It comes down to a question of bandwidth. People are submitting problems at a faster rate than you are willing to expose them on the front page. You have several options: increase the rate at which problems appear on the front page, allow fewer problem submissions, be more stringent in dumping problems, or allow problems to go live without necessarily appearing on the front page. Of those options, I would favor the last one, since I think it's best to get as many problems on the site as possible, but still allow the best problems to shine.

kuro5hin.org has a queue somewhat similar to this one with four voting options: Front Page (+1), Section Page Only (+1), Abstain (0), and Dump It (-1). You could adapt this strategy so that a TU could count either toward the front page or toward a particular puzzle category. 3 TU's in the front page option qualifies the problem to go on the front page, while 3 TU's in either option qualifies the problem for an appearance in a particular section.

There are two drawbacks that I can think of to this approach. One is that a good problem might get 2 front page TU's and 1 category TU, but then quickly get pushed to the category page when it might otherwise eventually qualify for the front page. Another issue is that a problem might quickly qualify for a category page and not get seen by many reviewers, which opens up the possibility that duplicates might slip by. Both of these would be fixed by requiring that problems spend a minimum of, say, 48-72 hours in the queue before they can be pushed (no minimum for getting dumped).

To get non-front page problems more exposure, I'd augment the Categories section to display something like "Riddles (2 new)", rather than just "Riddles" if two riddles had been pushed in the last 24 hours.

What do you guys think?

I still have mixed feelings about how to clear the queue more efficiently.
On one hand, it seems voters are too reluctant to vote TD on a problem unless it is a duplicate, or an extremely poor problem. As a result, seven of the ten problems in the voting queue are waiting for changes to be made or for more people to vote one way or another to move it, and the pages on the front page are becoming mediocre and uninteresting for discussion.
On the other hand, to new users who have submitted a few simple problems and waited a month or two just for the voters to see them, it seems rather harsh to TD them.
One way to move out a few 'lesser' problems without having them necessarily spotlighted on the front page, with the existing implementation, is to submit two problems one right after another. Then, the first problem is still made live, and found at the bottom of the front page as well as the digest page (which, by the way, I do look at), but it may not recieve as much attention as some problems that have been deemed 'better.' If we made this a regular practice, three or four problems a day could be made live, with one 'good' problem to be the feature problem for each day. That might be preferable, too, since (I'm assuming) many people get on only once a day at around the same time, so if two problems go live at different times of the day, they may be missed. I was also thinking something along the lines of what Gamer suggested, to make more recent problems besides the single latest more prominent on the front page, either by putting more than one problem in full text, or perhaps just displaying the two or three next-most recent problems above the 'Oldie' problem. I think either of those would make people more likely to look at more than just the current problem. Just some thoughts to throw around..

I offered an idea like the 'abstain' vote (with a different name) a few weeks ago in another thread, but no one commented on the idea then. Now that I can see the voting queue, though, it seems like that would only further impede the process. Also, I really think the minimum queue time is not a good idea, for two main reasons: first, we are trying to make things move more quickly. As it is, it usually takes a day or two for all the voters to see it and vote and quibble and whatnot, if not longer. Second, I don't like overexposure to the problems in the voting queue before they are made live for everyone to comment on and solve. It seems now that for even the lengthiest problems you cannot get to a newly posted problem without Charlie or someone having already posted a complex solution within a few minutes or hours of the problem being made live. While of course anyone is still free to comment and solve the problem on their own, going to a newly-posted problem and finding three comments that say "Solution" I think discourages people from commenting or even attempting the problem. If you look at many of the earlier problems, there is a lot of discussion and attempts at the problem before anything was ever concluded. Any site can list a bunch of problems with a solution; to make these problems 'live' and open for people to solve, I think it's important that people have a chance to look at a 'fresh' problem or one with only a few thoughts in the comments, or they will be discouraged from saying anything at all. I know that for me, it's disappointing to see a potentially interesting problem like the recent "Quaker Queens" problem, only to find that Charlie has not only posted a solution, but every possible solution, along with a 120-line program to find them. I found a couple arrangements with pen and paper and a little logic...but after that, why bother? .. That's pretty much an unrelated issue, except that moving problems through the queue to be live more quickly will hopefully inspire a little more discussion. A lot of it, too, is simply discretion on the part of those with access to the voting queue. It shouldn't be a race or competition to solve the problems first, but a collaboration and discussion at finding a solution.

Back to topic:
Personally, I think the voting queue should be more of a voting queue. As it is, it seems more like an editing team or something; throw away duplicates and tell authors how to modify the problems so that they will read better on the site. I think, if you don't like a problem, vote it down. There are a lot of better problems in queue, just waiting to make it into the voting, while we fiddle around with a slew of "okay," uninteresting problems. This might also be less of a problem if we make more problems live more quickly.

Those are my thoughts.

One of the things I often tell people who submit brute force solutions is "How could you use logic/number sense/smarts to figure this problem out, rather than trial and error (which is what the computer does)". I actually even put this in "Squares Probability", as that is the whole point of the problem.

Saying how you figured the problem out is almost more important than the solution (if not more), because knowing the solution feeds the man for a day, where knowing how to solve the solution feeds him until he gets tired of fish. :)

And I agree with the voting queue, the only problem there is being deadlocked with problems like Erin's airplane problem.

For the problem of 10 people collectively owning the queue, I don't know how difficult the following would be to implement, but my suggestion is that no problems be promoted to queue weight 1 until there are no other QW 1 problems left in the submission queue. The problem with this is that it might take a long time to get back to particular individuals, even if their submissions were made quite a bit earlier. Even more complicated would be combination weighting schemes involving recency of submission along with the present queue weight criteria.

These are all pretty interesting proposals that I will have to think on.

One additional thing occured to me today thogh that I'd like your input on:

It seems that a lot of the "lesser" problems we get are submitted by people who are trying to get ahead on the site. To that end, I think some of our volume may be alleviated by enabling a couple of policy changes:

1) Require a person to be a Studend before being able to submit a problem. This makes sense especially since Novices don't really have the ability to do anything with problems they have submitted - plus I believe that letting people submit on the first day is not the greatest of ideas. Let them hang around on the site and get a feel for it - it can take as little as 24 hours to become a Student here anyway (mind you I don't believe the way of becoming one should be advertised lest people start posting just for the sake of getting the rank)

2) Somehow enable people to get promoted without having submitted so many problems. Perhaps a nomination scheme: a journeyman would be able to nominate a Student to become an apprentice - given that the Student has met certain requirements (for example - a week of membership + X number of comments/ratings) After a Student has 3-5 nominations, a Scholar will be able to promote them. The process of becoming a Journeyman would still require problems on the site, but perhaps with a nomination, it could be set to lower than the current 6 - you could become a Journeyman with 3 problems + 5 nominations for instance.

What do you think?

DJ - your proposed system seems to me to be a step in the wrong direction - you are proposing we say that a comment is worth X ratings, and a problem - X comments. I think that just as you said, there's not a unified system that would work well, especially one based on points or any kind of rigid criteria.

For example, The Mighty Puck has come pretty close to becoming a journeyman by submitting nothing but riddles pasted in from The Hobbit.

The system of nominations I am proposing would probably do away with a need for a rigid and complicated structure of promotions, by replacing it with a human element.

I just wanted to end the idea that you have to submit problems to advance, and hence a lot of newcomers are submitting poorer problems merely to gain the next level. My point it, it is possible to be a contributing member without submitting problems, and in truth the members that I consider to contribute the most are not the ones that submit the most problems, but the ones that leave comments to more problems and more frequently. That's why with the point system, I suggested extra weight or consideration on the frequency of comments, as well as their number.

Personally, I think that while we don't want people doing just a poor minimum to earn a rank, the process should be more objective than subjective. People will want to know (I think, at least) that their contribution (comments, ratings, submissions) to the site are the basis for advancement, not the personal approval of some gentry. However, I can see that no matter how you change the criteria for advancement, users like Tim Axoy will be able to jump through the proverbial hoops to automatic advancement unless there is a human element involved.

Also, I recognize some potential problems and abuses with the method I suggested, so I have a few more suggestions to make that incorporates the ideas that everyone has pur together here (I recant whatever previous statements I made, and while I'm still sketchy about the idea of nomination/approval for advancement, it seems to be the most favorable way to go, and I'll stick with it).

First, to address the flooded queue that started this thread, I still think that the option to delete one's own problems would greatly diminish the queue, especially those by new users who submit a few problems just for the sake of advancement or for whatever reason, and then decide that they are not up to par with the site, or later realize that is a duplicate problem. Speaking from experience, the first few problems I submitted were not very good (although a few made it through the voting queue anyway), and three or four were duplicates of previously submitted problems (one of which was made live while mine was still waiting). Allowing people to delete their own problems (with, of course, due confirmations and warnings against accidental button clicks) I believe would both reduce the queue size and raise the quality of problems submitted, or at least those reaching the voting queue.

My next suggestion is a rearrangement of the user levels themselves. I agree that it would be prudent to only allow problems to be submitted to Students who have been around the site for at least a little while, and then to only allow apprentice users to modify submitted problems is fine. I did not see the reasoning behind disallowing the category change at first, but I assume it is to prevent people from submitting blank problems to save a spot in the queue, or later changing a problem to a better one (as I noted in a related thread, I did that, unwittingly violating the site's unwritten lawbook, because I noticed a duplicate problem and did not have the ability to delete it). If that is the case, I would suggest that Apprentice users would be allowed to change everything about a problem except for its title. It would be hard to change a problem to a completely different one, or to pre-save a spot in the queue, when you could not change the title to a problem once it has been submitted. Also, when a problem reaches the voting queue, a change in category is suggested by the voters much more often than a change in the problem's title (except for things like 'Leg Byes,' but that's a different story...), so that change would mean requiring less manual intervention from levik. I think we should eliminate the requirement on problem submission/approval for Apprentices altogether, replacing it with a standard based on how long a person has been a member and by how often they contribute to the site. Advancement this far, I think, can be done automatically, without the need for voters or higher-ranked members to intervene. Alternately, the human intervention in this step could be minimal, such as merely requiring two members ranked Journeyman or higher to click a button that says 'Approve' for qualifying members. Really, though, I don't think anything of that sort is necessary.

My last suggestion is mostly redundant of what others have said. To become a Journeyman, a user would have to be a member for at least a month or two. I would say an 'active' member, however that would be established, as I maintain that the singlemost determinant or most telling measure of a person's contribution to the site is the frequency and quality of their comments. In any case, I think requiring someone to be around for a length of time, as someone suggested earlier, would be good to make sure that not only is someone contributing to the site, but they are interested enough to stick around.
Then, they would have to be nominated by a Journeyman or Scholar (or Director, although I'm sure 'worthy' users will be noticed before levik feels the need to step in). This step would probably eliminate the need for any measurement of how often a person contributes, since I imagine that they will need to be quite active for someone to notice and nominate them. After nomination, there should be some sort of deliberation, perhaps in a separate voting queue, with a minimum time, say a week, for people to discuss and evaluate a person's contribution to the site, and what they would offer to the voting queue. The ultimate decision should be somewhat different; rather than a sum total of TU or TD votes, I can think of two ways to go. First, voters would have to reach some sort of unanimity (like a jury), or perhaps allow for only a single dissenting vote (no more than one 'nay' for every six 'yeas' or something). The other way to go would be, after the week of discussion, levik would have the final word. As to the method of deliberating..I think it would also be useful to allow nominees to speak on their own behalf, and to say what they think they would contribute to the voting process.

In sum, as everything I have just said is very long-winded (I wrote it, what would you expect..) and confusing, here is my suggestion for the new ranking system (much of which is copied from the current 'levels' page).

Novice
This is the basic level, assigned at signup. Most users are currently in this category. Novices can rate problems and post comments and solutions, as well as read and post to selected discussion forums.

Student
A Novice is promoted to Student if they have rated at least one problem and posted at least three comments [as is currently the case]. Students can submit new problems to the queue and modify the text, difficulty, and category (but not the title) of problems before they are approved and made live on the site. Students may also delete their own pending problems before they are approved, as well as add and edit the solution to their own submitted problems before the solution is posted to the live site.

Apprentice
A Student becomes an apprentice if they have been a Student for at least two weeks, and have posted at least 11 problems with timestamps on at least four different days during that time. [Tthat is the easiest way I can think of to put some sort of weight on how often a person visits the site as well as how much they have to comment. Two days a week is certainly not excessive. Further, I don't know why eleven comments, but it's the current standard, different than the overused standard multiple of five, and I like it.]

Journeyman
Journeymen are the stewards of the site. They have the ability [as now] to decide when to make the solutions to their own problems public, as well as access to problems submitted by others, which they can vote for giving them "Thumbs Up" (+1) or "Thumbs Down" (-1). An apprentice can become a Journeyman only if they are nominated by another user who already has the rank of Journeyman or higher, based on their contribution to the site through comments, problem ratings, forum discussion, and problem submissions. After being nominated, a special voting forum for Journeymen, Scholars, the Director, and the nominee to deliberate, after which time frame the voters must reach consensus, allowing for only one dissenting vote. If there is more than one vote agaisnt their promotion, a nominee is rejected (which would rarely happen, and only if there are very serious objections to a person's nomination), after which the user would again have to wait two weeks before being nominated again. At the expense of seeming elitist or clique-ish, we will reiterate that the responsibility of selecting problems to post on the site is an important one, indeed the foundation of the site itself. As such, there is no accurate rubric or criteria that can ratea user's contribution to the site, besides the other users themselves.

Scholar
Scholars are able to edit their problems that are "live" on the site (in addition to pending problems). They can also change the status of their problems from "pending" to "live" and back. Upon submitting a problem to the website, they are given the option of making this problem live immediately. Scholars can now also make "live" problems of others that have reached a score of 3 through voting. Like Journeymen, one cannot reach the rank of Scholar automatically. The Director of the site will manually promote Journeymen as needed based on their overall contribution and judgement. There is no need for more than a few active Scholars to help run the site, so this promotion does not happen often.

Director
The Director is the administrator of the site, given access to backend administration tools. Currently, the only Director on the website is Levik, and it will probably stay this way for a while.

In regard to the Queue Weight reworking, it would seem appropriate that no more than two Riddles get queue weight 1. Right now, Riddles seem to be building up in the queue, as they arrive there in the top 10, but with so many recently, they tend not to get pushed, and just build up in the queue.

Coming back to the original question, I believe that the only long term answer is to raise the bar on what's posted. This gives two major benefits: firstly, the overall quality of the problems (and therefore the site) improves - to use Gamer's expression the average "goodness" of the problems increases; and secondly the queue diminishes through not only the initial deletions, but also through fewer submissions as people decide not to bother submitting poor problems.

If you try the alternative route of simply increasing the rate at which problems are published then you are actively promoting quantity over quality which when combined with the current promotions system encourages some people to submit poor problems. Ultimately this choice will cause the queue to grow rather than reduce.

I'm a firm believer that what was once excess will become the norm. In Flooble's case this can either mean that today's poor problems (eg I am not a knight) become typical of the site; or that today's absorbing problems (eg Mind Boggling) become merely average. I guess I know which direction you'd prefer things to go and as uncomfortable as it may feel at the end of the day it's up to the Journeymen to become more ruthless to prevent the downward spiral.

There already is a "goodness rating". It's just that people rarely seem to use it, despite levik's flashy JavaScript stuff.

Sorry if I have missed something obvious, but I don't know of where the "goodness rating" is... (If it would be difficulty, I think I addressed that above. /|\)

The reason I wanted the goodness rating put in is to make the promotions based on problem quality, not problem quantity. This would make an boring problem (like I am not a knight) to be worth less than a better problem (like Letter cubes or Four digit number) would be, such that you could have 6 or 7 boring problems and still not be a journeyman, but you could submit 4 or 5 great problems and be a journeyman. (If we wanted to base it on problem submission.)

The problem is that it's frustrating to submit a problem, only to have to wait months for anybody to even look at it. Often, people will simply stop visiting the site by the time their problem becomes visible to reviewers. This makes it impossible to get last-minute fixes into problems (or sometimes to just get a solution).

If the problem is essentially the delay between submitting and posting i propose to review problems from the beginning of the queue, not the end.

This method fights flooding in the following way

1) invalid problems are deleted quickly, and will not pollute the queue. this applies both to low quality, and also duplicates. additionally, the problem with duplicate order inversion is eliminated (a concern voiced by ravi i think)

2) users receive instant feedback about their problems. this ensures that they
adapt in a timely fashion to the site's standards, and will not submit crap which quickly bounces back to them. the opposite also holds, users will know if their puzzle was in fact accepted and will be encouraged by it promptly. (even if it wont actually be posted promptly, altho they could be informed of ETA)

3) related to 2), problem acceptance can be better used as a metric for user contribution, because acceptance does not merely imply pressing a submit button, but an actual review. the concern that users spam problems to gain promotion (stupid as this is) is therefore addressed.

4) if to better control queue velocity (QW) for problems, review at the beginning of the queue can be used to control how fast a problem surfaces depending on how good it is. (this is related to the idea of rating problems outright in the review, appart from TU/TD which someone already suggested here) this mechanism also results in more feedback to the poster.

5) last minute fixes and problem editing is generally possible because the user will normally be around shortly after posting (this problem was suggested by fl above)

6) if the queue is still large, problem reviewers may have forgotten the problem since their review and may have more fun with it when it actually comes out!

The other end of the queue is only used to mechanically push the problems live, and perhaps control uniform category distribution. Anyway, just a thought.

Actually, my previous comment referred the two problms before the two preceding it, but this thread is moving faster than I am.

As to Cheradenine's (btw, hello, I haven't seen you around here in a few months) suggestion, I can certainly see the reasoning behind the theory, but in practice I'm not convinced it will work as intended.

First, The main reason for this is the amount of deliberation that each problem recieves in the voting queue now, not to mention the fact that to actually move the voting to the beginning instead of the end, there are still 500 problems that we need to go through before that is accomplished.
If the voting queue were still limited to ten problems at a time, then we would have the same bottleneck effect as we are seeing now, where a hefty chunk of problems sit there for quite a few days as the voters and the author debate the problem and any changes that need to be made. During that time, more problems would be submitting, and would only trickle slowly into the newly-moved voting queue, likely only to suffer the same sort of debate.

Second, if the size of the new queue were not limited, I can picture only chaos, with piles of pending problems, some requiring debate or editing, some simply requiring more votes one way or another. With the ten problems in queue now, I have trouble keeping track of what ones have been decided on, what ones have new comments or changes, etc. That could possibly be alleviated with "new comment" or "new change" indicators, like the red "!" next to updated threads in the forum, but I imagine that it would soon be found overwhelming.

Third, once problems have made it through the voting queue, they sit for just as long waiting to make it onto the site. That is, unless we push more problems live daily, and I'm sure most would agree that such is not a favorable solution. Also, there are a few problems with having an extensive queue of approved problems waiting to be made live. First, there is too much exposure to the problems for the voters, a problem even with the short-term queue as it is now. That would even be a problem if the problems were hidden once they were approved, I imagine.
Also, if that were the case (approved problems were hidden), there would likely be some issues with duplicate problems for which it would be necessary to look at some already approved problems ("This sounds familiar, is it the same problem as that old one?").
Whether or not reviewers recall the problems, I think it is still best to make them live as soon as possible after the voting.

Lastly, due to the wait now between problem approval and problem posting, it would be necessary to make sure that a solution is provided with the problem before it is approved, especially with newer users who aren't guaranteed to be around when the problem goes live. Levik can always manually enter something like "fl came up with a good solution here" as has been needed in the past with an absent author of a problem, but that, of course, is not favorable either.

There have been a number of suggestions regarding how to better move problems through the voting queue. I certainly agree that something could be done. However, I think we first need to deal with problems that are being submitted.

Hi DJ

The strategy i propose requires reviewing rate to match submission rate. Funnily, this requirement is made possible through the application of the strategy itself, because it should reduce submission rate and increase rejection rate. (For example, the cost of discard/deletion for further editing/fixing is almost zero, by virtue of the strategy itself). If it were the case that rates were in fact disparate, responsibility can be finally deferred onto director (as was in the beginning anyway) to ensure prompt resolution of inputs if journeymen etc could not cope.

Your third point assumes that the queue would be just as long when reviewed up front, when it could be reduced for just that reason. If in fact it were still long, at least there would be no trash, a problem which most suggestions here have been aimed at anyway.

As to the availability of solutions, the situation would at worst be exactly as it is now, but normally better because the time of review coincides roughly with submission.

It seems the good ideas are starting to gel. I really like DJ’s proposal for a user’s TD backlog, with those problems going back to the front of the voting queue when the user edits and resubmits the problem. That way, the voting queue gets weeded quickly, and editing is done on the user’s own time. In addition, I propose that puzzles that don’t get voted in within a certain period of time, say four days in the voting queue, get kicked out automatically to this backlog list. If Journeymen can’t decide they like a problem in that much time, it probably isn’t a great problem in the first place, and if the lack of voting was just due to a holiday weekend or something, the user can resubmit the puzzle, no harm done. This brings up a good point, however: a resubmitted problem from the backlog list should have the original comments attached, so Journeymen can tell if the areas for improvement were addressed. In any case, this proposal allows puzzles to be quickly approved or kicked out, so more puzzles will pass through the voting queue in a week. When Journeymen are active, I can imagine 8 of the 10 puzzles in the voting queue being new each day.

As for the issue of puzzle quality, I feel the TU/TD should be replaced with puzzle ranking, as proposed by fwaff et al. A puzzle receiving five or more numerical votes averaging at least 3.5 is in, 5+ votes averaging <3.5 is out, and less than 5 votes by the 96th hour in the voting queue causes it to get kicked back to the backlog, where its voting history is erased -- after resubmission, the puzzle should start with a clean slate and be judged based on its merits after being edited.

I really do like bryans idea as well but instead of getting rid of tu/td and having the time thing, we shoukld just put in and "edit vote" so it will be thumbs up or thumbs down or edit or note. after 3 edits and a time period expires, then the problem will be kicked back, an automatically generated e-mail will be sent and if the author does not edit it by the time it gets in the queue again, it is deleted.

When I right out my problems, I always seem to make loads of mistakes. But when I check it I don't find any. After about a fortnight if I check it again for some reason I can suddenly see all the mistakes. Perhaps if the problems were arranged in queue order in 'my problems' it would be easier to edit problems before they reach voting.

PS I know this isn't related to the topic of this thread, but it is linked to the comment above.

I agree and have said in the past that we need to be more discriminate in the problems that we let through to the site, and the only way to clear the queue with a fixed rate of posting is to TD more problems, and people are often reluctant to give a mediocre problem a vote down.
However, I think you have crossed the line and are being very much overcritical. Fine, you don't like L&K problems, or deductive logic problems, or cryptography problems -- but lots of people do.

As far as the queue backlog, Ravi still has problems coming into the queue from February, and the rest of the active voting queue is composed of problems submitted in June and July. It's great that you've submitted problems, and I'm sure they're good ones, but above all you need to learn the same patience that everyone else has. It's not a stranglehold on new members; people that are really interested and will stick around for a few months will see their problems (pending approval) make it onto the site.

I'm sure that once you've been around for more than three weeks, if your problems are as good as you suggest, you'll become privy to the voting queue and company car and all the wonderful benifits of being a flooble journeyman. In the meantime, please try to realize that the purpose of the site is to post interesting, challenging problems for a lot of people, and while we don't want to go the way of other sites in promoting quantity over quality and ridiculously easy problems, we want a variety of problem types and difficulties. Generally, the 'regular' users are the ones who vote on each problem, it seems, and the simpler ones also seem to generally recieve a lower score. That doesn't necessarily mean that they are bad problems. Also, a good problem can seem ridiculously easy once the solution is known, but that doesn't mean that it wasn't reasonably challenging to begin with. Just some things to consider as you [patiently] await voting on your problems...

It's easy to complain without having a solution. What would you suggest be done? I haven't taken mass submission to the extreme that Ravi has, but I have a number of problems waiting, each of which I honestly think is a fairly good one. Are you suggesting that only the first two problems I ever submitted should have QW1, and once those are posted, the weights should never be recalculated? That's the only way for what you suggest should happen could happen, that a new user could come in and submit a problem that magically jumps to the front of the queue in front of problems that people who have been here for months submitted months ago. And, once your first two problems have been laid to rest, I think you'll agree that that's not exactly the best way to go either.

I, for one, like L&K and deductive reasoning and crypto and stupid riddles for that matter. As I've already stated elsewhere I especially like those "who is in the green house?" grid problems that seem to be so loathed by the kids who sit at the cool table.
For the record I have been waiting ( I believe longer than SK) for the posting on my first three problems. They are not great massive mindblowing mots of mathematical mastery but they are cute and amusing (like me in my wife's estimation) and I made them up myself rather than cribbing them from somewhere else. That being said, I have a message for any scholars reading. If you find the fatboy's offerings lame, please do not hesitate to vote them down. I will not be offended (well, maybe a little, but not so much it'll show). I'd rather have everything I ever submit turned down than lower the quality level of the site.
BUT, dear scholars, please do not be brow beaten into restricting any particular types of problems. Yes, a site made up entirly of word melds would probably drive some folks away, but don't turn the site into some sort of number theory text book either.
I LIKE the site as it is. I think you've got a good mix. Keep up the good work.

PS I second DJ's thought about scholars commenting right away. Give us regular folk a shot, huh? Thanks

One thought that I have had regarding the turnover of top-10 spots is the idea of not changing any QW 2 to QW 1 until all the QW 1's were depleted (that is until there are only 9 QW 1). I don't know the statistics of how many submitters there are (and so how many QW 1's there are at any given time), and so I don't know if this would unduly harm the heavy-hitting posters. If there are dozens of QW 1's to get through, it could be a long time between pairs of offerings from the top few submitters, with resulting needlessly long waits for them. But if there are only a few QW1's then it might not be onerous, especially if some turn out to be easily thumbed down.

Regarding the early solution post, I misread the time since posting. I'll have to be more careful, especially on fridays when levik sleeps in for a while.

I think your scholar idea is much of the same idea as the "long feet make you smarter" Better solvers are more likely to stay with the site and eventually become journeymen (and maybe even scholars)

I think it's hard to weed out problems, because many of the problems we see ARE good... there aren't too many that should be weeded out.

My solution to the "problem submitted" thing is just to make it so that a user can only submit one problem a day. This would disperse the problems, and also make it so that some people *cough* can't post lots of problems in one day.

I really think that the queue position needs to be more "accurate", but I don't know a solution so I can't complain too much .

Gamer,

I think that your suggestion of "more accurate" description of queue position would go a long way to remove misunderstandings. Unfortunately, given the current workings of the system (at least as I understand them), it is indeterminate where exactly your problem lies. (In other words, it sort of depends on which problems that people post and discard in the future, and can't be uniquely determined at any given time.) So an estimate would have to continue to be calculated.

As for you likening the "scholar idea" to the "long feet make you smarter".... what scholar idea are you talking about? Are you advocating that scholars should be able to freely post to problems?

--- SK

I think you are confusing comments that DJ and FatBoy made with comments that I made.

Charlie solved this problem earlier today very quickly.

I commended him on it. But DJ and FatBoy criticized his answering so quickly. I was backing up a scholar's right to answer quickly. And I agree with your analysis that scholars are better problem solvers because they have larger feet.

--- SK

Man - this is a great discussion we have here (brough back from inactivity)

Let me weigh in on some things.

1. The queue mechanism currently offers a "benefit" for early submission. This benefit is lessened by the QW mechanism, but is not eliminated. Nor should it be. It has been suggested on a few occasions that to encourage new posters, their problems should be given top priority (even to the extreme of placing a person's first problem onto the head of the queue immediately). I disagree with such suggestions for the reason that statistically, people's first problems are not the best submissions we get on the site. Once a person gets posting priviliges (by becoming a Student) they are often so eager to submit that they do not bother to dupe check their problem, or figure out if it belongs on the site at all. Those who have not had access to the queue would be surprised to see the kinds of submissions we sometimes get - from ridiculously easy arithmetics problems to riddles that totally do not work once put down on "paper".

I find that statistically (no offense to SilverKnight whose problems I'm sure are very good) the more a person has been submitting to the site, the better suited their problems are for posting. Of course exceptions to these occur at both ends of the spectrum.

2. The purpose of the site is not to see your problem on the front page. It is to serve as a place to come in order to solve interesting, brain-stimulating puzzles. I believe that the site as it is now is serving this purpose really well, to great credit of everyone who has ever submitted a problem here, as well as those who make sure that these problems are up to par.

I know that seeing your problem on the front page is more satisfying than seeing somebody else's - especially if it's an interesting and thought provoking one. However, this is not the site's primary goal, and while it may be frustrating to have to wait two to three months to have your problems show up, I believe that it's worth it to maintain the great quality of the material this site is seeing.

3. Journeymen and Scholars get to see problems before anyone else. They are always asked to not try too hard to solve a problem before it going live, since usually just seeing it is enough to determine whether if fits the site well. However, some problems are really easy, and others are just too darn interesting to hold off on.

I have full confidence, that people who have reached these ranks on the site know better than to tak e advantage of them - especially since they recognize more than anyone else that this is not a competition to show yourself in the best light. As such, if any Journeyman or Scholar posts a solution to the site shortly after a problem goes "live", I fully trust that they were able to solve the problem on their own after it having been posted. (In particular Charlie's ability to solve even the most difficult problems quickly and correctly has been well demonstrated even before he reached the Journeyman rank.)

4. This site has over 650 problems as of right now. Most of the great classics that people tell each other at parties have been submitted in the early days - and it's a pitty that more people do not go browsing the archives for them. I believe that commenting on them (be it to post your solution, or just correct someone else's) is great since it brings them into the visible list of puzzles with new comments, and thus allows newer members to experience them perhaps for the first time. Much like this thread surfaced after 3 months of inactivity, I believe re-visiting old problems to be healthy for this site (also keeps the number of dupes down, freeing the queue for more new problems to come through quickly).

Well - that looks like it, sorry for the long rant :) - And thanks for caring enough to read this whole thread :)

I was actually thinking of doing this same thing lately. Anyone see any issues with it? This would be the second time we mess with QW since its introduction... I don't reember any adverse effects from when we switched it from 3 per user to 2, but if anyone does, please speak up...

Silver Knight--

There are two conditions affecting your analysis for which you are assuming the worst case (in terms of a non-mass-submitter getting through the queue fairly quickly), while the truth is that while not at the best case positions, these conditions are much better than your assumptions.

First, those journeymen and scholars with the hacking abilities to check it out have assured us that there are only five or six über-submitters, which means that once levik reduces QW1 to one question per submitter, there are still six or seven "top ten" spots for those with less submissions. (Because of QW re-calc-ing, there are usually 12 or thirteen problems which have been exposed in the "top ten" and recieved comments and/or votes.)

Second, even most of the über-submitters did not submit all of their problems at the same time. This means that there may well come a time that their oldest problem is newer than enough other QW1 problems that it does not make the "top ten" for a few days, or possibly even for a week or more. (Granted it's not likely it will stay out of the running that long, but it does have to face the same competition that your problem submitted on the same day does.

TomM,

With all due respect, I addressed both of your points earlier.

First, I am happy to hear that there are only 5 or 6 über-submitters. But all this means is that the "system" works PROVIDED we are fortunate enough to continue to have the number of über-submitters stay below 10 (or 11 or 12 with the re-calculating). If this site should, unfortunately ;-) , get popular enough to attract a few extra über-submitters, then this issue returns. (Note that I didn't use the term problem.)

And second, the notion of an über-submitter's problems being older than a newbie's submissions remains with or without QW (as I mentioned in a previous posting).
___________________

So, again, with respect to the two conditions affecting [my] analysis for which [I am] assuming the worst case, the "worst case" (as you put it), for the first condition, WILL occur eventually, so we are just postponing dealing with it (if indeed we wish to deal with it).

And your second condition has almost nothing to do with QW.

DJ,

Those are good points.

According to your description then, even when several users gradually (as opposed to all at once) submit problems... one or two per week.... you can see that they can easily mass a backlog of 50+ problems. It might get worse as the site becomes more popular. And perhaps... that's okay.

One partial suggestion to address the point you make, is to change the "order by" from submission date to some other number (let's call it SW, submission weight). Perhaps the submission date can strongly influence the initial SW, and then, for example, if a paricular problem gets held up for one of various reasons, one can simply do a +10 (or whatever is appropriate), to slide it a little away from the front of the list, thereby allowing other problems (even by the same author) to slide down.

And perhaps the SW could be influenced by the number of other pending problems the user has, etc.

--- SK

better problems. more categories, especially verbal... probs solved. pat/sat

The link on the home page for the digest page is just below the link for the forums and is labelled "Newest Problems."

I'm all for such a page, but unfortunately I am so familiar with the site (having programmed it and all) that I think I take a lot of things for granted that should really be explained.

I suggest we strart a help thread, where peole can point out areas of the site that need explanation for either myself, or anyone knowelegable enough to answer. Then I will compile it all into some sort of an FAQ and make a separate page of it.

Come to think of it, that's a good idea. I'm off to start that thread.