Problems with the Queues

[northsidegal]

I am bumping this thread rather than creating a new one because some of the issues brought up are topical, although some are outdated.

There are currently eight ongoing normal games and only one ongoing open game. While some of this may simply be due to more demand for normal games than open games, I believe a considerable amount is due to the Open Games Queue. I originally wrote lot more here about the nature of queues and why we have them, but it's probably pointless.

I believe the main problem with the open queue is that unpopular games can grind the queue to a halt. With normal and newbie games, there's really nothing to select before you sign up for a game, with the exception of maybe the moderator (and in the case of normal games, what size of game you want to play in). And we can see that those are currently the two most popular queues. With opens, micros and mini themes, you specifically select which setups you'd like to play before signing up for them. In the case of opens, I think this leads to major problems if games start to halt the queue (and the same case could be made for mini themes, but I'm not here to make that today).


I don't mean to turn this too much into specifics or into blame, but currently in the open queue is:

• Switch (Setup), a nightstart game with a cop and a serial killer. I think this setup has a lot of problems (funnily enough, this was another thing mathdino took issue with), and although I can't know this, I expect that a lot of potential players do as well.
  
• And Hard Boiled, a setup which, as I understand it, has a breaking strategy. This setup has also been in the queue for about a month, is not close to filling, and has not been pulled.

---

What am I suggesting to solve this problem? If anyone remembers when the Micro queue had a separate "Tiny Games" slot, I think that something similar should be done for the open games queue.

Currently, the way the queue works is that mods sign up, and then there are always two games in signups at once. When one game fills or is pulled, the next mod in the queue is put into signups. I propose that the open queue have three "slots" for signups, each "slot" being something specific: one slot for "mini" open setups, games greater than 9 players, one slot for "micro" open setups, 9 or less, and one slot for experimental setups. This is similar to what already happens in the Normal queue, where there is always one mini normal and one large normal in signups at once. So, for instance, if a micro open were filled or pulled, the next game in that "slot" wouldn't necessarily be the next mod in the queue, it would specifically be the next mod with a micro open setup. (Perhaps I'm overexplaining something that everyone already understands, but I'd just like to not be misunderstood.)

I think that this would at least do work to alleviate the issue of queue blockage. In the first place, having three setups able to be in signups at once reduces the chances of all games in signups being unpopular. Second, even if they were to still fill at a slow rate, micro games take less players to fill, and as such ensuring that there are never only larger games in signups should help the queue move faster. Third, and this is less to do with queue blockage but perhaps still relates, giving a dedicated slot for experimental (or "new") setups I believe will both encourage the creation of new setups as well as incentivize more people to play open games. It's been a noted problem of the monthly open game challenges that very few of the setups are ever played, even the winners.

There are some problems with my suggestion that I know. The first I can see is that of "splintering" -- I went more into this in a longer version of this post, but basically if we had every mod in queue in signups at once, it would likely lead to players being spread out across games rather than coalesced into one or two games, which could actually lessen the rate at which games fill. Having three games in signups at once instead of two would theoretically increase splintering. That being said, I think that micro games alleviate the problem of splintering, and ensuring one is always able to be in signups should help that. The second issue is that of overlap with the Micro queue. Micro open games can already be run in either the micro queue or the open queue, and it may be said that giving them a dedicated slot in the open queue will only serve to hurt the micro queue. It is true that this change would take some games away from the micro queue. However, having the open queue have a dedicated slot for micro opens should free the micro queue up and allow Micro Normals or Micro Themes to reach signups faster. In fact, this problem already exists in some form in the fact that Micro Normals can be run in either the Micro queue or Normal queue.

[northsidegal]

i assume just because noir isn't known about, although i could be wrong

[northsidegal]

I don’t really understand why we allow micro sized games to be run outside of the micro queue. I think it backlogs those queues.

to the contrary, i find that the open queue is so backlogged that non-open micros find themselves backlogged by the micro opens that otherwise would have been in the open queue. i think a situation like that is taking place right now.

[northsidegal]

I don't mean to turn this too much into specifics or into blame, but currently in the open queue is:
Switch (Setup), a nightstart game with a cop and a serial killer. I think this setup has a lot of problems (funnily enough, this was another thing mathdino took issue with), and although I can't know this, I expect that a lot of potential players do as well.
And Hard Boiled, a setup which, as I understand it, has a breaking strategy. This setup has also been in the queue for about a month, is not close to filling, and has not been pulled.

my personal opinion is the Open Queue should just have completely different setups. A lot of these setups really just suuuuuuuck

This is something that's been noticed, and actually a lot of work has been done in a subforum to re-do the current list of "Approved Open Setups". There are two problems:

1. As far as I am aware, the work done was never actually implemented into any meaningful change. Partially I am to blame for this as one of the people who worked on re-doing the list of Approved Open Setups, as I think I've let the subject sort of drop. I'm going to go back and look over the earlier work done and try to come up with a finalized proposal for that as well.
   
2. Also as far as I am aware, the "Approved Open Setups" are really only for first-time mods or mods who don't know in advance what setup they want to use. If someone wanted to mod a specific setup that wasn't approved or their own experimental setup, they could do. This is partly why I suggested an experimental queue, to ensure that two games that might be less than balanced or too weird for the general population to appreciate don't block the queue.

Queues will always suck because you get your turn in queue based on what you want to mod, and whether it fills quickly is based on whether you want to play. It's the Electoral College of forum mafia

Could you elaborate?

[northsidegal]

tbh, I just want a queue that is exclusive to only the moderators with the highest rep. Something sorta like a VIP queue.

everyone starts somewhere. the recently finished purgatory was dannflor's second game and i think it's exceptionally done.

Experimental queue isn't an awful idea, but a Open Setup Review Group could also solve the issue of checking setups before they're run.

If you look at those contests I'm sure there are plenty of willing folks.

they're already supposed to, as i understand it. having a separate slot for experimental setups has a twofold benefit, in my eyes: it allows new setups to be "fast-tracked", solving a problem which has been noted frequently by the monthly open queue designers, and it prevents the open queue at large from being blocked by two unknown setups which players may be reticent to sign up for.

[northsidegal]

The use of Weak in the Weak Hider role seems to conflict with the current Normal definition of Weak. We should name it a Pacifist Hider with the Pacifist modifier being one that dies if it targets a role capable of killing.
That might mess up Detective investigations and may “re-break” the setup. Either way, discussion for another thread I think.

That's actually a quirk of the wiki page rather than the setup itself (in the game mathdino ran it was just referred to as "hider"), although there'd be a contradiction between the normal definition either way.

i think what pops means is if there is a demand for a certain setup, it doesn't mean that a mod will come in and fill that demand

instead you have this weird dichotomy of mods running setups that they want to run

and players don't sign up unless they find a setup they want to play

so until there is a match, you just have artificially induced clogging of the queues until the signups are pulled and new setups can enter the market until something works

Yeah, that makes a lot of sense and I tried to get at part of that when discussing how the Normal queue doesn't have this problem as compared to other queues.

I think that having three games in signups at once would at least serve to lessen this problem.

My fear would be the sort of thing where if it's too frequent the appeal is lost and players stop coming

Like if we had a weekly marathon queue I imagine it'd be less crowded, yknow.

What do you think of just a dedicated slot in the open queue rather than a specific "open queue test day"?

everyone starts somewhere. the recently finished purgatory was dannflor's second game and i think it's exceptionally done.

Correct! This queue would function as
1. A place for someone like Varsoon or someone who is already known to mod these really fantastic games that people love to play
2. A place for new mods to get some rep.

I'm not sure I entirely understand the specifics of your suggestion if new mods also moderate there.

[northsidegal]

Okay. What do you think of my suggestions regarding changing the Open queue?

[northsidegal]

Queues will always suck because you get your turn in queue based on what you want to mod, and whether it fills quickly is based on whether you want to play. It's the Electoral College of forum mafia

Could you elaborate?

I can restate myself. Tomorrow I can make an alt named Rhodesy and one named Isla-chan and claim the main account is Ras but somehow the listmod is kewl with this. Rhodesy can join the open queue to run 11v2 Mountainous, Isla-chan will join the open queue to run 4p Dethy as a forum mafia game. It will be decided they both get 2 weeks to fill. On day 11, when no one has joined either game except that Isla-chan is in for Rhodesy's game and vice versa, Jingle will post that Cultist Recruiter or any of his other popular setups will be in queue next, 7 people will post pre-/ins in the thread instead of PMing him, and they will be hushed as we keep repose and meditate on Rhodesy and Isla's game for another 3 days. The system works perfectly to serve who it is intended to serve, Rhodesy and Isla, not Jinglefornia

Okay, I believe I understand what you're saying now. I think that to fully solve that problem in every single scenario would require turning the open queue into something more like the large-theme queue, where it's not really a "queue" at all. I went a lot more into that idea before deleting a lot of it in my original post, but basically I think that retaining a queue structure is still valuable. I think that eliminating it entirely increases "splintering" too much and is very disincentivizing towards new moderators as well as untested setups.

Okay. What do you think of my suggestions regarding changing the Open queue?

I think having three open slots could theoretically alleviate some issues with the Open Queue because what it does is it results in more "shots" at creating a game that meets demand.

But also the Open Queue is the least popular queue and I don't think it has a consistent enough playerbase to even generate consistent demand so fracturing might just be the death blow to the queue.

Having three signups doesn't necessitate a removal of stagnation in the queue, you have three games unable to meet demand, they sit there and block the queue. I guess on the positive side is you cycle through setups faster.

Basically the point I'm trying to make is that to implement your change is a massive risk by the list mods and administrators who I assume don't want to just kill the open queue while its still barely breathing in favor of a change that may or may not improve it.

The open queue wasn't always the least popular queue. Once upon a time it was somewhat thriving and had a fairly consistent group of players who played a lot of opens. It doesn't

have

to be the case that it's the least popular, and the fact that it is now is, I think, a product of the queue rather than the cause of the queue being stagnant. Basically, I think you have your causation backwards: you think that opens are unpopular and so the queue is slow, and I think that the queue is slow and so opens are unpopular.

It's true that having three signups doesn't necessitate a removal of stagnation. That being said, I think that my specific proposals all do work to help eliminate stagnation. Micro games are inherently anti-stagnant with their lower player size: we can even see that

right now

there are two micro opens in the Micro queue that could theoretically be in the open queue. In fact, you're the one modding one of them! It's thus a little surprising to see you say that there isn't a consistent enough playerbase to generate demand. Experimental setups have an inherent draw to them, I think, by the nature of the monthly setup challenges as well as people who work together on theorycrafting a setup. I think the recent popularity of Undertale Semi-open speaks to this, although that's obviously only one example.

I don't mean to be accusatory towards T-Bone, but I think that the decision to make the queue "self-automated" has done far more to kill the open queue than this change could possibly do. I think the impact of that decision is still being felt today, actually.

[northsidegal]

1 would be the setup of the month. Something that players can reliably play, similar to how the Newbie queue functions.

This has been proposed before. I don't think that there are enough games played in a singular month to really justify it, and games that do attain a certain level of popularity where players just keep wanting to play the same setup just get put into the queue over and over without any need for a dedicated slot for them (the time that Coalition was ran a few times in a row comes to mind). With that in mind, it seems like this is just asking for an unpopular setup to make it into the monthly rotation. I think that a dedicated experimental slot instead would do more to increase open game popularity.

i agree open queue was thriving once

its just i don't even check open queue thread unless to remove the red marker thing in ego posts anymore whereas once i did actively monitor the queue

i assume im not a special

thus i assume that if open queue continues to stagnate less people will check the open queue thread

I'm not quite sure what you're arguing here. The existence of some kind of negative feedback loop of popularity? I think we can both agree that the open queue, as it is right now, is stagnant. You say yourself that if it continues to stagnate that even less people will check the thread. I believe that it is likely that if things continue as they are now, the open queue will not gain popularity. As I understand it, you seem to disagree, and think that a change would be more of a risk of stagnation than keeping things the same?

[northsidegal]

It's not intuitive to me that a demand driven queue would "splinter" more. Could you explain why you'd expect that result?

I don't consider it an eventuality, but it seems more likely to me given that kind of structure.

Let's say that the normal queue went like the large theme queue, and you could sign up for any mod in line's game. Certainly we would expect a similar behavior as things already are to remain, where people simply sign up for whichever game is the closest to filling, or whoever's game is at the top of whatever list there may be of games. That being said, we might also expect some players to sign up for only mods that they like: for example, if I really like schadd's normal game design, I might only sign up for his games, even if other games are closer to filling. Of course, we might expect this behavior as things are now, where I would only sign up for a normal game if it were modded by schadd. However, I believe there's a subset of people who would only sign up for schadd's games if they could that otherwise (as in, right now) sign up for whatever normal games are in queue. A similar behavior might exist for certain playercount games, not just moderators.

And thus I would expect the total rate at which games fill to go down, even if only slightly. Of course, this analysis is for normal games, for which setup information doesn't really exist before you sign up.

[northsidegal]

This doesn't come up as much if you let people "/in - unless one of my other /ins in this category fires, then /out". That seems like a logical step of implementing a free market queue. If the mere presence of schadd's game makes people not even make that level of commitment to both schadd's game and another in queue, it's not really the structure of the queue itself but that schadd's spot in the queue is functioning too much like schadd tweeting "feelin like moddin another next week, don't commit to anything you guys" and I'm a hard sell on anything that's supposed to be better than something else by depriving users of information just because lack of information creates a desired side effect.

I'm not quite sure we're understanding each other. At the very least, I'm not sure I'm understanding you. The scenario I mentioned doesn't consider someone who would say "/in unless the other game fires first, then /out". It considers someone who, if given the choice, would only sign up for schadd's games, but otherwise just signs up for whichever normal game is first in queue.

I don't really understand your tweet analogy at all.

I don't think that a queue structure deprives users of information at all. You know who's coming up to mod things in the Normal queue, it's always listed in the first post. I agree that even I feel something unintuitive about arguing for less choices at once. It seems to run contrary to what I would argue in a more real life scenario. I think that the goals that I would be trying to achieve are different though. If I were to make a market analogy, in mafiascum I care about being fair to newer moderators and untested setups, whereas in real life I might say that if people only want to buy their favorite products (as opposed to "newer or untested products") then they should absolutely be able to.

[northsidegal]

The market analogy also breaks down when you consider that games of mafia are inherently social, community-based experiences in a way that buying a product isn't. In mafia, if 5 people only want to play micro A given the choice and 4 only want to play micro B, neither will ever play a game. In real life, if 5 people want to buy product A and 4 want to buy product B, their lack of coordination isn't

preventing

them from buying the product.

I think that some kind of queue structure is actually really important because on some level it forces people together when, if they were given more options, they would be apart. That seems fairly important for keeping a community alive to me.

[northsidegal]

I was trying really hard to state my extreme position and bounce, because I am 100% that person that thinks new mods should get brickwalled out of modding indefinitely because games are constantly filling with reputable mods because I think of the players as the customer. The current system treats the mods as the customer. Isla-chan got her name for a reason. I know Isla-chan wants to mod her first game but I don't want 9 players to pay any amount of setup preference to give her that.

-shrug-

I would probably agree with you in different contexts, but in the context of mafiascum as a community that we want to keep alive and see grow with the influx of new players, I don't think this is a good position to hold. New members need to be given a chance so that eventually

they

can become the old members, the reputable mods. Mafiascum isn't a service, there aren't customers. It's a community.

[northsidegal]

i am remiss at the fact that i have not yet worked choice overload ("the paradox of choice") into any of these posts yet about eliminating queues so i'm just gonna awkwardly bring it up here

[northsidegal]

That's fairly compelling, actually.

[northsidegal]

That being said, that seems more like an argument for either stricter punishments for moderator errors, or stricter oversight for newer moderators (or both, any of which I could easily be convinced on depending upon the specifics) rather than having to do with queue reform. Would you agree?

[northsidegal]

Currently 0 ongoing open games with the conclusion of Open 776.

[northsidegal]

@TBone-

I can't help but feel that, while changing the list of Approved Open setups should help the open queue somewhat, it doesn't really do much to stop the issue of queue blockage. I also think that it just leaves some other issues floating, like that of a lack of ability to test new open setups in an expedient manner.

I don't understand your reasoning for not wanting to increase to three games in signups at once, but I can imagine some reasons myself. That being said, what about at the very least changing to be like the Normal queue, with one slot for Micros and one slot for Minis / Larger games? This should immediately increase the amount of games being played in the open queue -- we can see at this very moment how many open games are being run in the Micro queue when they could potentially be in the Open queue. You may consider it a mere trick of accounting, shifting games from the Micro queue to the Open queue and not really increasing anything at all, but by having Micro Opens be run in the Open Queue that would otherwise be in the Micro Queue, other Micro games will reach signups faster. (The same could actually

also

be said about the Mini-Theme queue: it's only just occurred to me, but Undertale Semi-Open is being run in the Mini Theme queue rather than the Open queue. It's a heavily themed game, yes, but still an open one.)

Either in addition to or instead of that change, I would still suggest some sort of expedited way to run experimental setups -- potentially it could be a monthly or bi-monthly "event", if you are against a permanent slot. I believe we have been "stuck"--as you say--to test new open setups primarily because of long mod queue times. I don't think it's an overstatement to say that the average waiting time between signing up to mod in the open queue and having your game just enter signups is an entire month or more. We see that Large Theme Games that aren't even really Themes but are just Large Opens get run in the Large Theme Queue all the time -- I would hypothesize because of the queue times. I know that personally that's why I ran Baton Pass in that queue, and I expect that that's why a game like Gameshow Mafia might have been run there as well (this might just be due to moderator preference or because Large Themes are perceived as more of "big events" than just large open games, but I digress). I feel that I should also bring up the Marathon Weekend Open Setup group, who were able to blaze through a ton of prospective setups given the fast nature of marathon games.

The point I am trying to make here is that, the way I view this situation, one cause (a slow or blocked game queue) is having multiple negative effects: a

literal

lack of ongoing open games, a lack of ability to test new open setups, and a diaspora of setups that otherwise would have been run in the open queue.

[northsidegal]

Also, if I might make a suggestion: I think that you should still offer first time mods a random suggestion of three setups, but just make it clear that they're free to run any approved open setup that they want. Like I said, choice overload and all that. One thing that I think pre-designed normal setups has definitely proven is that some people just want a setup to be chosen for them, and are primarily focused on modding it rather than coming up with it. It's not difficult to just pick a random one from an Approved Setups list, sure, but having those immediate three options there is, I think, a benefit.

[northsidegal]

I'm going to copy paste my posts from when we went through the current Approved Open setups back in late August:

Spoiler:

okay, there wasn't really that much to add from what i could see. if anything the issue would just be some setups not having wiki pages

anyways, a few micros that i want to put forward for definite consideration:

The Coalition
8-Ball

Variable Setups:

Half Mast Nightless
Pop Goes the Nightkill

i could see the argument of not enough data for the last two, but i think the setups are pretty well designed and have had decent enough test runs that it's not a huge issue greenlighting them.

Vengecop is another one that'd i'd like to put forward that's variable, although I think it works best at 13 players.

I'm just gonna go through most of the setups listed in the "approved" category and give some thoughts on them.


The One I Cannot Kill — I don't think this one has actually been approved? I think it's probably in this category by mistake, although I do think it's a pretty good setup. Two scum wins out of the two runs I'm aware of, but I'm not sure that's entirely representative. Room for testing it further given the marathon weekend coming up.

Vengeful Mafia — A classic. More of a marathon-type setup than is often played, but I don't think there's any major problems that would warrant removing it?

Lovers Mafia — Largely the same as above. About 65% scum wins to town wins, but again – mainly a marathon-type setup. I think that this shares the problem of many of the lovers-type setups in that literal random voting has shown to be as good as if not more effective than actually scumhunting.

Carbon-14 — A classic and sitting at 48% town winrate going off of the wiki history. I'm not making any comments on the setup itself, but at least based on the winrate it seems fine.

No Lynching Town — Don't think this setup has been run for years, making it a decent candidate for getting cut in my opinion. If we want the open queue to improve from its current pace, the setups being run should be modern ones that people actually want to play and (while not always) have some kind of experience with.

Nomination Mafia — Honestly don't have many thoughts. It's been played once recently? Doesn't seem

terribly

unbalanced, looking at the history?

Kids With Guns — Largely the same as above. Been played once recently, no obvious winrate problems and in this whole diatribe I'm trying to keep somewhat away from analyzing the specifics of the setup itself. I don't think this setup has ever been played in the open queue.

Less Pressure — Another one that hasn't been played in years and that I've never heard of.

Scumhunter's Speed 8p — Fairly bad town winrate.

True Love — Another that seems to have not been played in years. I also don't think very highly of most of the "lovers" setups.

09:12 — Played somewhat often recently, near as I can tell it's balanced enough. Good setup.

2d3 — The old newbie setup. If it's good enough for the newbie setup it should be good enough to be approved, but do we want it in the rotation? That's a genuine question, not a rhetorical one – I really don't know.

3d3 — 2d3's prototype. I don't think this has ever been played and I think it's not actually approved. Don't think it should be.

Alternating 9P — Fairly bad town winrate, but played recently enough.

Backup6 — Modification of Matrix6, played once. Really, when it comes to a lot of these setups, it's going to come down to what philosophy you have as to the approved open setups list. I don't think there's much

wrong

with this setup, but there's still the question of whether or not it's worth including.

Chosen Mafia — Fairly popular, with a pretty balanced winrate. I also personally think it's well-designed.

CultD3 — Not an approved setup.

Desperation Day — I think it's fairly well designed, although it hasn't been played much at all, recently or in total.

Double Day Unlimited — Has been run recently, but probably has winrate issues given two town wins out of seven runs.

Even-Odd Killers — Another one that I'm pretty sure is not approved.

Forest Fire — Great setup. Played fairly often recently and fairly balanced.

Grey Flag Nightless — Winrate issues. Town won for the first time fairly recently, not sure if that's evidence enough to point towards anything.

JK9 — I don't think anyone has played this in a while. I would say it's also a pretty swingy setup that only makes the 2d3 problem of early scum deaths being a death sentence even worse.

Matrix6 — Old newbie setup, pretty much same thing to be said here as with 2d3.

Purgatory — Played fairly recently, town has won all three games. Not much of an opinion here without looking more closely at the setup itself.

The Coalition — There by mistake, not actually approved (although I think it should be).

Twin Trap — Pretty balanced winrate history, a few games played recently. Not a bad one to keep.

Overall, I think the following setups should be removed for sure:

• No Lynching Town
  
• Less Pressure
  
• True Love
  
• 3d3
  
• CultD3
  
• Even-Odd Killers
  
• JK9

These are setups that I think have reason against their inclusion, but I'm less sure about:

• Scumhunter's Speed 8p (winrate issues)
  
• 2d3 (for the sake of keeping interesting setups in the queue)
  
• Backup6 (same as above)
  
• Double Day Unlimited (winrate issues)
  
• Grey Flag Nightless (winrate issues)
  
• Matrix6 (interesting setups)

Removing all of those setups would, if I counted correctly, still leave us with 12 micro setups before even adding any new ones, which I think is a decent size.

The amount of times that I said "this isn't actually an approved setup" should also point towards reformatting the way the wiki is structured being a good idea. Perhaps having a category

specifically

for approved open setups would be a good idea – it's easy to just place newly created micro setups into the "micro open" category without realizing that that category is meant to be only for approved setups.

For the sake of time I'm gonna be a lot more concise on the mini setups and focus on problem setups.

Faith Plus One should be out, if it's still even in. Hope Plus One as far as I'm concerned is a superior version.
Hard Boiled has a breaking strategy and should be replaced by Noir, designed specifically to correct that breaking strategy.
Another Mathdino original, Polygamist should probably just go.

Enemy of My Enemy has only been played three times, with the last time being years ago.
Masons and Monks has

never

been won by town. This was reworked a bit in that same thread that I linked to for polygamist, but in it's current state it should go.

Switch (Setup) should go, and Sharing is Caring is another one I'd like to put forward for consideration.

If you don't care to read all that, it would leave the Micro Approved list as follows:

Spoiler:

Returning Micros:

Vengeful Mafia
Lovers Mafia
Carbon-14
Nomination Mafia
Kids With Guns
09:12
Alternating 9P
Chosen Mafia
Desperation Day
Forest Fire
Purgatory
Twin Trap

New Additions:

The Coalition
8-Ball
Half Mast Nightless
Pop Goes the Nightkill

Keep in mind that this was compiled in late August last year, so it's probably missing some new setups.

[northsidegal]

so standard to fill games is 2 weeks


what if we make it so you have to hit breakpoints? like it must be at half capacity at one week or its pulled, that way we can "cycle" through the queue faster? Also if mods are looking for experience, maybe we create some way for new mods to find out whats in demand? Like a game request thread/discord?

I'm going to quote Something_Smart on that, he had a pretty good analysis on the subject in a similarly themed thread:

Spoiler:

*Dongempire Friends & Enemies & Enemies 2/11 - PULLED 3/1 (5p in 18d)
ChibiBear Donner Party 2/11 - 2/23 (12p in 12d)
*ejjinami Nomination 2/26 - 3/2 (11p in 4d)
brassherald Fire and Ice 3/6 - 3/14 (13p in 8d)
Kaiveran Two-Fold C9++ 3/16 - 4/3 (16p in 18d)
mcqueen Hope Plus One 3/16 - 4/10 (13p in 25d)
Jingle Panic Room 4/7 - 4/18 (11p in 11d)
*Clemency Duck Duck Goose 4/8 - PULLED 6/3 (11p in 56d)
schadd_ House Party 4/25 - 5/18 (14p in 23d)
BuJaber Wayward Bullets 6/2 - in signups, unlikely to fill (3p in 18d)
*NotAJumbleOfNumbers Grey Flag Nightless 6/3 - in signups, likely to fill (8p in 17d)

(* = first-time mod)

Assuming Wayward Bullets is pulled and Grey Flag Nightless fills:

Signups per day, FILLED GAMES:
*0.47, 0.52, 0.61, 0.89, 1.00, 1.00, 1.63, *2.75

Signups per day, PULLED GAMES:
0.17, *0.20, *0.28

Proportion of first-time mod games that filled: 2/4
Proportion of non-first-time mod games that filled: 6/7

This is an analysis of only games in the new queue thread. It might be worth it to extend this a while back into the previous thread to get a higher sample size, but I do think this is helpful.

Ideally, what we want is a way to recognize a game that will get pulled without waiting three weeks of it clogging up the queue.

We might be able to say something like "if a game doesn't have at least 4 signups after 10 days, it gets pulled immediately."

Another thing that might help, according to these data, is to revisit how first-time mods work. In my opinion, the setups Dongempire and Clemency were trying to run are both bad, and I'm not surprised they didn't fill. Usually you need either a high-profile mod or an interesting setup to get fast signups, and it's not surprising that the two games with the lowest fill ratios (mcqueen and NotAJumbleOfNumbers) are a first-time mod and someone who hasn't modded in years. Maybe this means first-time mods shouldn't be allowed in the Open Queue at all, or maybe the list of approved setups needs to be shortened to only ones that have been run successfully many times (Friends and Enemies, Fire and Ice, Stack the Deck, C9++, etc.).

My final idea is related to that. We want mods to be running setups that players want to play in, so maybe there should be more player involvement in what setups get run. Could be as extreme as an open game where the players vote for what setup to play after the game fills, or it could just be a monthly poll of what setups players want to play, to be posted for anyone considering modding an open.

Also I notice that nearly a full year ago he also noticed that it's faster to just run experimental opens in other queues:

Experimental Open setups can just be run in the Mini Theme queue, to be honest. That queue usually moves way quicker, and open setups sometimes get run in there anyway.

I mean, this really shouldn't be happening, right? It's just not efficient, right?

[northsidegal]

I do notice that there is some element of hypocrisy in my argument. I'm claiming that it's inefficient for opens that could be run in the open queue to be run elsewhere, and if we would only make it easier for them to be in the open queue then the total amount of games being played would increase. But this in some sense contradicts my earlier point that queues are valuable because they prevent splintering.

Let's do a thought experiment. Let's say that we got rid of the Mirco queue entirely. All of the games that would've otherwise been in the micro queue could still be run, they'd just be run in different queues: Micro Normals in the Normal queue, Micro Opens in the open, so on. The opposite of doing this would be introducing

more

queues--if, say, we split the Micro Queue into two separate queues, say the Micro General queue and the Micro Theme queue. Theoretically, one could argue for this using the same arguments that I'm using now: that having a Micro Theme queue would increase total games played because it'll free up games that would otherwise be in the Micro General queue, and so on. At the extreme end of this, we can imagine a separate queue for every game that would be run, whereupon we find that we don't really have any "queue" at all, we've actually gone to the "entirely separate signups" model. The same arguments that I'm using now with regards to efficiency and how Opens are blocking other queues, brought to an extreme conclusion, leads to exactly what I argued against before--a total lack of queues!

My only thought in response to this is probably just "meh, there's really no need to bring it to an extreme conclusion, there's probably some optimal middle ground".

anyways, this is all totally pointless musings and none of it matters at all i think with regards to the current discussion.

[northsidegal]

I think the Mini Theme queue is hit particularly hard by the moderator/player disparity, a problem that normals don't have at all and a problem that is fairly limited when it comes to opens. With Opens, if you're modding a pre-existing setup, there's almost certainly some subset of people who will be interested in playing: it may be a highly specific audience, as with a multiball setup, but

most

existing setups should have people willing to play. With a Mini Theme, your setup idea could potentially be a total miss with prospective players, leaving you "dead in the water" in the queue, I guess.

(this is somewhat unrelated to the open queue, although i guess the thread

is

"problems with the queues" and not "open queue reform")

[northsidegal]

The Marathon Weekend Open Setup group already exists as a method to test new open setups during marathon weekends. As far as just having more marathon weekends I'm not going to comment, but as for whether or not it would serve to increase the amount that experimental setups are tested, I don't think that making it "official" would do anything that isn't already really being done.

[northsidegal]

As far as long term health of the site I have little to say. I'm certain that it's already apparent from my posts, but really my primary concern here is the Open queue. I will say that I think, regardless of the general matter of overall site health, the disparity in games being played between the Normal queue and the Open queue is indicative of some specific issue with the Open queue rather than just a symptom of a dying site.

Actually, I think I have little else to say on anything until T-Bone or other site administrators respond.

[northsidegal]

You’re correct.

what do you think about all this penguin

[northsidegal]

Is a more reliable model for ensuring that marathon-speed games fire to plan them well in advance, or to assemble a playerlist directly at the time of playing, as with marathons and MU turbos? The latter seems unsuited to a "queue" model as it would likely be too quick to reliably be able to be overseen by the listmod, but it also seems to sidestep some issues that I can see with games planned well in advance, such as people forgetting that a game is taking place. If that were to happen then it would likely require the game mod to scramble to find a replacement player anyways, in some sense imitating the marathon/turbo model.

I think that a queue would help make it so that 9 players

could

make time in their schedule to sign up for and play that game, but I also think that it would run into some issues. That being said, maybe the fact that replacements would be necessary and that issues would arise doesn't mean that the idea itself is an untenable one.