torah questions? I read the first link and I still have no idea. – SAH Feb 26 '17 at 17:20
6

This option was implemented on Feb. 24, 2017.


Low-key alternative:

Instead of making a rate-limiting rule, just include a gentle reminder to not post junk, and then remind any users who post many low-quality posts to try to adjust their quality/quantity balance.

The reminder could go at the bottom of the "What" section of the policy, and could take the form:

Please try to put as much attention into question quality, answer quality, and above all, Being Nice!, as you would for regular Mi Yodeya content.

5

I like this limitation. I think a reasonable cap would be one question per day per user, without roll-over questions.

This would both cause users to weigh their PTIJ posts for selection of their best material, and create some built-in spreading of PTIJ posts over the season, so that they aren't all flooded in at once.

PTIJ season lasts about 16 days, so this doesn't seem like an overly-restricting limit. I think only someone pushing the bounds of desirable PTIJ traffic would run into this limit.

  • How do you propose that this cap would be policed and enforced? – Isaac Moses Jan 14 '16 at 19:07
  • 1
    @IsaacMoses It could be policed by everyone keeping an eye out and enforced by custom flags or immediate mod deletion. – Y     e     z Jan 14 '16 at 19:11
  • @IsaacMoses I began to update the question to include your question, but decided not to, because I would rather not have someone's poor suggestion on implementation detract from their suggestion of the basic policy. – Y     e     z Jan 14 '16 at 19:12
  • Is this UTC date? Or 24 hours breaks? – Double AA Jan 14 '16 at 19:13
  • 2
    @DoubleAA, I would suggest formulating it as a 24-hour minimum waiting period between posted questions. That's intuitive, universally fair, and easy to check. – Isaac Moses Jan 14 '16 at 19:14
  • 1
    @IsaacMoses UTC date is also fair, intuitive and easy to check. – Double AA Jan 14 '16 at 19:17
  • @DoubleAA If the latter would be equally as easy to check, I think it would be preferable towards the end goal of uniform distribution of questions, although only marginally so. – Y     e     z Jan 14 '16 at 19:18
  • @DoubleAA, UTC as a cutoff may be more favorable in some way to people in some time zones. I agree that it would be easy to check. I think that if the goal is to reducing crowding of these questions, a measure that explicitly spaces them out more intuitively fits that goal. – Isaac Moses Jan 14 '16 at 19:23
  • @DoubleAA, also, if formulated as a gap, rather than a rate, there's no ambiguity that would require explicitly nixing roll-over. – Isaac Moses Jan 14 '16 at 19:36
  • 3
    A 24-hour break fits more fairly with posting before and after Shabas than a once-per-UTC-day rule does: it treats those for whom sunset is about 00:00 UTC (who miss a day of posting PTIJ on the once-per-UTC-day rule) the same as those for whom sunset is 12:00 UTC (who don't). – msh210 Jan 14 '16 at 19:36
4

Though I still think that the implementation of a limit on PTIJ questions is unnecessary per my other answer to this question it seems that my position is not the majority based on how voting has gone so far on this question. The currently top-voted answer proposes a minimum waiting time between posts.

If we do go this route, I'd like to propose suspending that minimum waiting period starting on Taanit Esther. The reason for this proposal is because right before Purim is when people are likely to be thinking about Purim Torah the most and if you come up with 3 great questions during the day of Purim, there's not going to be enough time to post them all if there's a minimum waiting period.

  • 1
    נתת דבריך לשיעורין The more details and exceptions the rules have, the more confusing they are, the less people learn and follow them. Very much like this other proposal of yours: meta.judaism.stackexchange.com/q/3958/759 – Double AA Feb 22 '17 at 13:51
2

I support YEz' proposal of a 1-day waiting period. Slowing people down is, I suspect, more likely to result in higher quality than putting a cap on total posts. I think a 1-day waiting period is the most intuitive, easily-stated, and easily-followed of the proposed rules.

Most of the enforcement would probably be self-regulation. If someone notices someone else asking within 24 hours of their last PTIJ question, they can comment, and hopefully, the over-poster will respond by temporarily deleting the new question and then reposting it after waiting long enough. If someone flagrantly over-posts or ignores peer suggestion, mods can step in and delete as necessary, perhaps with an eye to preserving whichever posts are higher-quality.

Here's a possible formulation for insertion into the PTIJ policy:

Add to the "When" section:

After posting a Purim Torah question please wait 24 hours before posting your next Purim Torah question.

Add to the "Moderation stuff" section:

  • In cases of flagrant over-posting of Purim Torah questions within 24 hours of each other, moderators may selectively delete questions to bring a user closer to one per 24 hours.
  • +1, I like this very much, simpler is better usually and this is much simpler than what I proposed. If we made it two days, we'd get to my 7 question limit but maybe this is pushing too far :-> – mbloch Mar 29 '16 at 15:16
  • 1
    Although maybe a one-hour rule would be sufficient. – Isaac Moses Mar 29 '16 at 16:11
  • 1
    I like this. I'd tweak the moderator guidance to say "...down to an average of one per 24 hours", so that if there do happen to be two good ones on the same day (but a bunch of poor ones on other days), we aren't forced to delete the good ones in favor of the poor ones. Also, we might prefer to delete unanswered ones over answered ones, if the answers are positively-voted. – Monica Cellio Mar 29 '16 at 18:00
  • 1
    @MonicaCellio How about "to bring the user closer to one per 24 hours."? I think we needn't specify what criteria mods would use to select questions to delete and just trust y'all to look out for the good of the site. If a user is really sad about one of the deletions and would prefer to substitute another question, they can always appeal. – Isaac Moses Mar 29 '16 at 18:35
  • 1
    @IsaacMoses yes, that's a good way to phrase it. I don't want to overly specify the criteria either; it won't happen often anyway, so when it does, mods should use their best judgement. If somebody doesn't like a specific choice, there are ways to let us know that. :-) – Monica Cellio Mar 29 '16 at 18:46
  • 1
    @MonicaCellio I thought the idea was to slow down posting, not to permanently delete a question. It can be undeleted once the waiting period is up, but deleting a bad question from 4 days ago does little to reduce the clutter of bad questions. I'd say that moderators could informally be lax in enforcement of this rule if the latest questions seem to be pretty good (regardless of the quality of previous questions). – Daniel Mar 31 '16 at 20:00
  • I'd love to see this restriction be lifted in the last few days of PT to allow people who think of a bunch of great questions on and right before Purim to post them. – Daniel Feb 16 '17 at 14:56
  • @Daniel I'd suggest that you write an alternative answer including that provision. – Isaac Moses Feb 16 '17 at 14:57
  • 1
    Done! – Daniel Feb 16 '17 at 15:08
2

msh210 provided some very helpful analysis from the 2015 Purim season (here with user IDs). In the context of question limits, I see the following take aways in that data

  • asking more than 7 questions produces low overall quality (but sample is small)
  • the top 6 users by average score all asked 2 questions or less
  • if one goes in the detail of the questions asked by the top posters (by nr of questions), one sees some very good quality (high scores) and lots of low quality
  • a limit to the number of questions would imply that users with lots of questions would be able to guess in advance which of their questions are strong - otherwise we will lose strong questions

In this context I would be very supportive of

  • limiting users to 5-7 questions per Purim season
  • or possibly, doing the same with an exemption that as long as one maintains a positive question record (e.g., average vote per question >=4), one can keep posting more questions

The deletion of question with negative scores could be done once a day by the moderators. Enforcing the limit of 5-7 questions could be done via flags from the whole community.

Note that DoubleAA came up earlier with a similar policy proposal strengthened by the idea to remove questions with negative scores after e.g., 12 hours.

What does the community think?

  • 3
    I like the positive question record idea. It allows users who are good at the game to keep playing, while limiting users who aren't so good. And, coupled with AAAA's idea of deleting negatively-scored questions, we might see more good. – Scimonster Mar 28 '16 at 19:40
  • How would you recommend this policy be enforced? I.e. by whom – Y     e     z Mar 29 '16 at 0:42
  • @Yez the deletion of question with negative scores could be done once a day by the moderators. The limit to 5-7 questions could be done via flags from the whole community – mbloch Mar 29 '16 at 3:42
  • For point 1 "asking more than 7 questions produces low overall quality (but the sample is small)"... The sample size is exactly 3. There are 3 people who have ever posted more than 7 PTIJ questions in a single year (actually 2 because two of the results are the same user in different years). I don't think there's anywhere near enough data to support your claim. Making this limit would essentially be a rule targeting one or two users which I think is a bad thing. – Daniel Mar 29 '16 at 4:27
  • Point 2 is simply the law of small numbers doing its thing. If lots of users post one or two PTIJ questions and a couple of them have hit questions, that explains this phenomenon. It's unsurprising that asking more questions results in less extreme average scores (the bottom 8 users by average score also asked <= 2 questions). – Daniel Mar 29 '16 at 4:30
  • @Daniel I think the question some are struggling with is how to avoid lots of low-quality material during PTIJ season. The fact that some users will post 1-2 non funny questions is unavoidable. The question is how to avoid some users posting lots of non-funny questions. You make very good points about statistical validity (and over in chat about the risks of having rules for 1-2 people). The challenge remains though – mbloch Mar 29 '16 at 4:38
  • @mbloch My point was simply the statistical validity point. I was just pointing out that the same reasoning that you have applied in point 2 could equally apply to argue that all users who post PTIJ must post at least 3 questions. – Daniel Mar 29 '16 at 4:43
  • @Daniel I hear you and you are right. I still believe question limits will help minimize noise, and that allowing users with strong question records to keep posting will maximize the fun to noise ratio. I wish the data would help make this a stronger case but it doesn't appear to be the case – mbloch Mar 29 '16 at 4:45
  • @mbloch I agree with you in theory (not that we should do this but that it theoretically should reduce noise). But the fact that the data doesn't support the theory is a strong reason not to. If your proposal had been in effect since the beginning of PTIJ, it would have affected exactly 2 users ever. Given the significant amount of effort and bureaucracy that such a policy would take to enforce, I think it's simply unreasonable. Like I said in chat, "Are we going to run a SEDE query every time a user asks a question?" – Daniel Mar 29 '16 at 4:51
  • To be clear, I think the rule with the exemption is a generally good idea but that it would take too much effort to enforce. – Daniel Mar 29 '16 at 4:55
  • 1
    @Daniel impacting few users is good. Ideally we'd like to reduce much of the noise while impacting fewest users. Monica said there were 100 questions this season and the top 2 users generated 35 of them! On effort to enforce, deleting negative score questions once a day is not a big deal, flagging questions over quota can be left to the community. But maybe someone will come with an even better idea that will be easier to make happen – mbloch Mar 29 '16 at 5:01
  • @Daniel Note that the 2 users it would impact make up about 35% of total PTIJ questions for this season... – Y     e     z Mar 29 '16 at 5:09
2

Why I disagree with a limit


A premise of this proposal is that there is a negative correlation between number of questions that a user asks and the quality of the question (or at least that this is true for some number of questions greater than N). The problem is, the data do not support this premise. I created a SEDE query (h/t to msh210 on whose query this one is based) that plots the average of all of the average question scores for a given number of questions asked (e.g. if there were 3 people who answered 5 questions and the average score of the 5 questions was 3, 4, and 5 for the 3 users respectively, the average score for 5 questions would be 4). Looking at the data, there is AFAICT no correlation between number of questions asked and question quality for N <= 7. But the data set for N > 7 is far too small to make any generalized statements (seven questions in one season has only been exceeded 3 times ever and two of those times were by the same user). I take a couple of points out of this:

  1. People generally do a pretty good job of self-regulating
  2. Any number that is chosen for a hard limit would be completely arbitrary. We have no reason to believe that it's impossible (or even particularly difficult) for someone who asks 15 PTIJ questions to maintain a high average score.
  3. Related to the previous point: any hard limit would basically be targeting one or two individual users. I think that's generally a Bad Ideatm. It's not our job to make regulations for individual users. This is really just a sneaky way of telling a couple of users to quit posting so much.

One proposal here suggests an exemption to the hard question limit. I think this idea avoids most of my issues in theory; however, I still do see a few problems with it.

  1. The proposed minimum average score threshold of 4 could be difficult to maintain, especially if someone asks one question that he or she thought was funny but turned out to be a dud. A score of -1 or -2 can do some serious damage to an average of 6 or 7 scores.
  2. Enforcement is difficult. It would require someone to manually tally a user's average score whenever he or she posts a new question. That's annoying. Plus, if that information is not communicated publicly, multiple people will end up doing this. Finally, that tally becomes outdated every time someone votes on one of the user's questions.
  3. It's easily circumventable. A person who has been capped can simply delete some posts and write some new ones. This doesn't help reduce noise in any way because by the time the question is deleted, the damage is already done.

My Proposal


I propose that we have no limit on PTIJ questions. The data don't support it and it would really only affect a very small number of people. Instead, when we see that someone is posting lots and lots of low-quality PTIJ questions, let's just point them to this discussion. People generally seem to be pretty good about self-regulating so hopefully they would comply once they become aware that people are getting annoyed. If they do not tone it down at that point, I would not be opposed to limiting that user from posting more PTIJ questions. I simply think an arbitrarily-chosen limit won't help.

2

Note that this answer was penned before the 2016 PTIJ season.

I suspect that the proposed change to the PTIJ policy, in a form like Daniel's or Isaac Moses's, will suffice, and no question limit will be necessary. We should not impose one now. If we find (perhaps this year) that those changes did not suffice, we can impose a question limit for the following year (or make some other change: whatever we decide at that time).

The downside of a question limit is, of course, the possible loss of some good questions. Oh, and the annoyance of some users.

  • 1
    Out of curiosity, do you still hold by this answer? Or I guess I should ask, were your suspicions confirmed or refuted? – Y     e     z Mar 29 '16 at 0:47
  • 1
    @Yez, I don't think the current rules sufficed for this year, but I don't know that a question limit is the right fix. I'm sure we as a site will discuss this some more over the next almost-a-year. – msh210 Mar 29 '16 at 1:22
  • You might want to incorporate this into your answer. – Daniel Mar 29 '16 at 4:34
2

Hybrid proposal:

  1. So long as your average score per question this season is at least 4 (I'm open to changing this number), no limit. People who are good at the PTIJ game can keep doing it. (Inspired by option 2 here.)

  2. Otherwise, you must wait at least 24 hours after your last question, per this answer.

Monitoring would be by the community; if you see what you think is too many PTIJ questions from somebody, check and speak up as needed. You can search questions by tag + by user + within a date range, and then do a little arithmetic.

I don't have specific policy verbiage yet.

  • This seems too complicated. Who keeps track of their current average score to that degree? Maybe a strictly positive average is doable. – Double AA Feb 22 '17 at 18:48
  • @DoubleAA people who think there's a problem take on the burden of checking and reporting, same as with flags. A search for user:### is:q [ptij] created:2017 produces what should be a short-enough list to scan and add up, and you get a total at the top of the search results to divide by. – Monica Cellio Feb 22 '17 at 18:53
1

Since I seem to be the main "culprit" this year, I'll chime in...

First, my apologies if I was a nuisance, this year.

I'm not great at stats, so I have to implicitly trust Daniel's expertise, here. (It's always hard for me to say no to someone who has the same 1st name as me, anyway.)

I think that when people see downvotes on many of their questions, it tends to discourage them from posting that quickly - most of the time. I think that in my case, there were a few days when I posted several questions, and some days, esp. almost when approaching Purim, when I posted none at all. Personally, I have no problem if a moderator steps in and requests me to stop. I also have no problem with the one per day disclaimer. I think that most people are willing to follow a policy that's pretty clearly stated esp. in a disclaimer. For the few that really don't obey the policy, you can always contact them, directly.

BTW, doesn't the system auto delete questions that have a huge negative vote number? In most cases, when I see a large negative number, I usually will delete the question, myself. Depends if I notice / remember and also I will read the comments to get a sense of why the negative votes are there. (Sometimes, but rarely, I strongly disagree, and will still leave the question there even with a low number.)

While I don't suggest that moderators close or delete a question that is lower than some number (say -5 or less), I have no problem if any of you want to do that with any of my questions, whether PTIJ or not. So, go ahead - consider me your "exception"!

  • The system only auto-deletes highly downvoted questions if they're also closed (it also deletes moderately and slightly downvoted questions that are closed) – Daniel Mar 31 '16 at 19:51
  • Also, I think the problem that people have isn't that the questions stick around. Deleting questions doesn't help because the thing that people don't like is the "noise" of lower quality questions being asked in the first place and taking up valuable space on the front page. By the time they've moved well down the front page, deleting them doesn't make much difference. – Daniel Mar 31 '16 at 19:53
  • @Daniel After writing my idea above, I realized that the vote count changes, anyway. Maybe a moderator should delete a low voted question that remains that way for a few days? It's very hard for a writer to realize a poor-quality question at the time he writes it. If he did, he wouldn't write it in the 1st place. What did I miss, here? – DanF Mar 31 '16 at 19:54
  • I think the idea is that perhaps with a bit of thought it's not really that hard to recognize a poor-quality question at the time of writing. The people downvoting don't have any secret information that the OP doesn't have. – Daniel Mar 31 '16 at 20:07
  • @DanF I like your Purim Torah questions and have no idea why people have a problem with their quantity or quality. – SAH Feb 26 '17 at 17:17
  • @SAH Thanks, and I like my own stuff too! But, if I don't keep a level head before Purim, then having a big head on Purim itself, won't mean much. Hence, a bit of restraint, probably isn't such a bad thing, esp. this time of year. – DanF Feb 27 '17 at 1:06

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .