60 replies to this topic

[Gagis]

Spawn points
Volumetric rescaling favouring smaller sizes. Mechs are too big for cover that used to be good for them.

[Horseman]

SirSmokes, on 18 November 2020 - 11:20 AM, said:

Why there forum warriors who fight to the death too keep quick play group and solo together. They love to stomp and get easy matches. I say take match size down too 8v8 and make another quick play that 2 groups of 4 lances

Stop being disingenuous. If the population was large enough to support separate solo and group queues I'd be the first person calling to split them again. It doesn't, and before the merge a lot of casual players were complaining about their inability to play with 1-2 buddies, taking their ball and ****ing off to games where they can do that. That's for whom really the change has been made, and you know it.

Edited by Horseman, 18 November 2020 - 12:35 PM.

[The pessimistic optimist]

Horseman, on 18 November 2020 - 12:33 PM, said:

Stop being disingenuous. If the population was large enough to support separate solo and group queues I'd be the first person calling to split them again. It doesn't, and before the merge a lot of casual players were complaining about their inability to play with 1-2 buddies, taking their ball and ****ing off to games where they can do that. That's for whom really the change has been made, and you know it.

If we go down to 8v8 we can make it work

[Knight Captain Morgan]

SirSmokes, on 18 November 2020 - 12:36 PM, said:

If we go down to 8v8 we can make it work

let's take it all the way down to 2V2 and have 7 different brackets so that the "I just wanna play with my buddy" crowd can pick the mechs that best suite them and then play with other "I just wanna play with my buddy" types.no, the merge was to appease the premade farmers who love roflstomping pugs but got tired of queue dodging each other in faction play.

[The pessimistic optimist]

Knight Captain Morgan, on 18 November 2020 - 01:38 PM, said:

let's take it all the way down to 2V2 and have 7 different brackets so that the "I just wanna play with my buddy" crowd can pick the mechs that best suite them and then play with other "I just wanna play with my buddy" types.no, the merge was to appease the premade farmers who love roflstomping pugs but got tired of queue dodging each other in faction play.

What are you even talking about that is all nonsense

[DevinMace]

SirSmokes, on 18 November 2020 - 11:20 AM, said:

Why there forum warriors who fight to the death too keep quick play group and solo together. They love to stomp and get easy matches. I say take match size down too 8v8 and make another quick play that 2 groups of 4 lances

Interesting point of view from a person in a unit that was created because of groups in quick play...

[Horseman]

SirSmokes, on 18 November 2020 - 01:57 PM, said:

What are you even talking about that is all nonsense

Not to throw any stones, but that's the same argument you were making a few posts ago.

[The pessimistic optimist]

Horseman, on 18 November 2020 - 02:29 PM, said:

Not to throw any stones, but that's the same argument you were making a few posts ago.

No I said 2 GROUPS of 4 making 16 player 8 on each side

[Anomalocaris]

Horseman, on 18 November 2020 - 12:33 PM, said:

Stop being disingenuous. If the population was large enough to support separate solo and group queues I'd be the first person calling to split them again. It doesn't, and before the merge a lot of casual players were complaining about their inability to play with 1-2 buddies, taking their ball and ****ing off to games where they can do that. That's for whom really the change has been made, and you know it.

So what about us solos who said f*** it and left the game because of the merge? Was that really a good decision? Reduce quality of play in the last viable queue and get rid of one set of players for another. So now everyone suffers more? Is there anyone arguing that quality of play in the merge queue isn't worse than it was solo?

Mixing groups with solos is a s***show, as you well know. You frequently end up with strong groups making everyone else irrelevant (saw a match the other day where the 4man did 8 kills and all were above 600 dmg while none of the rest of the team broke 300 dmg), or weak groups dragging down the efforts of the remaining players on the team (guys scoring 1000 pts dmg and multiple kills only to be hamstrung by a bunch of 100 pt dmg potatoes dropping together).

Yeah, all those guys existed in solo queue (good and bad), and matches like that happened too. But they were more rare, because the pros and the potatoes rarely got grouped in similarly skilled units to bork the matchmaker. A group in solo is like an exponent on the influence of the grouped players. A group with players operating at 50% of the average match performance ends up effectively playing like they are at 40% of the average (0.5^1.25) while a group that's running at 50% above has an impact like 4 solos playing at 66%% above normal (1.5^1.25). Obviously I'm pulling the exponents out of the air, but that's the general trend I see. Whereas in old solo queue, those guys would be split on opposite teams at least some of the time, balancing their impact.

And we haven't even gotten into guys trying to sync groups to get more of their buddies on the same team.

Look, I despise groups in solo. If I was still playing I'd probably be frustrated enough to start legging them just to make them feel my pain (good thing I stopped playing, eh?). But I'd be willing to compromise with one (1) 2-man per side so people can drop with a friend and show them the ropes. Otherwise throw together an 8v8 group queue with solo opt ins and start convincing people that group drops are fun and engaging. If its really what a substantial portion of the player base wants, it'll do fine. I suspect it won't, simply because a lot of guys dropping groups are doing so with intentions that they won't admit to, and a true group queue would foil those intentions quite effectively.

So yeah, my vote would be fix match maker and give me a solo queue back again.

[Horseman]

Anomalocaris, on 18 November 2020 - 04:44 PM, said:

So what about us solos who said f*** it and left the game because of the merge? Was that really a good decision? Reduce quality of play in the last viable queue and get rid of one set of players for another. So now everyone suffers more?

Losing some of the exclusively solo players vs losingall of the exclusively group players, most of the primarily group players plus a giving fair portion of the casually grouping ones reason to find other games to play with their buddies (and potentially lose them entirely to those games over time).
It's damned if you do, damned if you don't
The ability to drop together as a group allowed that last group to bring in a trickle of new players from their other social circles, which probably weighed in on PGI's decision as well.

Quote

Is there anyone arguing that quality of play in the merge queue isn't worse than it was solo?

Nope. It seems a lot more prone towards extreme results now.

Quote

Mixing groups with solos is a s***show, as you well know. You frequently end up with strong groups making everyone else irrelevant (saw a match the other day where the 4man did 8 kills and all were above 600 dmg while none of the rest of the team broke 300 dmg), or weak groups dragging down the efforts of the remaining players on the team (guys scoring 1000 pts dmg and multiple kills only to be hamstrung by a bunch of 100 pt dmg potatoes dropping together).

Yes, but at the same time failing to balance the concentrated badass or concentrated stupid sounds like a matchmaker problem to me.

We also have the relaxed tier separation to thank for some of the matchmaking shitshow on display right now, let's not pretend that didn't make a difference. MM was doing pretty well after the reset (even if it was taking a little longer to get a match) until the peanut gallery started stomping their feet that they want faster matchmaking no matter if it's worse quality. They got their wish, didn't they?

Quote

And we haven't even gotten into guys trying to sync groups to get more of their buddies on the same team.

Not as if this wasn't happening before, though.

Edited by Horseman, 18 November 2020 - 10:55 PM.

[bilagaana]

I'll come back when they get the groups out of QP.

[The pessimistic optimist]

Horseman, on 18 November 2020 - 10:52 PM, said:

Losing some of the exclusively solo players vs losingall of the exclusively group players, most of the primarily group players plus a giving fair portion of the casually grouping ones reason to find other games to play with their buddies (and potentially lose them entirely to those games over time).
It's damned if you do, damned if you don't
The ability to drop together as a group allowed that last group to bring in a trickle of new players from their other social circles, which probably weighed in on PGI's decision as well.
Nope. It seems a lot more prone towards extreme results now.
Yes, but at the same time failing to balance the concentrated badass or concentrated stupid sounds like a matchmaker problem to me.

We also have the relaxed tier separation to thank for some of the matchmaking shitshow on display right now, let's not pretend that didn't make a difference. MM was doing pretty well after the reset (even if it was taking a little longer to get a match) until the peanut gallery started stomping their feet that they want faster matchmaking no matter if it's worse quality. They got their wish, didn't they?

Not as if this wasn't happening before, though.

Can't fix the match making with no players smart guy as in if there no players don't work so great. 8v8 made up of two groups 4 to make 16 match maker to do that easier and faster and honestly I thing the game would play better with 4 less mechs on the field. Also do solo in 8v8 but that just me

[Gagis]

A match maker will work just fine all the way down to the point the game has a total of 24 players. It just needs to be able to assign some value to each player and try to divide the numbers between teams as evenly as possible.

[The pessimistic optimist]

Gagis, on 19 November 2020 - 09:17 AM, said:

A match maker will work just fine all the way down to the point the game has a total of 24 players. It just needs to be able to assign some value to each player and try to divide the numbers between teams as evenly as possible.

NO the match maker is not magic it can't make good matches out of a tiny player pool

[Hunka Junk]

Pick 2?

I choose either Gearbox or 343 Industries.

[GARION26]

Put me in the mechdad who wants to play with 1-2 buddies group.

I didn't call for the addition of groups to quick play but it's been well received in our household where I get to play with my 9 year old son - previously we were sync dropping and hoping for the best.

I do (sorry Horseman) favor faster drop matchmaking - though I recognize quality of matchmaking has certainly dropped with that change. I think the 4 minute wait time before the matchmaker opened the gates was too long for what is often an 8 or so minute game (and there is a around another minute of drop screen/map selection as well.) I'd rather play more games faster rather then wait longer for better quality matches.

I'm a Tier 5 bum - but I really enjoy this game. I'd love to see new maps, but I have a hard time imagining what brings someone back if they aren't enjoying the game fundamentally.

Clearly some people haven't liked what PGI has offered for a long time. I don't think most of the changes suggested would bring a plurality of the players who have left back. For example I don't think volumetric rescaling so some mechs get better really brings people back unless the only reason they left was an inability to play a specific beloved mech competitively.

I do think as a community we could talk about ways to adjust weighting in the matchmaker so it weights things that are more likely to result in better matches. I don't think we'll agree mind you - the current matchmaker that prioritizes group balancing on each side is the result of player requests on this forum. I think it doesn't work as well as PSR balancing IMO would as the primary matchmaker. For example a Tier 5 group pulls down the team, a tier 1 group pulls it up - treating those two very different groups as equivalent for the primary matchmaking criteria isn't going to work. On the other hand others feel that the matches are better matched by tonnage slot - would you put it above the group vs non group criteria in your personal optimized matchmaker?

[Horseman]

SirSmokes, on 19 November 2020 - 09:19 AM, said:

NO the match maker is not magic it can't make good matches out of a tiny player pool

But it can make better matches than it is now

GARION26, on 19 November 2020 - 10:03 AM, said:

On the other hand others feel that the matches are better matched by tonnage slot - would you put it above the group vs non group criteria in your personal optimized matchmaker?

I would put player quality as the first criteria. Groups are ultimately just a "block" of players that cannot be separately juggled between teams by the MM.
Tonnage is at best wonky as a metric, as mechs in the same tonnage range can have wildly different performance, so a different metric is needed to establish how much of a threat is the player x mech worth.
Taking last month of the player's performance in the chassis (or, if that's insufficient, global performance of similarly ranked players with the chassis) would provide more useful indication (with the aim being that the W/L, K/D and AMS of both teams are equal)

Edited by Horseman, 19 November 2020 - 10:20 AM.

[The pessimistic optimist]

Horseman, on 19 November 2020 - 10:07 AM, said:

But it can make better matches than it is now

I would put player quality as the first criteria. Groups are ultimately just a "block" of players that cannot be separately juggled between teams by the MM.
Tonnage is at best wonky as a metric, as mechs in the same tonnage range can have wildly different performance, so a different metric is needed to establish how much of a threat is the player x mech worth.
Taking last month of the player's performance in the chassis (or, if that's insufficient, global performance of similarly ranked players with the chassis) would provide more useful indication (with the aim being that the W/L, K/D and AMS of both teams are equal)

You can't make better matches with no player to make the matches better still not getting it

Edited by SirSmokes, 19 November 2020 - 11:32 AM.

[Gagis]

SirSmokes, on 19 November 2020 - 10:56 AM, said:

You can't make better matches with no player to make the matches better still not getting it

You can. It just needs really really accurate data to know which players really are the 2 best ones out of the 24 to put them in opposing teams and then go the list all the way down to the worst 2 players.

That accurate data is something it doesn't actually have, since PSR is based on Match Score and will thus unavoidably under-estimate just how much better than average the best players are and just how much worse than average the worst players are.

A matchmaking algorithm can do just fine with 24 completely randomly chosen players. The closer the matchmaker can make the chances for winning to 50-50, the better the algorithm is. This is why Win/Loss Record is also the best possible metric for ranking players, since it directly tells the matchmaker about a player's ability to shift the chances in their teams favour.

TIERS only make the job of a matchmaking algorithm easier by making sure the outliers above and below the curve are not too far away from each other.

There is absolutely no reason why teams in any given game would have to consist of 24 as IDENTICAL as possible players, as long as both teams have even distribution of different skill levels.

Edited by Gagis, 19 November 2020 - 12:18 PM.

[The pessimistic optimist]

Gagis, on 19 November 2020 - 12:17 PM, said:

You can. It just needs really really accurate data to know which players really are the 2 best ones out of the 24 to put them in opposing teams and then go the list all the way down to the worst 2 players.

That accurate data is something it doesn't actually have, since PSR is based on Match Score and will thus unavoidably under-estimate just how much better than average the best players are and just how much worse than average the worst players are.

A matchmaking algorithm can do just fine with 24 completely randomly chosen players. The closer the matchmaker can make the chances for winning to 50-50, the better the algorithm is. This is why Win/Loss Record is also the best possible metric for ranking players, since it directly tells the matchmaker about a player's ability to shift the chances in their teams favour.

TIERS only make the job of a matchmaking algorithm easier by making sure the outliers above and below the curve are not too far away from each other.

There is absolutely no reason why teams in any given game would have to consist of 24 as IDENTICAL as possible players, as long as both teams have even distribution of different skill levels.

*sigh* Nah they need to except reality and go own to 8v8