104 replies to this topic

[Alan Hicks]

I hope they delete those, or at least reduce them to three people or two...whatever, the match-maker still sucks

[The pessimistic optimist]

Alan Hicks, on 18 February 2021 - 04:25 AM, said:

I hope they delete those, or at least reduce them to three people or two...whatever, the match-maker still sucks

They need to fix the population to fix the matchmaker. Letting groups farm baby seals hurts the population and by extension the matchmaker

Edited by SirSmokes, 18 February 2021 - 04:28 AM.

[TENTACLE BOSS]

Samial, on 17 February 2021 - 11:52 PM, said:

As a person that plays with friends every night as a 4 man i agree pre mades should not be influencing the pug solos.. They literally destroy the matchmaker and the game experience..

If i have to play alone so be it if it makes the matches a better experience.. Having multiple tier 1 people exploiting the matchmaker daily with the group queue so they can stomp seals needs to go asap.

So completely gut the ability to play with your friends because you might lose occasionally to a group of four? That's ******* stupid. Playing socially is an important part of MWO being a quality experience for me, I can take the occasional stomp if it means I can also reliably find matches with a friend or two in a few minutes.

If they take away my ability to do this then there are greener pastures for social shooters where I don't have to wait 45 minutes to get put up against a team of sweats, just to be able to play with friends. That sentiment will be echoed and this community is already small. The community manager can't expect financial support from the community this year if they are planning to make moves to stop us from being able to reliably play MWO together with our friends. At that point their updates are actively working against my interests and hurting the game experience.

[Wildstreak]

My best belief for matchmaker would require 3 queues.

4 v 4 for Solos only, they prefer fast matches anyway.

8 v 8 for mixed.

12 v 12 for groups.

[The pessimistic optimist]

TENTACLE BOSS, on 18 February 2021 - 04:45 AM, said:

So completely gut the ability to play with your friends because you might lose occasionally to a group of four? That's ******* stupid. Playing socially is an important part of MWO being a quality experience for me, I can take the occasional stomp if it means I can also reliably find matches with a friend or two in a few minutes.

If they take away my ability to do this then there are greener pastures for social shooters where I don't have to wait 45 minutes to get put up against a team of sweats, just to be able to play with friends. That sentiment will be echoed and this community is already small. The community manager can't expect financial support from the community this year if they are planning to make moves to stop us from being able to reliably play MWO together with our friends. At that point their updates are actively working against my interests and hurting the game experience.

Why they need two 8v8 queue's one for solo and one for groups. Let people drop in groups of 2 to 8 on the group queue and there a 8v8 solo queue just my two cents

[PocketYoda]

TENTACLE BOSS, on 18 February 2021 - 04:45 AM, said:

So completely gut the ability to play with your friends because you might lose occasionally to a group of four?

If it means the game prospers from the less toxic behavior then so be it.. I'll deal with group que with friends or just play solo.

MeanMachinE, on 18 February 2021 - 03:25 AM, said:

I really have not noticed this. In my experience the matches were equally or possibly even more lopsided in the solo queue before groups joined in them. I would still say the result will be close to same if groups are removed. The real problem is with the matchmaker which should be fixed.

I'm seeing multiple or more stomps a night..

Edited by Samial, 18 February 2021 - 07:08 AM.

[Willard Phule]

Aidan Crenshaw, on 17 February 2021 - 10:13 PM, said:

I did. And cherry picking and fitting it to your own weird logic is what you do.
See below.



I'll leave that latest post of Paul stand for itself, for every one to read. As you can see, the matchmaker issues Group PSR in some way, but we are NOT told how exactly it does that.


Not happening. See above.

How is the PSR of a group counted? Is it average, as my personal experience suggests, or is it max PSR of all members, as it should? Does being in a group apply a flat raise of PSR, which would be even better in my eyes? We don't know.
Now, what do we know? From the charts Paul provided, we can see that the Matchmaker does not really use the Tier in itself for matchmaking, but rather PSR value and Tier is just an abstraction. So being in Tier 1 at PSR value 4002 does not make you equal to the maxed-out Tier 1 player at PSR value 5000.
I also know, by experience, that groups present on a match kick the weight class balance out of the window. Yet I have seen matches where no groups were present, where this balance still was achieved.
We don't know much about the matchmaking algorithm and I don't bank on PG ever letting us in on all the dirty details.


We are still waiting on that talk, but I guess that's on Daeron now and pushed back with the Roadmap 2021.

The weight part is easy to understand. That was part of the description as well. Groups are given a tonnage limit based on the size of the group. If the group decides to drop light, the team they get placed on will be light as well. That much we've seen on a daily basis, since the group is placed in the same lance.

As for the PSR/Tier thing, he doesn't explain how the MM is taking it into consideration beyond "If a group of 3 Tier 1 players decides to add a Tier 3 or lower player to their group, there's nothing we can do about that." Does that mean all four are considered Tier 1, or all four are considered Tier 3?

From my experience, being nothing but a solo, I've seen the low tier groups get slapped into a T1 match and get run over like they were just standing still enough that it's fairly obvious that the MM does not try to match groups with an equivelent PSR/Tier level as the solos. It's also why you don't see many groups outside the T1 comp guys anymore. And I can't blame them. If I were a new T5 struggling in T5 matches, then decided to team up with three other guys and ended up in a T1 lance, I wouldn't group up again either.

[Wolfos31]

I'd really like to see them release the math behind matchmaker. Like, is that copyrighted code that they're trying to protect?

If there isn't any balance applied to groups then that is a major issue in my eyes. It should be an average of the players IMO. I don't think it should be the top player's PSR used. Because if I'm tier 2 and I drop with my tier 5 friend they're going to get absolutely steamrolled. I'd rather we drop in like tier 3. I may be overskilled, but my friend is underskilled for that tier. So in theory our contribution should average out.

But not knowing what goes on behind the curtain is very frustrating. But I don't think that this info is commonly shared for other games. I just did a quick google search on Overwatch matchmaking and couldn't seem to find an official source on how it works. Though they seem to have a tier & rating system similar to MWO. Visible tiers and invisible ratings.

[CFC Conky]

I think this game will always be tough on new/newish/casual players, soupQ or not.

I play other PvP games and the reality is that until one gets a certain amount of experience/competency, one's opponent doesn't have to be all that much better in order for you to get rekt. Probably the fastest way to get better is, ironically, for one to play with a group that can guide them and watch their six until they get used to the maps/mech movement/weapons/aiming/etc.

Nobody likes to get stomped (I think ). How often does it happen and how many times does it have to happen before you get turned off? 1, 10, 20 times? Does it happen multiple times in a row? I'm pretty sure that in any PvP shooter you're going to get stomped a whole lot at first.

If in every match I had to go up against a four-man from JGx, Goon, EmP, 228, etc, it would be very discouraging, but let's be honest here, does this actually happen to most new and/or low Tier players? I don't think so.

All in all, while I believe soupQ has been a net-negative to pugs, I don't see it as a game-killer.

Imo it's more of a symptom of other issues with MWO.

Good hunting,
CFC Conky

Edited by CFC Conky, 18 February 2021 - 10:50 AM.

[Vlad Ward]

MWO is a game with an extremely low performance floor. 90% of your in-game performance is decided in the Mechlab. Matchmaking was always going to be a crapshoot at best.

On top of that, no respawn means one misplay and you're out of the match. No amount of nerfing other players is going to make unhappy folks happy. It's just the nature of the game.

[The pessimistic optimist]

Vlad Ward, on 18 February 2021 - 11:55 AM, said:

MWO is a game with an extremely low performance floor. 90% of your in-game performance is decided in the Mechlab. Matchmaking was always going to be a crapshoot at best.

On top of that, no respawn means one misplay and you're out of the match. No amount of nerfing other players is going to make unhappy folks happy. It's just the nature of the game.

Why do army's fight together and not solo? Is it because they don't want to nerf the army?

Edited by SirSmokes, 18 February 2021 - 01:48 PM.

[Vlad Ward]

SirSmokes, on 18 February 2021 - 01:47 PM, said:

Why do army's fight together and not solo? Is it because they don't want to nerf the army?

I think you're missing my point.

Stomp rate in the pure PUG-only queue was >30%. It's a completely expected consequence of the core game mechanics. No amount of Matchmaker patches or redesigns will change that. Folks unhappy with MWO because of stomps will never be happy with MWO.

Edited by Vlad Ward, 18 February 2021 - 02:30 PM.

[The pessimistic optimist]

Vlad Ward, on 18 February 2021 - 02:28 PM, said:

I think you're missing my point.

Stomp rate in the pure PUG-only queue was >30%. It's a completely expected consequence of the core game mechanics. No amount of Matchmaker patches or redesigns will change that. Folks unhappy with MWO because of stomps will never be happy with MWO.

That is just silly it that + the other factors on top like groups. Those factor don't happen in a vacuum

Edited by SirSmokes, 18 February 2021 - 02:35 PM.

[LowSubmarino]

Wolfos31, on 17 February 2021 - 12:29 PM, said:

To give an example. On Jarl's List Low Submarino is ranked 1024 for Jan 2021 and I'm ranked 2335 for Jan 2021. I am JUST outside of Tier 1. I expect I'll get into Tier 1 and still be ranked somewhere around 2000 even. There are a 1000 players better than me and weaker than Low Submarino, and above them are another 1000 players to the peak of Jarl's List. People just barely in T1 are much worse than the folks at the peak if T1. It's a very broad definition of skill.

I also recognize that Jarl's isn't the end all be all of skill ranking, but we work with the tools we have.

Edit to add:

I would say, I've never claimed to be great at this game. But I wouldn't characterize myself as low skill or "truly bad". If I get matched up against peak players I'm sure I'll look like hell. But in the matches I play I consistently pull 400+ damage and a couple of kills.

I still feel that the new PSR system is a more accurate representation of skill. I believe that MOST T1 players are better than MOST T2 players. (assuming their PSR has more or less stabilized)

Thats cute...no, man. Jarl's isnt accurate. For a variety of reasons. I know some very strong players that are somewhere between rank 2000 and 5000. where you are. Theyre really good. And I know players that swim around somewhere in the first 200 ranks and theyre average at best.

But if you wanne believe that the tier system or your jarls rank means much then suit yourself. Jarls can be an indicator of skill but it disuiguises or fails to adress a lot of other variables.

And you need to look at consistency. Not at individual seasons. A high dmg outpout combined with like 4 to 5 kills on average....is probably pretty good. Thats a pretty strong player. And someone that on average gets one kill or less, is probably not very good all in all.

Thats jars.

But the tier system literally means ****. It doesnt say anything about a player. Youll find players that shut down cause of overheating 5 times in a row in so called tier 1. And you got very good players staying in tier 3 to 5 on purpose cause they wanne test builds or are juts a bit on the evil pug-slaughtering side. Or you have ppl that play 10 games per season and get lucky or unlucky with their win loss ratio and suddenly jump to rank 50 or even lower or to rank ten hundred kazillion even though theyre very good.

And everything in between.

The tier system has one purpose. It serves to give you the impression of progression. Almost everyobdy gets a bit emotional and proud when breaching tier 1. Thats exactly why it is there. But it doesnt say much. Its basically an experience bar.

[Vlad Ward]

I mean, the OG PSR bar was absolutely an experience bar.

Post-2020 PSR is zero-sum, but movement is so slow that it takes hundreds of games to place people correctly. For casual players who can click on robots, it's going to behave just like an experience bar for weeks or months until they finally level out.

Edited by Vlad Ward, 18 February 2021 - 06:55 PM.

[TENTACLE BOSS]

Samial, on 18 February 2021 - 07:03 AM, said:

If it means the game prospers from the less toxic behavior then so be it.. I'll deal with group que with friends or just play solo. I'm seeing multiple or more stomps a night..

Alright have fun prospering with less players on a game that barely attracts newcomers. Nothing is prospering when newcomers can't play with their friends, you are left with a hollow, isolated experience, all in an attempt to tackle "toxic behavior."

Co-ordinating with a friend to do something, like "lets try out this mech combination and see how it goes" is a rewarding element of the game and shouldn't be viewed as toxic, and I can take the occasional stomp if it means I get to retain the social experience.

[Wildstreak]

PSR / Tier is now more like old ELO which means back to the pyramid system instead of the cylinder.

Still online games need something better.

[Anomalocaris]

TENTACLE BOSS, on 18 February 2021 - 06:57 PM, said:

Alright have fun prospering with less players on a game that barely attracts newcomers. Nothing is prospering when newcomers can't play with their friends, you are left with a hollow, isolated experience, all in an attempt to tackle "toxic behavior."

Co-ordinating with a friend to do something, like "lets try out this mech combination and see how it goes" is a rewarding element of the game and shouldn't be viewed as toxic, and I can take the occasional stomp if it means I get to retain the social experience.

Once again, people are not actually looking at the numbers.

First of all, before we get to the numbers - again - telling people they should play the game your way is just silly. After the queues were merged we lost player population 3 of the following 4 months (the month of PSR reset being the only gain). Only when PGI announced with the hiring of Daeron that they were going to start developing (lol) again did population start to climb.

But when you delve into the numbers, you get some interesting detail. We can only count players who play 10 or more games in a season as that's what PGI reports, but TBH, if they are playing less they really aren't contributing much to the population available for matchmaking anyways. From April 2020 till now the game added about 1800 new players per month (1500 to 2300 is the range). Some could have been smurf accounts, but no way to tell. If you look at overall population numbers, those new players often don't even balance out the loss of existing players. Simply look at population growth from one month to the next and subtract the new players added.

Existing players gained/lost (rough numbers):

May - +100
June - -3200
July - -550
August - -2900
September - -2900
October - -500
November - +100
December - +1000 (giveaways?)
January - -2100

Given that in most months we are losing old accounts, it is clearly new a t that are keeping the game alive, assuming they aren't all smurfs enjoying seal clubbing in Tier 4 and 5.

Furthermore, look at the population difference between April, after which the queue merge took place, and now. Despite the player loss that occurred after the merge, we are 50% higher than we were then. And average player numbers online at any given time are also up 50%. Compared the lows in February 2020 (which was 3 months before the queue merge) we are on pace to have 133% more people playing at any given time. We are now on par with April 2019 in terms of players online, and June 2019 in terms of population size, and no one was talking about merging queues then. We didn't even try 8v8 group queue until April 2020.

Unless PGI believes the current population is unsustainable due to disillusionment with their inability to deliver real change going forward, there is clearly a big enough population to reinstitute a group queue. 8v8 was very well received and would do even better now, especially with solo opt-in. I'm ok with letting one (and only one) 2-man group on each team in solo queue to protect those shepherding new players from more predatory groups. But that's as much as the queue can tolerate IMO.

I think we'd also see more returning players with a solo queue relaunch. We lost a lot to the crapiness of the soup merge.

Oh I know some will argue that it just won't work, but I suspect that most of them want to play in groups without having to face good players on the other side, which will happen more often in a group only queue.

See ya next month.

[TENTACLE BOSS]

Anomalocaris, on 18 February 2021 - 07:57 PM, said:

Once again, people are not actually looking at the numbers. First of all, before we get to the numbers - again - telling people they should play the game your way is just silly. After the queues were merged we lost player population 3 of the following 4 months (the month of PSR reset being the only gain). Only when PGI announced with the hiring of Daeron that they were going to start developing (lol) again did population start to climb. But when you delve into the numbers, you get some interesting detail. We can only count players who play 10 or more games in a season as that's what PGI reports, but TBH, if they are playing less they really aren't contributing much to the population available for matchmaking anyways. From April 2020 till now the game added about 1800 new players per month (1500 to 2300 is the range). Some could have been smurf accounts, but no way to tell. If you look at overall population numbers, those new players often don't even balance out the loss of existing players. Simply look at population growth from one month to the next and subtract the new players added. Existing players gained/lost (rough numbers): May - +100 June - -3200 July - -550 August - -2900 September - -2900 October - -500 November - +100 December - +1000 (giveaways?) January - -2100 Given that in most months we are losing old accounts, it is clearly new a t that are keeping the game alive, assuming they aren't all smurfs enjoying seal clubbing in Tier 4 and 5. Furthermore, look at the population difference between April, after which the queue merge took place, and now. Despite the player loss that occurred after the merge, we are 50% higher than we were then. And average player numbers online at any given time are also up 50%. Compared the lows in February 2020 (which was 3 months before the queue merge) we are on pace to have 133% more people playing at any given time. We are now on par with April 2019 in terms of players online, and June 2019 in terms of population size, and no one was talking about merging queues then. We didn't even try 8v8 group queue until April 2020. Unless PGI believes the current population is unsustainable due to disillusionment with their inability to deliver real change going forward, there is clearly a big enough population to reinstitute a group queue. 8v8 was very well received and would do even better now, especially with solo opt-in. I'm ok with letting one (and only one) 2-man group on each team in solo queue to protect those shepherding new players from more predatory groups. But that's as much as the queue can tolerate IMO. I think we'd also see more returning players with a solo queue relaunch. We lost a lot to the crapiness of the soup merge. Oh I know some will argue that it just won't work, but I suspect that most of them want to play in groups without having to face good players on the other side, which will happen more often in a group only queue. See ya next month.

In the past, less than a year ago you could literally wait for multiple hours before you found a single group game which was often stacked, according to your extrapolated take on limited data we're on par to a population size roughly the same as what we had early/mid last year and you think the solution is to go back to that. According to you the queue can only tolerate one two man team per team in a game where you could frequently find your friends in the same matches on the opposite team without queuing together. (Because it wasn't possible back then)

Based on ******* what? All enforcing that would do is lead to an annoying algorithm where people who are trying to play together with a friend are arbitrarily unable to join matches looking for players when those matches need people. Your "solution" would slow down queue times for everyone and force people in groups of two to wait for extended periods due to the limited amount of simultaneous matches being played at once. This idea is an over-complication that creates a bigger problem than the occasional stomp. Teams that communicate are always going to have an advantage, even if there are no groups in the quickplay queue in the future, you will still see one sided stomps all the time when cadets go up against tier 1's, one team communicates while another does not etc.

The definitive consequence is that when you remove the ability for people to play together in quickplay you kill the ability for people to reliably play matches with their friends. Best case scenario, people then play solo, while others who wanted to play with their friends drop off to titles that actually support social play. That will undoubtedly affect new players, especially those looking to play the game together, and any spikes you are seeing will drop off as those players launch the game, try to queue, wait an hour (if they are that patient), go "**** this" and play something else.

[Anomalocaris]

TENTACLE BOSS, on 18 February 2021 - 11:47 PM, said:

In the past, less than a year ago you could literally wait for multiple hours before you found a single group game which was often stacked, according to your extrapolated take on limited data we're on par to a population size roughly the same as what we had early/mid last year and you think the solution is to go back to that. According to you the queue can only tolerate one two man team per team in a game where you could frequently find your friends in the same matches on the opposite team without queuing together. (Because it wasn't possible back then)

Based on ******* what? All enforcing that would do is lead to an annoying algorithm where people who are trying to play together with a friend are arbitrarily unable to join matches looking for players when those matches need people. Your "solution" would slow down queue times for everyone and force people in groups of two to wait for extended periods due to the limited amount of simultaneous matches being played at once. This idea is an over-complication that creates a bigger problem than the occasional stomp. Teams that communicate are always going to have an advantage, even if there are no groups in the quickplay queue in the future, you will still see one sided stomps all the time when cadets go up against tier 1's, one team communicates while another does not etc.

The definitive consequence is that when you remove the ability for people to play together in quickplay you kill the ability for people to reliably play matches with their friends. Best case scenario, people then play solo, while others who wanted to play with their friends drop off to titles that actually support social play. That will undoubtedly affect new players, especially those looking to play the game together, and any spikes you are seeing will drop off as those players launch the game, try to queue, wait an hour (if they are that patient), go "**** this" and play something else.

Aww shucks, you didn't read the post did you?

Population size is on par with where we were nearly _2_ years ago. Last February was the lowest recorded population since data has been published. We are now well over twice that and on par to see the population hit 133% of where we were at the low. 2 years ago we were not discussing merging queues because it wasn't needed.

I don't want any groups in _solo_ QP. But I'm willing to compromise, are you? I can understand the rationale for not wanting to subject a new player to group queue, which will generally have more sharks in it. So bring your new friend in a 2-man to solo.

But for bigger groups, we clearly have enough players now to reinstitute 8v8 group queue with solo opt-ins. Your assessment of what would happen is completely wrong. I mean entirely, without exception, wrong. If we have an 8v8 group queue you can actually choose to drop in solo or group with a 2 man, or you can get a bigger group and drop 8v8 only.

Returning to split queues would bring back a ton of solo players who left the game because they got tired of the absolute dumpster fire that soup queue created. Horrible matchmaking, huge tonnage mismatches, screwed up spawns and being forced to drop with groups which alternately swing from the extremes of high level killers in max tonnage meta lances to complete clownshoe wearing, legally intoxicated, mech-dads playing their lore builds in a severely under-tonned lance.

And for groups of 2, you'd have an even bigger population to choose from since you could drop in solo or group. And for solo players who want minimum wait time, they can roll the dice and opt in to group drops. And they already did 8v8, so the software engineering is done, plus people really liked it. Seems like a win-win to me, and we get to undo this ridiculous soup queue matchmaker. Hell, PGI wouldn't need to worry about fixing the spawns anymore either. More resources for other stuff!! Win-win-win