Punish bad lobbies

@cocucka said in Punish bad lobbies:

@thomashiatt , @ftxcommando , explain me please how can I play ladder at desired position without full share on dual gap map.

Sounds like you’re part of the problem then. Can’t moan about wanting a specific slot and then moan that host wants to play with a specific player.

Likewise can’t have these 80 prerequisites for you to play and then be flabbergasted host takes a while to balance games.

Consider there should be interface like this:

  • Select map: random, small pool, medium pool, huge pool, <check manually>.
  • Select size: random, 1v1, 2v2, 4v4, 6v6.
  • Select role: random, mid, air, navy, eco, non-mid.
  • Toggle ranked: on, off.
  • Button: Enqueue
  • Hint: Most wanted roles (faster search): mid.
  • Hint: Most wanted sizes (faster search): 4v4, 6v6.

Algorithm of matchmaking:

  1. Search for available players for a while.
  2. If not enough players prompt to extend role/size/map (without loosing priority).
  3. If not enough players prompt to extend role/size/map(without loosing priority).
  4. If not enough players prompt to extend role/size/map(without loosing priority).

@ftxcommando said in Punish bad lobbies:

@cocucka said in Punish bad lobbies:

@thomashiatt , @ftxcommando , explain me please how can I play ladder at desired position without full share on dual gap map.

Sounds like you’re part of the problem then. Can’t moan about wanting a specific slot and then moan that host wants to play with a specific player.

Likewise can’t have these 80 prerequisites for you to play and then be flabbergasted host takes a while to balance games.

If host really wants balance then he should use opti. Otherwise he should respect the role. At least, he should react somehow in chat about this.
Also when the highest rank writes something like "move me to 5" host moves but when the same is written by low skill players host kicks them.

@cocucka said in Punish bad lobbies:

Consider there should be interface like this:

  • Select map: random, small pool, medium pool, huge pool, <check manually>.
  • Select size: random, 1v1, 2v2, 4v4, 6v6.
  • Select role: random, mid, air, navy, eco, non-mid.
  • Toggle ranked: on, off.
  • Button: Enqueue
  • Hint: Most wanted roles (faster search): mid.
  • Hint: Most wanted sizes (faster search): 4v4, 6v6.

Algorithm of matchmaking:

  1. Search for available players for a while.
  2. If not enough players prompt to extend role/size/map (without loosing priority).
  3. If not enough players prompt to extend role/size/map(without loosing priority).
  4. If not enough players prompt to extend role/size/map(without loosing priority).

Too many similar options.
allowing so many choices will cause players to fragment and not be able to match until enough time passes. (If you aren't allowing hard vetos)
but to go into the specific options:
@cocucka said in Punish bad lobbies:

Select map: random, small pool, medium pool, huge pool, <check manually>.

A map veto system has been talked about and might come if someone finds the time to implement it. allowing such big choices here will make it hard for people to match up and their ratings will deform to that type of map requiring seperate ratings to accurately match players.

@cocucka said in Punish bad lobbies:

Select size: random, 1v1, 2v2, 4v4, 6v6.

This is already there (except 6v6 for stated reason)

@cocucka said in Punish bad lobbies:

Select role: random, mid, air, navy, eco, non-mid.

This only makes sense for dual gap. Other maps have different slots and most maps that aren't DG or Setons aren't played enough to have such hard roles. (most maps have an air slot and maybe navy if there's water)
also hard to define for mapgen maps.

@cocucka said in Punish bad lobbies:

Toggle ranked: on, off.

This option makes no sense. Unlike most online games FAF has no hidden rating that it uses to match you vs others. The rating you see is the rating you are matched at. unrating a match makes sense if you play a severly unbalanced match or try something which could otherwise fall under rating manipulation. If you play mostly unrated against random opponents none of the players involved players will have fun.

You overestimated the amount of players FAF has or underestimate the amount of players needed to allow for such options without deteriorating queue times massively.

Nothing like this would ever exist for the matchmaker. The whole purpose is making it easy to play not spending 45 minutes going through a hospital entry form in order to just launch your weekend game.

Singular map queues are disliked conceptually because it isn’t what is viewed as the sort of game experience FAF should be endorsing. Why not just have a wonder queue (6 variants of it), dual gap queue (3 variants of it), astro queue (7 variants of it), senton queue, map gen queue for every size of game, so on.

If you ONLY want to play one map, that’s the whole point of custom lobbies. You choose it when it is literally a precondition for you to want to play. Same reason you have this whole “I only want to play air slot” or whatever.

If you dislike other people having the right to do as they please in the lobby they host, then host your own. There is no prerogative on the host and he can do whatever he wants so long as he is willing to spend forever waiting for people to play the game he wants.

All welcome host with 7 0s and 1 1500? He can

(1) launch
(2) kick a 0
(3) kick the 1500

all have drawbacks, up to host to choose which tradeoffs he prefers.

@ftxcommando said in Punish bad lobbies:

If you ONLY want to play one map, that’s the whole point of custom lobbies

No.
The point of custom lobbies is to fill whatever gap the matchmaking system can't fill.
This is almost always playing with specific people and then everything your matchmaker can't do.
But you want the matchmaker to be able to handle as many scenarios as possible, because that makes it easier for new players to join in on the fun.

@ftxcommando said in Punish bad lobbies:

Singular map queues are disliked conceptually because it isn’t what is viewed as the sort of game experience FAF should be endorsing.

But that's conceptually bad, because the matchmaker mustn't be an iron hard competitive environment. It can be casual as well, so there is no need to exclude modes because they are not deemed competitive enough. If lots of people want to play it, there should be a queue.

@ftxcommando said in Punish bad lobbies:

Why not just have a wonder queue (6 variants of it), dual gap queue (3 variants of it), astro queue (7 variants of it), senton queue, map gen queue for every size of game, so on.

That's why I also mentioned you don't want mutliple queues that are too similar, so you just pick one version of each important map etc.
That way you also reduce the same fragmentation in custom games, where now people host 5 different versions of the same map, but by selecting one for the matchmaking queue you lower the entry barrier to play that specific version, reducing the incentive to host a different version and compete for players.

@nex said in Punish bad lobbies:

The point of custom lobbies is to fill whatever gap the matchmaking system can't fill.

Correct. And rather than having 50,000 matchmaker queues for the 50,000 individual maps, we can delegate it out for the free market of players to handle and instead maintain the queues for the ways FAF softly encourages itself to be played. This is already seen with how balance decisions operate and maps are just as large a role in deciding the utility of units as statlines.

This is almost always playing with specific people and then everything your matchmaker can't do.

You can party queue in the matchmaker.

But you want the matchmaker to be able to handle as many scenarios as possible,

No.

because that makes it easier for new players to join in on the fun.

Then we would remove all the matchmakers as they currently stand and simply have a pve queue that leads to astro and culminates in dg.

But that's conceptually bad, because the matchmaker mustn't be an iron hard competitive environment.

Nobody said you need to try your ass off, but the game is ultimately balanced around some overton window of acceptably deemed situations. These are the matchmakers.

That's why I also mentioned you don't want mutliple queues that are too similar, so you just pick one version of each important map etc.

These are all big, simple teamgames. They are all too similar.

That way you also reduce the same fragmentation in custom games, where now people host 5 different versions of the same map, but by selecting one for the matchmaking queue you lower the entry barrier to play that specific version, reducing the incentive to host a different version and compete for players.

This already happens organically.

@ftxcommando said in Punish bad lobbies:

rather than having 50,000 matchmaker queues for the 50,000 individual maps

Please provide some statistics if you have access.

I can see that the most popular maps are: seton, astro, gap. Maybe generated. Maybe other curated huge maps. But there should be at least one queue for these maps.

Should be based on what rationale? By popular sovereignty half of all queues wouldn’t even involve players facing players.

Ultimately these queues have never been about some popularity poll. The first queue in GPG was 1v1, this carried out into FAF, now it’s different. Back when it was only 1v1, this was basically the entire rationale for balance decisions with some sentons sprinkled in. Now the schema of games viewed this way has gone beyond 1v1 and you have a roster of teamgames that have legitimacy for considerations about the game.

1v1 wasn’t chosen because it was popular, never was. It was easy dev resource wise, it had historical precedent, and it was built upon using said precedent.

@ftxcommando said in Punish bad lobbies:

just play matchmaker

Sad past-10PM EST player noises

@ftxcommando said in Punish bad lobbies:

Should be based on what rationale? By popular sovereignty half of all queues wouldn’t even involve players facing players.

Ultimately these queues have never been about some popularity poll. The first queue in GPG was 1v1, this carried out into FAF, now it’s different. Back when it was only 1v1, this was basically the entire rationale for balance decisions with some sentons sprinkled in. Now the schema of games viewed this way has gone beyond 1v1 and you have a roster of teamgames that have legitimacy for considerations about the game.

1v1 wasn’t chosen because it was popular, never was. It was easy dev resource wise, it had historical precedent, and it was built upon using said precedent.

Why do you speak about history? This is not about history, this is about actual problem.

I do not understand your point. You say lobbies are too bad to play them, matchmaker is for strange different (then play game and get fun??) purpose. But currently faf is lobby simulator with rare gameplay inclusions. So what do you suggest to solve this issue?

I don’t see an actual problem. I see you wanting to have the powers of the dude that hosted a lobby but only to enforce the things you want. If you want that, host it yourself. Otherwise, lower your list of criteria in terms of slots, balance, maps, etc and if you lower it enough, you can use a matchmaker for the process to be automated.

Host by definition had to spend the most lobby sim time, if he doesn’t want to play with somebody or play in some sort of way and increase said lobby sim time, it’s his prerogative.

You don’t get to choose the map. The rating limit. What slots are open. If it’s optimal balance. Why would you suddenly get a choice over who he kicks or anything else?

Has anyone even thought about why there exist five different gap lobbies at the same time? Why did four people host another lobby, enduring the maximum amount of lobby sim possible, instead of joining the existing lobby? What makes you confident that a matchmaker queue would reduce this fragmentation? Would these people even play in a gap matchmaker?

Honestly, wouldn't be even surprised if after getting gap Q they would come to bitch about FAF balance team not working on the balance for TMM gap, player balance in game, TMM using wrong map version, or the fact that they can't pick they starting position lmao.

@ftxcommando said in Punish bad lobbies:

This is almost always playing with specific people and then everything your matchmaker can't do.

You can party queue in the matchmaker.

Yes, but if you want to play against specific people. (There are people that don't enjoy competing with stranger, but like to compete with their friends)

@ftxcommando said in Punish bad lobbies:

But you want the matchmaker to be able to handle as many scenarios as possible,

No.

Why wouldn't you want that?
Or rather why would FAF be the literally only online game that doesn't want people to primarily use the matchmaker?
Every other modern game has a quickplay button and only a negligible amount of custom games

@ftxcommando said in Punish bad lobbies:

Then we would remove all the matchmakers as they currently stand and simply have a pve queue that leads to astro and culminates in dg.

If the PvE players would like to play with random people and wouldn't care which of the coop maps they play, a PvE queue might also make sense. (Maybe survival could work as a queue🤔 )
And I don't see how this would lead to the removal of the currently existing queues.

@ftxcommando said in Punish bad lobbies:

Nobody said you need to try your ass off, but the game is ultimately balanced around some overton window of acceptably deemed situations. These are the matchmakers.

It's not about needing to try your best. It's about the complexity of the game which is just lower in an astro game. Sure the game isn't balanced around astro, but it doesn't need to be balanced for every possible queue.
You could then also add specific ratings for the astro/DG queues and just hide those ratings from the players, to make the feel even more casual.

@ftxcommando said in Punish bad lobbies:

These are all big, simple teamgames. They are all too similar.

How is Astro 4v4/3v3 too similar to dual gap and those two compared to setons?
I don't really see many wonder hosts, so I don't think it plays a role in this, but a Setons queue could also work.

@ftxcommando said in Punish bad lobbies:

This already happens organically.

How is anything here happening organically?
There are always multiple different lobbys of the same or almost the same map hosted.
The problem exists and a queue might solve it.

@blackyps said in Punish bad lobbies:

Has anyone even thought about why there exist five different gap lobbies at the same time? Why did four people host another lobby, enduring the maximum amount of lobby sim possible, instead of joining the existing lobby?

I don't know why people do that. Maybe because every second advice is host your own custom game if you don't want to get kicked?
Also they get hosted for different rating ranges, but that could also be solved more efficiently by a matchmaker.

@blackyps said in Punish bad lobbies:

What makes you confident that a matchmaker queue would reduce this fragmentation? Would these people even play in a gap matchmaker?

Would they play a matchmaker queue? I don't know and probably nobody will ever know until such a queue is tested. But I'd think that a queue attracts people more than custom game lobby sim, since it requires less effort. They don't have to fear getting kicked, they don't need to check if the settings are correct or change in between you joining and the game starting. You don't need to fear getting the spot you want, just for it to be taken away a minute later.

@endranii said in Punish bad lobbies:

Honestly, wouldn't be even surprised if after getting gap Q they would come to bitch about FAF balance team not working on the balance for TMM gap, player balance in game, TMM using wrong map version, or the fact that they can't pick they starting position lmao

You'd need to use the most popular version and at least most dual gap (all of the currently hosted games) use the same version and the queue would then naturally decide what the correct version is once you get people to start using it. Since the "correct" version is just what people are used to playing.

To get around people complaining about their queue not being considered in balance talks, you could split the queues between a "competitive" and a "casual" section, to clarify the difference.

For starting positions: If the queue only hosts one map, allowing for preferences in starting positions is actually doable and more reliable than hoping the host of the custom game finds the best solution to all players preferences.

@Nex agree. Moreover this is the simplest solution could be - just add more queues like other queues already exist nearby.

No the simplest solution is to start hosting lobbies yourself, not sure why that’s constantly ignored when other people in this thread bring it up.

I generally play setons and I would be against a setons queue and wouldn’t use it unless it killed custom lobbies and I had no choice, and if it did it would be a downgrade. This is how I’ve felt about that for a long time now and I’ve said as much before.

Sometimes people are reluctant to host themselves because it seems like the slowest option to get a game. @Cocucka all I can suggest is make your lobbies fun and fair and hopefully people will flock to them. It's slow at first but it gets better as you get better at hosting. Once enough people add you to their friends list, they get notified when you make a game and it fills a lot faster, even compared to matchmaking at prime time.

This whole thread in painful.

@exselsior the OP is indeed trying to host themselves already, I've seen their lobbies in the list.

Maybe the thread is partly about that solution not being entirely frustration free, and that might have something to do with people's reluctance to join lobbies hosted by lower rated players, but I'm just guessing.

Personally, I have to admit whatever gut feelings I might have about a hypothetical gap only queue are worthless since the death of 4v4 tmm came as quite a surprise to me. It was vindicating, but surprising as hell, based on what I perceived the consensus to be at my rating level at the time. The fact that it even existed at all is proof that fullshare was at the very least controversial, forcing the devs to launch tmm with both, which was the right call.

I wonder if no-share's utter failure came as a surprise to any them as well, and if so, maybe it's worth reminding ourselves our intuition is not perfect at predicting how large, mostly silent communities will react and can't replace actual experiments.

On the other hand, if it came as no surprise at all to anyone on the dev team, more the reason to trust their judgement on this, but it's interesting to think that they were willing to entertain no-share then just to prove to people hardly anyone wants to play it, so it must not have been that huge of an effort.