113 replies to this topic

[Bud Crue]

Kroete, on 01 December 2020 - 03:08 AM, said:

And we still see groups syncdropping more then 4 in a match.
Seems both is true.

Lets see how this is supposed to work:

Me and mine are typically a 4-8 man group.

"Oh! you're a group! You must go to Faction Play!" say the gatekeepers of this topic. "That is the queue for you, as all others shall be forbidden, so that you and your elite friends can't farm us pure and noble solos who just want to play the game and have fun."

Me and my friends look at each other, then at our stats, and think: "What the hell are these morons on about? Elite? The best of us barely manage a 1.0 w/l ratio and most of us rarely survive a match or even break 300 damage. But whatever, the gate keepers know best!"

So off we go to drop as an 8 man into faction play. Wait what is this message screen? Hmm...lets read this..."something something, highly competitive, something team cohesion, specialized mechs, something about a focus on tactics and team work, well coordinated teams, yada, yada, yada." "Oh yeah! Faction play the mode for the truly competitive folks in the game! Duh! Well that is certainly not us but the gatekeepers of soup queue all say this is where we belong!"

Half an hour later, after finding a match, and getting rolled, down 24-3; one of the nice fellows on the other team helpfully types out that maybe you guys should play better mechs/builds, and another red mentions that we should maybe tell our pugs to stop reinforcing (we did, we did), and another says ya know if you see that its conquest you need to try to control... and so on.

At which point we all collectively remember why we don't play faction play. It's because we aren't good enough players to cut it in a mode designed by its very premise to be where the elite go to play in a highly competitive environment specifically designed for them! We learned that 3 years ago. Faction Play is no place for a bunch of casuals who just play the game as a means of hanging out after work and doing a common activity while we drink some beers and chat.

So where do we go? Well we used to play GQ. Every night. Until it died. Now we are told that soup queue is where we should go. But no! The gate kaeepers of sportsman ship and fair play say that two four mans sync dropping against each other are cheating (somehow...yes hey throw matches to take turns winning!)! or are just playing to club seals (despite most so-called seals having better stats)! or are trying to avoid other groups (despite playing match after match against other groups)! or some other nonsensical and internally contradictory thing.

But boy, it sure is nice that you guys have this all figured out and are doing your part to contravene small casual groups from playing the game as PGI specifically designed it to be played these days.

[R Valentine]

Monkey Lover, on 27 November 2020 - 07:12 AM, said:

You basically have same win ratio as I do and I only play in groups in most cases of 4 lol

Groups of 4 bad players determine the outcome of the game the same as the groups of 4 amazing players. The only difference is they make their team lose instead of win. Either way, it was the group that skewed the odds.

[Anomalocaris]

Kiran Yagami, on 01 December 2020 - 06:28 AM, said:

Groups of 4 bad players determine the outcome of the game the same as the groups of 4 amazing players. The only difference is they make their team lose instead of win. Either way, it was the group that skewed the odds.

Thank you. Thank you so very much. All you group drop supporters who say "but we're not good enough to club seals" or "I beat groups all the time" are missing the point. Unless the average skill level of your group is approximately the same as the rest of the team (and the same must be true for the enemy team), you are having a matchmaker breaking effect.

Whether you are good or bad is irrelevant. It's the fact that you are aggregating your skill level outside of the normal solo MM parameters. The further your group skill level from the mean of the rest of the players in the match, the worse the mismatch becomes. Are we clear now? Groups are bad in mixed queue because the MM can't account for your skill. And that's not going to change because PGI are incapable or unwilling to build the systems to account for it. So the only solution is to separate the queues, or at least minimize group size to 2.

Yes, I think you should have a place to play. Bring back group queue 8v8 with group sizes of 2,4 or 8. But your inclusion into solo queue has objectively reduced match quality and arguably had a negative effect on population.

[RickySpanish]

Anomalocaris, on 01 December 2020 - 06:41 AM, said:

Thank you. Thank you so very much. All you group drop supporters who say "but we're not good enough to club seals" or "I beat groups all the time" are missing the point. Unless the average skill level of your group is approximately the same as the rest of the team (and the same must be true for the enemy team), you are having a matchmaker breaking effect.

Whether you are good or bad is irrelevant. It's the fact that you are aggregating your skill level outside of the normal solo MM parameters. The further your group skill level from the mean of the rest of the players in the match, the worse the mismatch becomes. Are we clear now? Groups are bad in mixed queue because the MM can't account for your skill. And that's not going to change because PGI are incapable or unwilling to build the systems to account for it. So the only solution is to separate the queues, or at least minimize group size to 2.

Yes, I think you should have a place to play. Bring back group queue 8v8 with group sizes of 2,4 or 8. But your inclusion into solo queue has objectively reduced match quality and arguably had a negative effect on population.

Oh Puhleeease. The match maker HAS NEVER been good at balancing teams. NOTHING has changed. Back in regular solo queue you would frequently see 1k damage and 4/5 kills from one guy quite blatantly carrying the game. The vast majority of groups in game are already exactly the same skill level as the rest of the solo players, you just have a little panic when you see three or four identical unit tags show up is all. Don't go labelling "group players" as some zeitgeist of quick play's current woes, the problem is ding-dongs who actively refuse to cooperate in a team based game. The players who don't use comms, don't call targets, don't spot flanks, don't stick together, bring obviously daft builds - they are the reason your team is falling apart. Sometimes that's the 4 man Urbie group, but usually it's not. Ask yourself: "How many times has a four man group on my team coordinated the rest of the team, drop called, or even stuck together in the same grid square?". Almost never, right? Well unless you are always ending up with the duff group in the match, guess what? That four man group is playing for a laugh and are no better equipped to win or lose than you are.

Edited by RickySpanish, 01 December 2020 - 07:04 AM.

[R Valentine]

RickySpanish, on 01 December 2020 - 06:57 AM, said:

Oh Puhleeease. The match maker HAS NEVER been good at balancing teams. NOTHING has changed. Back in regular solo queue you would frequently see 1k damage and 4/5 kills from one guy quite blatantly carrying the game. The vast majority of groups in game are already exactly the same skill level as the rest of the solo players, you just have a little panic when you see three or four identical unit tags show up is all. Don't go labelling "group players" as some zeitgeist of quick play's current woes, the problem is ding-dongs who actively refuse to cooperate in a team based game. The players who don't use comms, don't call targets, don't spot flanks, don't stick together, bring obviously daft builds - they are the reason your team is falling apart. Sometimes that's the 4 man Urbie group, but usually it's not.

There's a world of difference between solo queue matchmaker being bad and implementing a queue in which it is technically impossible for solo queue matchmaker to ever be good. At least when everyone drops on their lonesome, there exists the possibility of a system that can create a half decent match. In soup queue, that's not even a pipe dream. It's outside the realm of possibility. The solo queue match maker can't account for teams of 4 extreme players, whether extremely bad or extremely good, and putting all 4 of those players on a single team robs it of the ability to level the playing field. Even if matchmaker, by some cosmic miracle, could accurately calculate the skill+force multiplier of being in a coordinated group, the odds it will find another team of equal value to pit them against is virtually non-existent. Our player pool is too small. You've created a queue that cannot be balanced in even the best of circumstances. Solo queue at least gave matchmaker a chance.

Edited by Kiran Yagami, 01 December 2020 - 07:08 AM.

[Gagis]

I'm optimistic and think that the MM might do fine if groups were counted as 4 players sharing the highest individual PSR out of its members.

[Brauer]

Gagis, on 01 December 2020 - 07:14 AM, said:

I'm optimistic and think that the MM might do fine if groups were counted as 4 players sharing the highest individual PSR out of its members.

Why though? As Kiran said due to the size of the playerbase it would not be surprising if there just aren't players in the queue to create balanced matches with groups at the extremes. When you have four players from a Div A/upper B comp team in a group, or four players with sub-1.00 WLRs in a group there isn't necessarily another group or group of solos who are in the queue to balance that out. Yet it's not uncommon as a solo to run into the same groups over and over who either throw their games extremely hard or hard-carry 90% plus of the time.

[Nightbird]

Brauer, on 01 December 2020 - 07:29 AM, said:

Why though? As Kiran said due to the size of the playerbase it would not be surprising if there just aren't players in the queue to create balanced matches with groups at the extremes. When you have four players from a Div A/upper B comp team in a group, or four players with sub-1.00 WLRs in a group there isn't necessarily another group or group of solos who are in the queue to balance that out. Yet it's not uncommon as a solo to run into the same groups over and over who either throw their games extremely hard or hard-carry 90% plus of the time.

Wrong, the player base is plenty big enough to create balanced matches.

[R Valentine]

Nightbird, on 01 December 2020 - 07:36 AM, said:

Wrong, the player base is plenty big enough to create balanced matches.

Even with soup queue? Because soup queue requires not only a player pool big enough, but a pool of groups big enough to have enough groups of similar skill level. You really think we have that many people?

[Nightbird]

Kiran Yagami, on 01 December 2020 - 08:22 AM, said:

Even with soup queue? Because soup queue requires not only a player pool big enough, but a pool of groups big enough to have enough groups of similar skill level. You really think we have that many people?

I made a whole thread 5 months ago proving that Matchmaking can be much better.

https://mwomercs.com...cy-with-graphs/

In one of the latest posts I showed how the current PSR and MM performed exactly as predicted all those months ago.

https://mwomercs.com...ost__p__6361878

[Anomalocaris]

RickySpanish, on 01 December 2020 - 06:57 AM, said:

Oh Puhleeease. The match maker HAS NEVER been good at balancing teams. NOTHING has changed. Back in regular solo queue you would frequently see 1k damage and 4/5 kills from one guy quite blatantly carrying the game. The vast majority of groups in game are already exactly the same skill level as the rest of the solo players, you just have a little panic when you see three or four identical unit tags show up is all. Don't go labelling "group players" as some zeitgeist of quick play's current woes, the problem is ding-dongs who actively refuse to cooperate in a team based game. The players who don't use comms, don't call targets, don't spot flanks, don't stick together, bring obviously daft builds - they are the reason your team is falling apart. Sometimes that's the 4 man Urbie group, but usually it's not. Ask yourself: "How many times has a four man group on my team coordinated the rest of the team, drop called, or even stuck together in the same grid square?". Almost never, right? Well unless you are always ending up with the duff group in the match, guess what? That four man group is playing for a laugh and are no better equipped to win or lose than you are.

Time to burn some strawmen.

1. Claiming that the "vast majority of groups are exactly the same skill level as the rest of solo players" is unsupportable. If you average all the groups together? Sure. But we're talking about 4 man subsets of the player base, who tend to aggregate for various reasons. You cannot make claims about their skill levels with the data available.

2. Nobody labelled group players as the source of all woes. The problem is that mixing groups and solos objectively reduces the quality of games. Per PGI's own published data. And I guarantee that if they had some good news about it since then they would have shared it - but they haven't.

3. The matchmaker cannot account for groups that substantially deviate in average skill from the rest of their teams. It should be able to. Shouldn't be that hard to implement. But this is PGI and Paul here, and they cannot do it, or simply refuse to. And because the MM can't account, when group skill deviates from team skill you screw up matchmaking. Even if it only happens some of the time, it's a net negative on matchmaking balance.

The rest about team communication, etc. is irrelevant, because we're talking about how groups break the soup queue matchmaker. Argue that they don't, we'll wait.

[General Solo]

RickySpanish, on 01 December 2020 - 06:57 AM, said:

Oh Puhleeease. The match maker HAS NEVER been good at balancing teams. NOTHING has changed.

Never is such a harsh word and I disagree.
Back when the ole PSR matchmaker was new, it was OK and did a decent job even in group queue.
If you remember it was the upward bias that reduced its effectiveness.

Don't even go where they fixed it with that switch and bait MM solution and then the merge on top. Ridiculous, OUTRAGEOUS!

Back to MM, back before upward bias took hold of the old MM, it did a decent job.

In my old unit, if we ran a mixed group with mostly lower Tier pilots plus me, we did not get EMP in our games.

When we ran a Tier 1 and 2 group when got EMP very often in our games.
So it worked back then, but was not monitored, maintained and so upward was allowed to set in.

Once upward started, well everyone was Tier 1 and the MM could not do its job good.

So in fact Match maker did work. Just maintained by an Arthouse Dev studio so what you expect.

It went downhill eventually.

Art and mech pack monetization can only take the game so far.

[General Solo]

Kiran Yagami, on 01 December 2020 - 06:28 AM, said:

Groups of 4 bad players determine the outcome of the game the same as the groups of 4 amazing players. The only difference is they make their team lose instead of win. Either way, it was the group that skewed the odds.

Yes

And if those same groups of 4 keep having the experience of being constantly out matched, walloped around the cockpit with skillgap match after match.

Well that can't be good for their morale or the games chances of retaining them.

Which is why Match making is so important.

Especially in a mixed queue.

[Monkey Lover]

Kiran Yagami, on 01 December 2020 - 06:28 AM, said:

Groups of 4 bad players determine the outcome of the game the same as the groups of 4 amazing players. The only difference is they make their team lose instead of win. Either way, it was the group that skewed the odds.

At the extremes of both sides. But when you see most groups are average players it does not skew the odds.

Again i will point out im a average to slightly above. The people I play with are average to slightly above. We have about a 1 win/ratio.


You people want to ban all groups because there is a dozen top lever players kicking your butt. Maybe you should try to focus them down. That's what we do....

Edited by Monkey Lover, 02 December 2020 - 05:51 AM.

[R Valentine]

Monkey Lover, on 01 December 2020 - 11:10 PM, said:

At the extremes of both sides. But when you see most groups are average players it does not skew the odds.

Again i will point out im a average to slightly above. The people I play with are average to slightly above. We have about a 1 win/ratio.


You people want to ban all groups because there is a dozen top lever players kicking your butt. Maybe you should try to focus them down. That's what we do....

Even in a world where most groups are "average", the force multiplier of being in a group skews the odds already. It's extremely common to find a group on one team and no group on the other team because our player pool is so small. There's no good reason to mix the queues. If you want a group queue, make one. 8v8 was very popular for the brief time it was out. Put 8v8 group queue back in, let solo queue be as it was. We only lost content when the queues were merged. We didn't gain any.

[Monkey Lover]

Kiran Yagami, on 02 December 2020 - 06:38 AM, said:

Even in a world where most groups are "average", the force multiplier of being in a group skews the odds already. It's extremely common to find a group on one team and no group on the other team because our player pool is so small. There's no good reason to mix the queues. If you want a group queue, make one. 8v8 was very popular for the brief time it was out. Put 8v8 group queue back in, let solo queue be as it was. We only lost content when the queues were merged. We didn't gain any.

Almost every time I play there is a team on the other size. More often now the teams are mix units this due to the population of teams decreasing. So its hard to tell unless you ask them. I can see it because we play them game after game and at this point we know their names. Our teams often have 1 or 2 different unit tags when people jump in from the discord. Anyone here is welcome to join the Marik discord too.

Other than removing teams from the game I would support teams only being able to play when there are teams on both sides. No teams ver solo teams would be fine with me.

Sorry to say even with 8v8 at the end of the test we started not finding games again. There is no population for this. If there was I would support it too.

Edited by Monkey Lover, 02 December 2020 - 02:09 PM.

[Anomalocaris]

Monkey Lover, on 02 December 2020 - 02:05 PM, said:

Almost every time I play there is a team on the other size. More often now the teams are mix units this due to the population of teams decreasing. So its hard to tell unless you ask them. I can see it because we play them game after game and at this point we know their names. Our teams often have 1 or 2 different unit tags when people jump in from the discord. Anyone here is welcome to join the Marik discord too.

Other than removing teams from the game I would support teams only being able to play when there are teams on both sides. No teams ver solo teams would be fine with me.

Sorry to say even with 8v8 at the end of the test we started not finding games again. There is no population for this. If there was I would support it too.

If almost every time you play there is a team on the other side, then clearly there are enough teams to support a dedicated 8v8 group queue. Remember it only takes four 4man teams to make an 8v8 match. During the lowest population times (early morning North America) there are still 500-600 players online. At peak there are 1800-2000 players. If only half those people are actually playing at one time there are still 10-40 matches going on. If there are groups in half those matches (on both sides) that's 10-40 groups playing. That's plenty for an 8v8 group queue, especially where groups are limited to 2 or 4 man only for easier matchmaking.

Assumptions here:
(1) total players online is equal to 2x the steam charts numbers. It may be higher as PGI has indicated in the past. Non-steam players may be up to twice steam player #s.
(2) Only half of the players online at any given time are actually playing soup queue. Some could be in the mech lab, some could be in faction, whatever. I suspect the number in solo are actually higher.
(3) Only half the soup queue games have a group in them.
(4) Those groups are 4 man (in reality they probably average between 3 and 4 players).

Pretty conservative estimates, yet we could easily have 5-20 group matches going on at any time in an 8v8 setup with easy matchmaking if we limit to 2 or 4 man groups.

Edited by Anomalocaris, 02 December 2020 - 02:36 PM.

[East Indy]

Reset PSR is actually not half bad, especially in Tier 1 games.

Mismatched groups are almost entirely the problem. In this corner, ringers in their matching-set Vapor Eagle meta chic. And in this corner...not that.

[Monkey Lover]

Anomalocaris, on 02 December 2020 - 02:21 PM, said:

If almost every time you play there is a team on the other side, then clearly there are enough teams to support a dedicated 8v8 group queue. Remember it only takes four 4man teams to make an 8v8 match. During the lowest population times (early morning North America) there are still 500-600 players online. At peak there are 1800-2000 players. If only half those people are actually playing at one time there are still 10-40 matches going on. If there are groups in half those matches (on both sides) that's 10-40 groups playing. That's plenty for an 8v8 group queue, especially where groups are limited to 2 or 4 man only for easier matchmaking.

Assumptions here:
(1) total players online is equal to 2x the steam charts numbers. It may be higher as PGI has indicated in the past. Non-steam players may be up to twice steam player #s.
(2) Only half of the players online at any given time are actually playing soup queue. Some could be in the mech lab, some could be in faction, whatever. I suspect the number in solo are actually higher.
(3) Only half the soup queue games have a group in them.
(4) Those groups are 4 man (in reality they probably average between 3 and 4 players).

Pretty conservative estimates, yet we could easily have 5-20 group matches going on at any time in an 8v8 setup with easy matchmaking if we limit to 2 or 4 man groups.

You understand group q died correct? This is only 2 more 4 mans than 8v8 but still didn't work.

How many team games did you play when it was 8v8? Because I know we ended nights early and went to sync dropping towards the end when we couldn't find matches.

People will not sit in q for 20 minutes when they can just sync drop.

[John Bronco]

I agree completely, dropping group size to 2 would be an acceptable compromise, in the short term.