200 replies to this topic

[Horseman]

DAEDALOS513, on 25 April 2020 - 08:18 PM, said:

I feel like we're goin in circles..

No, you are going in circles.

DAEDALOS513, on 25 April 2020 - 06:44 PM, said:

All i'm saying is MM has a complicated job to do and you need more than just the 24 players in queue as was stated earlier by a few guys.. MM can not do it's job with the current population numbers..

The current matchmaker could not do its' job with double the population numbers.
Arguably, due to MASSIVE flaws in how the matchmaking works (hello, playtime-based?), it could never do its' job in the first place, the gaping hole in design just took a few years to show just how bad it is. Too many potatoes floated up to Tier 1 by sheer amount of matches played.

Yes, matchmaker systems work better with more players in queue.
But that doesn't mean that - as you insist - the quality of matchmaking cannot be improved in any other way.
A better matchmaking algorithm would not just fill out the teams with players from broadly compatible tiers (as the current one does), but also attempt to equalize the teams' expected performance against each other.
And it would not need more than 24 players to do so.

[Nearly Dead]

Since it sounds like the options are change or die, it might be that doing a couple of things might help it work. This is a bit of a rehash of stuff I had posted earlier but I hope to express it a little better.

Matchmaker needs to look at some key metric from the players last 100 games. I propose measuring (Damage / Tonnage) in each match. Period. Damage done, divided by mech weight. So I need to do 300 damage in my Kit Fox and an assault pilot needs to average 1000 damage in an Atlas to be in the top 10% or Tier 1. Tier 2 would be that score times 0.9 etc. Once you see the distribution they would need to adjust to get a good distribution of skills. Maybe the 10 per ton is too low, as a potato I have done more but have done a lot less and I don't know what my average is.

I still think that having a selection button to opt out and wait for a solo only match would provide a haven for casual players who find the experience overwhelming or unsatisfying. I think this would have value on its own and also as a warning flag for PGI to say, "Hey 70% of solo players are opting out of combined group/solo matchmaking, what's wrong?"

I still have real reservations about building a 12 player "team" that is actually a 4 or 8 player team plus solo filler. Player experience isn't just winning and losing it is also about contributing and sharing in the win. If existing teams come in, communicate in discord or some other outside comm channel and just ignore the solo players then those solo players will leave and even if they don't they aren't going to improve their game unless they are so astute that they can figure it out despite the fact that no one talking to them.

I could see the best new A players coming in and getting snatched up into units, B and C players being overlooked and underserved and leaving after a while because they know enough to know something is missing, leaving D players just running around shooting rocks, each other and sometimes a red mech. Pretty much the definition of a death spiral.

[Davegt27]

I will drop solo in group

but I want a guaranteed pay out
like 1 mil C-bills per drop and 500 XP and 500 GXP per drop

[Anomalocaris]

Davegt27, on 26 April 2020 - 07:46 AM, said:

I will drop solo in group

but I want a guaranteed pay out
like 1 mil C-bills per drop and 500 XP and 500 GXP per drop

How about 1 GSP for playing, 2 if you win (gotta incentivize people to actually try)? And nothing if you disco or afk. But I like how you think

[DAEDALOS513]

justcallme A S H, on 26 April 2020 - 01:23 AM, said:

Not really. I am explaining the flaws in what you are saying and explaining the reasons why.

You are just ignoring them and sticking to your point. A point that is factually incorrect.

It is incorrect because it is absolutely possible to balance matches better, right here / right now, in MWO. It just requires a bit more 'smarts'.

I could say the same about my points being ignored.. and around and around we go.. it's called disagreeing and it's ok..

There are too many variables to consider and you guys are living in la-la land if you think all you need is 24 players to create even a reasonable match-up. You're hanging on a hope and a dream for MM to work wonders when it's impossible because of lack of player base.. I've seen this problem in other games.. players whine that their mm needs to be tweaked and reworked when the problem of lack of player base must be solved first and foremost. It's a catch-22..

Edited by DAEDALOS513, 26 April 2020 - 08:53 AM.

[DAEDALOS513]

Horseman, on 26 April 2020 - 05:39 AM, said:

The current matchmaker could not do its' job with double the population numbers.
Arguably, due to MASSIVE flaws in how the matchmaking works (hello, playtime-based?), it could never do its' job in the first place, the gaping hole in design just took a few years to show just how bad it is. Too many potatoes floated up to Tier 1 by sheer amount of matches played.

Yes, matchmaker systems work better with more players in queue.
But that doesn't mean that - as you insist - the quality of matchmaking cannot be improved in any other way.
A better matchmaking algorithm would not just fill out the teams with players from broadly compatible tiers (as the current one does), but also attempt to equalize the teams' expected performance against each other.
And it would not need more than 24 players to do so.

Back when we had double the population we also had many more game modes going at once.. Solaris, Scouting, CW, QP and Group queue.. so even with the 'double' population, MM still struggled because of how the population was so divided..

Too many buckets was one of the nails in the coffin for this game... when the egomaniacs in the game won out and Solaris was born.. it was at the cost of Clan Warfare, the heart of the game. All attention should have been placed on CW before more buckets were created. This was the pivotal turning point in the game.. dev's took a wrong turn and a great game started to die.

Edited by DAEDALOS513, 26 April 2020 - 08:41 AM.

[Anomalocaris]

DAEDALOS513, on 26 April 2020 - 08:16 AM, said:

I could say the same about my points being ignored.. and around and around we go.. it's called disagreeing and it's ok..

There are too many variables to consider and you guys are living in la-la land if you think all you need is 24 players to create even a reasonable match-up. You're hanging on a hope and a dream for MM to work wonders when it's impossible because of lack of player base.. I've seen this problem in other games.. players whine that their mm needs to be tweaked and reworked when the problem of lack of player base must be solved first and foremost. It's a catch-22..

Let's try a different approach then. Post a screenshot of the end of game team stats of any solo queue match you've played in the last 30 days. Preferably a one sided stomp. And I'll show you how simply rearranging the players in that match would have given a much better chance of a close match. I'll also keep tonnage and mech classes as close as possible between the two teams. I'll simply rearrange by avg match score of the last 5 seasons on Jarl's.

That way we're not just throwing conjecture around, we'll actually demonstrate the inadequacies of the current MM and how it can be better - with the existing player base.

edit - let's make it even more interesting. All I want to see is the player name and mech columns for starters. Don't want to have any possibility of bias based on their match performance. Oh and please select a match without any AFKs.

Edited by Anomalocaris, 26 April 2020 - 10:03 AM.

[Bud Crue]

DAEDALOS513, on 26 April 2020 - 08:24 AM, said:

when the egomaniacs in the game won out and Solaris was born.. it was at the cost of Clan Warfare, the heart of the game. All attention should have been placed on CW before more buckets were created. This was the pivotal turning point in the game.. dev's took a wrong turn and a great game started to die.

I think I get what you are saying here, but it is a bit of a a simplification. The game has been shedding population for as long as they have given us the data to track it.
So looking at global stats:

https://leaderboard.isengrim.org/stats

That peak in the middle is the month leading up to and just after Solaris was introduced (Solaris patch was April 17, 2018). Solaris brought a lot of folks back but within three months the game had returned to the status quo of nominative loss rate with monthly ups and downs. Its also kinda interesting that the population spike of December 2016, which corresponds to the release of Faction Play phase 4.1 was actually less of a spike both in absolute terms and relative to the average population loss despite the game having had a much greater population back then. Point being, that this game has been shedding players at a decent clip way before Solaris. The short term gains and losses in population that corresponded to Solaris, were just far more dramatic than others. Solaris wasn't a "pivotal turning point" for the game, but just a more dramatic example in a long line of examples,of PGI not understanding what their player base wants in a mode.

[DAEDALOS513]

Anomalocaris, on 26 April 2020 - 09:51 AM, said:

Let's try a different approach then. Post a screenshot of the end of game team stats of any solo queue match you've played in the last 30 days. Preferably a one sided stomp. And I'll show you how simply rearranging the players in that match would have given a much better chance of a close match. I'll also keep tonnage and mech classes as close as possible between the two teams. I'll simply rearrange by avg match score of the last 5 seasons on Jarl's.

That way we're not just throwing conjecture around, we'll actually demonstrate the inadequacies of the current MM and how it can be better - with the existing player base.

..even if you're right.. you don't think there will be occasions when even in a perfectly matched game, there won't be stomp every now and then?

[Anomalocaris]

DAEDALOS513, on 26 April 2020 - 10:12 AM, said:

..even if you're right.. you don't think there will be occasions when even in a perfectly matched game, there won't be stomp every now and then?

I never said that. I don't think anyone in this thread has even come close to asserting that. Those of us that say we need a better MM and it can be done regardless of population size are saying exactly what we said. We can do a better MM with existing stats and population. A much better MM. Don't let the perfect be the enemy of the good.

An AMS, or WLR, or some combo of stats addition to the MM is low hanging fruit and will make everyone's solo queue experience better. Let's get it done. Then we can move on to other improvements. We can't go anywhere if we don't start taking steps.

Edited by Anomalocaris, 26 April 2020 - 10:21 AM.

[DAEDALOS513]

Bud Crue, on 26 April 2020 - 10:05 AM, said:

I think I get what you are saying here, but it is a bit of a a simplification. The game has been shedding population for as long as they have given us the data to track it.
So looking at global stats:

https://leaderboard.isengrim.org/stats

That peak in the middle is the month leading up to and just after Solaris was introduced (Solaris patch was April 17, 2018). Solaris brought a lot of folks back but within three months the game had returned to the status quo of nominative loss rate with monthly ups and downs. Its also kinda interesting that the population spike of December 2016, which corresponds to the release of Faction Play phase 4.1 was actually less of a spike both in absolute terms and relative to the average population loss despite the game having had a much greater population back then. Point being, that this game has been shedding players at a decent clip way before Solaris. The short term gains and losses in population that corresponded to Solaris, were just far more dramatic than others. Solaris wasn't a "pivotal turning point" for the game, but just a more dramatic example in a long line of examples,of PGI not understanding what their player base wants in a mode.

Yes and those stats make sense.. Solaris was a dead end game mode.. it didn't provide the players the opportunity to progress and grow.. it only provided a system to compare each others phallus size, and it did that poorly because the only way to truly do that is with two pilots bringing in the same build and mech. Had they focused on CW, players would have stayed with the game for the long term because it would have provided a chance for growth.. personal growth as well as growth for the unit you belonged to.

I say that solaris was the main pivotal deciding point because unlike other failed and disappointing patches, Solaris was extremely costly for PGI to implement. It took PGI very long to implement it, so much money and resources, that afterward they were depleted.. they didn't have the resources to continue work on CW because Solaris didn't provide the long term income they thought it would. Also, players waited so damn long for Solaris that when they saw it added no long-term 'fun' factor to the game they gave up faith in PGI (more than usual, because of how long they waited). No way were they gonna wait another year for the 'update that would save the game.'

After the failed Solaris release, this is when I think PGI decided that this game truly had an end-date... they decided to focus most of their resources on mech paks and keep the money train goin for as long as they could before the players woke up and realized that PGI was giving up on our beloved game.

Edited by DAEDALOS513, 26 April 2020 - 10:35 AM.

[DAEDALOS513]

Anomalocaris, on 26 April 2020 - 10:20 AM, said:

I never said that. I don't think anyone in this thread has even come close to asserting that. Those of us that say we need a better MM and it can be done regardless of population size are saying exactly what we said. We can do a better MM with existing stats and population. A much better MM. Don't let the perfect be the enemy of the good.

An AMS, or WLR, or some combo of stats addition to the MM is low hanging fruit and will make everyone's solo queue experience better. Let's get it done. Then we can move on to other improvements. We can't go anywhere if we don't start taking steps.

I say that because not every match is a stomp.. I'd say one out of 5.. but when they DO happen, they stand out more in your mind because of the aggravation factor.. so psychologically, you perceive it to happen more often than not and so MM may be working better than you realize..

I dunno.. imho, if we can't even get two 12 mans going in group queue, I don't think there is enough population for MM to work..

I agree with the low hanging fruit idea.. they should at least add some easy metrics and see how it goes. At the very least it will stir more interest in the game and increase population for a bit..

Edited by DAEDALOS513, 26 April 2020 - 10:37 AM.

[Paul Meyers DEST]

It has been said a gazillion times, a win W/L based MM is far better.
MM thinks ASH and i are at the same level and it is just not true. Period.

[Anomalocaris]

DAEDALOS513, on 26 April 2020 - 10:31 AM, said:

I say that because not every match is a stomp.. I'd say one out of 5.. but when they DO happen, they stand out more in your mind because of the aggravation factor.. so psychologically, you perceive it to happen more often than not and so MM may be working better than you realize..

I dunno.. imho, if we can't even get two 12 mans going in group queue, I don't think there is enough population for MM to work..

I agree with the low hanging fruit idea.. they should at least add some easy metrics and see how it goes. At the very least it will stir more interest in the game and increase population for a bit..

Well, participate in my experiment and I'll see if I can convince you....

Regarding stomps, that's not the biggest issue bothering me. I'm a decent player. Not elite or worldclass, but I get my dmg and kills and generally win more than I lose. But last week over a 3 day period, I went 2-19. Yeah, it was really, really bad. I'd have games where I was getting 4 kills, 2 solo and 700-800 dmg and still losing 12-9, 12-8, etc. Not stomps, but very disheartening when you're playing close to your max ability and still can't pull out wins. And it was because I had teams where 6-7 players weren't even able to break 200 dmg. I think a better MM will help even out streaks like that.

[justcallme A S H]

DAEDALOS513, on 26 April 2020 - 10:12 AM, said:

..even if you're right.. you don't think there will be occasions when even in a perfectly matched game, there won't be stomp every now and then?

Of course there will be.

However that is not the point a many number of people have tried to point out to you so far.

The point is simple:

Match quality - if the MM actually balanced players by actual skill and not "PSR" - would improve.


Is it perfect? Of course not.
Is that the point? Not at all.

Read and re-read till it makes sense.

[Bud Crue]

DAEDALOS513, on 26 April 2020 - 10:20 AM, said:

Yes and those stats make sense.. Solaris was a dead end game mode.. it didn't provide the players the opportunity to progress and grow.. it only provided a system to compare each others phallus size, and it did that poorly because the only way to truly do that is with two pilots bringing in the same build and mech. Had they focused on CW, players would have stayed with the game for the long term because it would have provided a chance for growth.. personal growth as well as growth for the unit you belonged to.

I say that solaris was the main pivotal deciding point because unlike other failed and disappointing patches, Solaris was extremely costly for PGI to implement. It took PGI very long to implement it, so much money and resources, that afterward they were depleted.. they didn't have the resources to continue work on CW because Solaris didn't provide the long term income they thought it would. Also, players waited so damn long for Solaris that when they saw it added no long-term 'fun' factor to the game they gave up faith in PGI (more than usual, because of how long they waited). No way were they gonna wait another year for the 'update that would save the game.'

After the failed Solaris release, this is when I think PGI decided that this game truly had an end-date... they decided to focus most of their resources on mech paks and keep the money train goin for as long as they could before the players woke up and realized that PGI was giving up on our beloved game.

Your premise, that if PGI threw their Solaris efforts into CW (rather than Solaris) players would have stayed with the game long term, is pure conjecture that ignores several factors, namely:

-The previous three years of the game, wherein PGI updated CW, more than four times (phases1-4.1) and yet players only flocked to CW at the beginning of every phase, then quickly fled as the good teams consistently dominated just like the mode is intended to encourage, and as always, the bads and casuals fled. This happened everytime. And throughout each progression the player loss continued unabated. There was never any potential for "long term" "player growth" from day one in the CW mode, as PGI provided it, and that has never changed.

- Meanwhile according to PGI (back during Tukayyid 2 at the height of Pre-Solaris CW popularity) no more than something like 17% of the overall population played CW. As an aside, I would bet that right now -post Solaris- more of the population as a percentage is playing CW than then (though this last week with 8v8 GQ that might no longer be the case).

-For years, a fairly vocal group of the player base was asking and begging for Solaris. Perhaps not the Solaris PGI gave them, but Solaris none the less. Damn near every developer update, roadmap, "townhall" or whatever Russ and/or Paul featured NGNG stream they did from 2015-2018 had someone asking about "when do we get a Solaris mode". If PGI would have come right out way back when and said "No. we will not be wasting any development efforts on this mode you keep asking for. All you get is CW, which though only 17% of you are willing to play even during a reward padded special event, we will continue to develop only that mode in addition to quick play." How do you think that would have played with the player base?

Anyway, I think the history and the data speak for themselves here. Solaris was indeed a mistake, but CW was never the popular and long term draw you are suggesting that it was, or could potentially have been, given the level of effort PGI historically puts into, well, anything.

Edited by Bud Crue, 26 April 2020 - 02:24 PM.

[Nearly Dead]

"Perfect is the enemy of good enough".

I will play whatever hand PGI deals unless it makes things bad enough that I can't take it. I will say that in the meantime, I am not spending cash beyond what I already have unless they show a good faith effort to make this change positive. I will probably buy more mechs for C-bills, I bought a ACW yesterday in fact, but no more money unless this change at least makes some kind of sense.

They need to improve player tier ranking. It is absurd that my tier goes up if my team wins despite carrying my 30 ton *** and goes down if I do top damage and team score in a loss in the same mech.

They need to monitor how all the players are reacting to the change and be prepared to address concerns realistically. No one knows if it is all going to come up daisies and unicorns or be a big youknowwhat sandwich.

My hope is that they are really trying to improve the player experience. My fear is they are pulling a WKRP and looking for an excuse to pull the plug.

[BALIander]

justcallme A S H, on 25 April 2020 - 06:35 PM, said:

BALIander, on 26 April 2020 - 02:59 AM, said:

Oh nice, you also see the need and benefit of only allowing max. 4 players to play together for sure.
- no clubbing if your other 4 mates that also wait in cue come out in the opposing team, ;-)[/color]


justcallme A S H, on 26 April 2020
Realistically that'd going to do nothing. A limit of 4 down from 8 or 12 makes no actual difference because the less mechs you have the more tonnage available.

thats not what I said if you read the full thing - Its still 3 lances vs. 3 lances (12 players vs. 12 players). ;-)

Edited by BALIander, 26 April 2020 - 03:06 PM.

[DAEDALOS513]

Bud Crue, on 26 April 2020 - 02:23 PM, said:

Your premise, that if PGI threw their Solaris efforts into CW (rather than Solaris) players would have stayed with the game long term, is pure conjecture that ignores several factors, namely:

-The previous three years of the game, wherein PGI updated CW, more than four times (phases1-4.1) and yet players only flocked to CW at the beginning of every phase, then quickly fled as the good teams consistently dominated just like the mode is intended to encourage, and as always, the bads and casuals fled. This happened everytime. And throughout each progression the player loss continued unabated. There was never any potential for "long term" "player growth" from day one in the CW mode, as PGI provided it, and that has never changed.

- Meanwhile according to PGI (back during Tukayyid 2 at the height of Pre-Solaris CW popularity) no more than something like 17% of the overall population played CW. As an aside, I would bet that right now -post Solaris- more of the population as a percentage is playing CW than then (though this last week with 8v8 GQ that might no longer be the case).

-For years, a fairly vocal group of the player base was asking and begging for Solaris. Perhaps not the Solaris PGI gave them, but Solaris none the less. Damn near every developer update, roadmap, "townhall" or whatever Russ and/or Paul featured NGNG stream they did from 2015-2018 had someone asking about "when do we get a Solaris mode". If PGI would have come right out way back when and said "No. we will not be wasting any development efforts on this mode you keep asking for. All you get is CW, which though only 17% of you are willing to play even during a reward padded special event, we will continue to develop only that mode in addition to quick play." How do you think that would have played with the player base?

Anyway, I think the history and the data speak for themselves here. Solaris was indeed a mistake, but CW was never the popular and long term draw you are suggesting that it was, or could potentially have been, given the level of effort PGI historically puts into, well, anything.

NONE of the CW updates ever fulfilled the original purpose of Faction Warfare.. that was to give real purpose to being in a unit and wearing a faction's flag. THIS was key in locking down their playerbase and PGi failed to do it. The vocal Solaris group you speak of was completely overshadowed by the massive CW group that was loyally standing in the wings while PGI kept lying through their teeth, postponing and breaking promises about finishing CW.. CW was the reason most players were sticking it out with the game.. certainly not Solaris. Even if there was a sizable group that wanted Solaris, PGI should have prioritized CW and finished it before beginning a new project. Any potential player loss due to not implementing Solaris and focusing on Faction Warfare would have been miniscule compared to the losses they experienced for not finishing C-dub. On top of all this, finishing Faction Warfare would have brought in new players because units would be clamoring for members.. Solaris doesn't have that same benefit.

BTW, my opinion isn't 'pure conjecture', it's educated opinion. I've been with the game almost since the beginning (less a few months), in that time I have stuck it out with PGI's good and bad decision making... I've dropped around 50,000 matches and own over 460 elited mechs.. during most of that time I also created and led one of the top units in the game. When I speak it's coming from experience.. and despite what you might be thinking.. i'm not trying to brag.. i'm trying to put some weight behind my words by showing you what a big MWO fan I am.

Edited by DAEDALOS513, 26 April 2020 - 06:56 PM.

[General Solo]

This is kinda for Paul, hope your reading man

TLDR I:Side effect of PSR's upward bias is that it reduces the granularity of the skill measure (PSR).
Given infinite time the majority would be in Tier one
So you get this huge group of Tier 1's

And four much smaller groups of Tier 2, 3 ,4 and 5 who even with time could not get carried as much as they lost.
So functionally their are two groups

Tier 1 - Consisting of "Carry people", "Competent people" and people that can be "Carried pretty ez" and

Permanent <Tier 1 - Consisting of people that are difficult to carry and so due to loses lose PSR more often than gain PSR over the long term and so never climb to reach Tier 1.
As you only lose PSR when you lose and can't when your team wins.

Theirs not enough granularity of the skill measure to make a balanced team, as their is the additional three groups within Tier 1, the carriers, the competant and the carried which at present match maker can not filter to make good balanced teams because all three of those groups are Tier 1 and match maker cannot distinguish between them.

This is due to the upward bias over time of the cannot go down on a win despite playing poorly, rewarding play time with PSR (which is fair enough but not this way) and funnily enough using that data as input into a matchmaker thats supposed to make balance games flawed logic.
Just give them a title or badge, rather than undermine the matchmaker and ultimately your profits.

The key to fixing this is to treat Match Score the same win or lose. This would reduce the "got carried effect"

You may think a measure of skill other than match score is better and maybe your right.

However Im thinking of the cheapest minimal effort fix that will work good enough, and that would be adjusting the existing system. (May have to hire a whizz short term, but worth it imo imagine if the pop started to grow)

The flaw and also fix is pretty simple
The flawed assumption that play time is correlated with player skill is the reason for our match making situation. (MS while not perfect is better)

Players have been rewarded PSR for getting carried due to the inability to lose PSR even when they perform very poorly on a win, and they only won because their team mates won the game for them aka got carried.

Hardly an indicator of individual player skill, what you say? Yet rewarded PSR supposed to measure pilot skill and supposed to be used by a matchmaker to that effect.

So as Ash said, garbage in garbage out, the match maker can only do so much with faulty input data.
Match maker wasn't designed for that.

As the skill measure PSR is derived from MatchScore, PSR is corrupted when a players PSR is inflated due to winning despite a low matchscore due to current logic used.

500 MS is 500 MS of skill win, lose or draw, same in all circumstances, an amazingly good pilot - Currently rewarded

50 MS is 50 MS of skill win, lose or draw, same in all circumstances, a poorer pilot - When loses demoted, wrongly rewarded on a win, should be demoted as on a loss.

Same thing should happen in all circumstances win, lose or draw as this match maker uses match score to measure "individual player skill"

Winning is the result of the TEAM, sure players contribute but TEAMS WIN!
No player 12v1 360 no scoped a win in this game with balanced teams.

TLDR II: By treating PSR to matchscore movement differently for wins and loses the individual skill measure PSR is coruputed wid chaos team contributed data for being carried to a win which must be PURGED! for the EMPpra!

The righteous MatchScore Data must be followed instead. iamirite? Cause Matchscore is how a pilots INDIVIDUAL skill is measured in this game. Iamirite!

Otherwise they got carried and mm shouldn't be measuring that as a skill. Also Carriers must be differentiated from competent pilots to balance teams guid. I play the game a lot long time, this must be done, please.

So the solution is pretty easy but worth it in my opinion.
Make match score get the same increase or decrease in PSR regardless of win or lose. The current loss conditions will do, make it the same for wins, done.
Reset not necessary as MM will fix itself over time but a reset would make the fix instant.

May have to hire a whizz to find the code if you don;t know where it is, but once found its only a few simple edits .


QA it and monitor the results and compare with your current baselines

Edited by OZHomerOZ, 27 April 2020 - 01:47 AM.