104 replies to this topic

[LordNothing]

problem with psr system is that more than the top 1/5 of the population live in t1. in theory each tier should represent 1/5th of the population so that actual skill is at the top and not people who play a lot of yolo games and just trickle up. instead use the stats (empirical data) to base tiers off of with each tier maintaining 1/5 of the active population. if its good enough for shaming why isnt it used for the match maker? you would be giving everyone better games.

of course the real puprpose of the tier system is to shield new players from the tryhard horde. to get them to buy a few mech packs before yanking the rug from under them.

Edited by LordNothing, 18 April 2019 - 02:39 PM.

[Nightbird]

Same topic gets brought up every few months, falls on deaf ears, and at this point in the life cycle it's just not worth the effort to implement a MM.

[Bloodwitch]

LordNothing, on 18 April 2019 - 02:37 PM, said:

problem with psr system is that more than the top 1/5 of the population live in t1. in theory each tier should represent 1/5th of the population so that actual skill is at the top and not people who play a lot of yolo games and just trickle up. instead use the stats (empirical data) to base tiers off of with each tier maintaining 1/5 of the active population. if its good enough for shaming why isnt it used for the match maker? you would be giving everyone better games.

Most certainly not. tier system should be based on actual skill and performance.
The problem with the current system is a.) low playerbase and b.) upward bias.
a tier system based on population percentage is nonsense. Calling the upper 1/5th of the polulation does nothing good since the really good players will totally declassify the lower end (this goes for every tier then).

What would help the matchmaking in general is
1.) tier reset
2.) fixing upward bias
3.) some time and games to ajust everyone into the tier they belong

this will come with a lot longer waiting times for games, tho.

[Daggett]

LordNothing, on 18 April 2019 - 11:09 AM, said:

then that puts it directly into the situational category, which is just as bad. without team support to set up that situation you might as well just throw your mech away.

As i said, your concerns are correct in theory, but reality looks different:



I bought the Bushwacker X1 for it's perfect triple RAC2 hardpoints and played it exclusively with that configuration. So those stats are pure solo QP, RAC2 only without a single exception. Does this look like a situational build or that i throw my mech away regularly?

For comparison you see my laser Hellbringer. I'm not able to reach the same stats with this heavier meta mech.

So for me the triple RAC2 Bushwacker is the MOST reliable and consistent build i have ever played. It never disappoints regardless how potato my team is.

RAC2s are way better that they may look on paper since their few weaknesses can easily be worked around with proper positioning and aggression.

Edited by Daggett, 18 April 2019 - 03:42 PM.

[LordNothing]

Kunato Developments, on 18 April 2019 - 03:22 PM, said:

Most certainly not. tier system should be based on actual skill and performance.
The problem with the current system is a.) low playerbase and b.) upward bias.
a tier system based on population percentage is nonsense. Calling the upper 1/5th of the polulation does nothing good since the really good players will totally declassify the lower end (this goes for every tier then).

What would help the matchmaking in general is
1.) tier reset
2.) fixing upward bias
3.) some time and games to ajust everyone into the tier they belong

this will come with a lot longer waiting times for games, tho.

sure but it seems easier to get rid of the current psr formula and just use the average match score over the last 100 or so games. top 1/5 average match scores are t1, the next 1/5 is t2 and so on. you also arent locked into a high tier if you get worse or have to level crappy mechs.

[Tarl Cabot]

Scout Derek, on 18 April 2019 - 10:38 AM, said:

I mean, he did say he was average. he never claimed he was a good player. Bad Stat Shame is bad.

He's complaining about MM, which is a legitimate reason to be annoyed with as the matching of players is absolutely garbage.

OP, do you recall when you hit Tier 2? Was it last month after playing almost 350 games, or was it sometime last year around Sept 2018 or so?

Stats show he is averaging approx 200 MS games w/near a 1-1 W/L (0.95) as of last eason/month (he played lots of games) overall 0.86 W/L AND he is in tier 2......

There are other players who have played longer and/or played more games with sub-190/180 MS and who are in Tier 1 and 2. And that is not just the population but due to who PSR is calculated. it has static thresholds and has a upward push to it, ie participation trophies for many. And even with the current setup has PGI has not considered reseeding the players on a regular basis since this went live several years ago, when it was initially seeded based on a player's Elo score. Nor does PGI MM utilize any other stats other than Tier level.

But it is not stat shaming him, it is showing the affect of PSR (remember- static thresholds). He is primarily running heavies/assaults and now in a tier where mistakes are more likely to be capitalized on than not. Stomps may likely happen more often T1/2 matches then in the lower tiers, higher tiers players are more likely attempt to capitalize on mistakes, and the side to lose a few mechs early then it is downhill from there.

Should the OP be in Tier 2 or I in Tier 1? No to both. I should be in Tier 2, and bouncing around in it. The OP likely should be in Tier 3 with the current setup. Tier 1 I am likely on the low end of T1's bell curve, the OP the same for T2 bell curve. My alt that I have recently been playing just reached T2. T3 I was definitely on the high end of the bell curve, avg MS 75 higher than my main, with around 350 QP games on that alt.

Nor do we know what type of builds he runs though, IS or Clan mechs, etc. And he may have older stats but may have changed his name sometime early 2018 or 2017.

And population is only part of the issue. If we had more players a number of people who are currently in Tier 1-2 may have taken longer or gotten there faster but it would not have changed the PSR static thresholds. On a win a player can move up with a 101 MS, or no change with a 100 MS or less. On a loss 250 MS to have no movement. Breaking even on win/loss (it is a team game), even with a higher population, would that mean that a Tier 3 seeded drop would have fewer or no Tier 1 players in it? Less likely to get turned into a 12-0 stomp but into a 12-4 or 12-5 game, so more players on the losing side have a higher match score, most players on the winning side have a higher and more evenly spread match score?

Crap, where was I going with this.. I need a drink..!!!

[GeminiWolf]

Bushrat, on 18 April 2019 - 12:50 AM, said:

OK, there are several threads that touch on this, but these are my frustrations.

Match maker needs a revamp.... PGI please put some resources on this that will improve gameplay for average player like myself. Many nights I go without a win in 20 + matches, sometimes I'll get on a roll and win several ina row. I don't think my sole efforts have much to do with it. It can't be my fault every time on a 12-0, 12-1, or 12-2 match. Nor can it be my fault on astomp the other way, I think ELO is not a good way to track players skill,,, K/D or, Match score or some combo of those factors,, but ecpereince is not enogh to balance a good match. It make me very discouraged to play the game.

RAC is OP....ad some heat, more jams, or camera shake or smoke to block vision or something to balance it out.

Lights, Lights in Stealth are slightly OP. A good light pilot should be able to do great things, but they seem to be more effective in the hands of a good pilot than an assault is with an equally skilled pilot.

I've joined a great unit, I pay with groups, I try to contibute and play wisely. I try to build effective META builds and successful build for my play style. What else can I do?

PGI please listen to me.... I have spent THOUSANDS of dollars on this game. Don't discourage me from playing, I'll be less likely to spend money on the game if these frustrations continue.

The Devoted,

Bushrat

Again PGI can't control the players play

[Bloodwitch]

LordNothing, on 18 April 2019 - 03:41 PM, said:

sure but it seems easier to get rid of the current psr formula and just use the average match score over the last 100 or so games. top 1/5 average match scores are t1, the next 1/5 is t2 and so on. you also arent locked into a high tier if you get worse or have to level crappy mechs.

I don't think it's a good system but at least there is a mechanic to fall in tiers.
Why not have the psr formular adjusted, tho?
The problem is that everyone climbs regardless of performance (some faster then others but still).



Question is: are we willing to wait longer for more balanced games?

LordNothing, on 18 April 2019 - 03:41 PM, said:

Crap, where was I going with this.. I need a drink..!!!

Call me stupid but i don't think there is much difference between tier 1 and 2. It's literally the same playerbase.
Yes, you occasionally get a tier 4 match as tier 2 but more often tier 1-3.

i've had this for another topic not to long ago.
tier 2
1: 2-3 mix
2: 1-2 mix
3: 1-2 blue
4: 2-3 mix
5: 2-3 blue
6: 1-2 mix
7: 2 (no info)
8: 1-4 mix
9: 1-2 mix
10: 1-2 blue / red no info (blue 12:0)
11: 1-2 mix
12: 1-4 mix (red 12:11)
13: 1-2 mix
14: 1-2 mix (red 12:0)
15: 1-2 mix
16: 1-2 mix
17: 1-2 mix
18: 1-2 mix
19: 1-3 blue
20: 1-4 (4blue)

Edited by Kunato Developments, 18 April 2019 - 05:58 PM.

[Variant1]

Meta builds should not be encouraged. The mechwarrior mechlab is all about building the mech to suit your playstyle not something the pros use. Otherwise whats the point of running any build if that one build will simply do better?

Scout Derek, on 18 April 2019 - 10:38 AM, said:

I mean, he did say he was average. he never claimed he was a good player. Bad Stat Shame is bad.

He's complaining about MM, which is a legitimate reason to be annoyed with as the matching of players is absolutely garbage.

#remove jarls list

[FupDup]

Variant1, on 18 April 2019 - 06:26 PM, said:

Meta builds should not be encouraged. The mechwarrior mechlab is all about building the mech to suit your playstyle not something the pros use.
#remove jarls list

What if someone's playstyle just so happens to coincide with what the meta is? And if someone is a pro themselves, then one of them using one of those "pro builds" isn't really surprising...

[Variant1]

FupDup, on 18 April 2019 - 06:29 PM, said:

What if someone's playstyle just so happens to coincide with what the meta is? And if someone is a pro themselves, then one of them using one of those "pro builds" isn't really surprising...

Coincidence perhaps but highly unlikely, not that their build should do bad, its just other builds should have a chance to shine ya know?

[LordNothing]

Kunato Developments, on 18 April 2019 - 05:51 PM, said:

Call me stupid but i don't think there is much difference between tier 1 and 2. It's literally the same playerbase.
Yes, you occasionally get a tier 4 match as tier 2 but more often tier 1-3.

i've had this for another topic not to long ago.
tier 2
1: 2-3 mix
2: 1-2 mix
3: 1-2 blue
4: 2-3 mix
5: 2-3 blue
6: 1-2 mix
7: 2 (no info)
8: 1-4 mix
9: 1-2 mix
10: 1-2 blue / red no info (blue 12:0)
11: 1-2 mix
12: 1-4 mix (red 12:11)
13: 1-2 mix
14: 1-2 mix (red 12:0)
15: 1-2 mix
16: 1-2 mix
17: 1-2 mix
18: 1-2 mix
19: 1-3 blue
20: 1-4 (4blue)

t2 seemed transitionary at best, after spending a very long time in t3, i shot up through t2 and was t1 pretty quick. if you look at my stats i barely belong in t3. its just by brute forcing through a lot of games i got there anyway.

[Y E O N N E]

Variant1, on 18 April 2019 - 06:26 PM, said:

Meta builds should not be encouraged. The mechwarrior mechlab is all about building the mech to suit your playstyle not something the pros use. Otherwise whats the point of running any build if that one build will simply do better?

That's not how things work.

There will always, always, always be some combination of things that is better no matter what you do. You cannot stamp it out.

[Bloodwitch]

LordNothing, on 18 April 2019 - 06:35 PM, said:

t2 seemed transitionary at best, after spending a very long time in t3, i shot up through t2 and was t1 pretty quick. if you look at my stats i barely belong in t3. its just by brute forcing through a lot of games i got there anyway.

I am switching between a tier 1 and a tier 2 account and i get the same players every game.

[LordNothing]

Kunato Developments, on 18 April 2019 - 08:34 PM, said:

I am switching between a tier 1 and a tier 2 account and i get the same players every game.

thats because the valves are always open and most of the players are in t1/2 anyway.

Edited by LordNothing, 18 April 2019 - 08:48 PM.

[Sjorpha]

Variant1, on 18 April 2019 - 06:26 PM, said:

Meta builds should not be encouraged. The mechwarrior mechlab is all about building the mech to suit your playstyle not something the pros use. Otherwise whats the point of running any build if that one build will simply do better?

I'm not sure what you mean by not encouraged. "meta builds" isn't any specific kind of build, it's just whatever builds has proven to work best and therefore becomes popular. You can't really avoid encouraging strong builds and mechs, as the game itself encourages it, and if you change what is strong you just change which builds are "meta". You also can't stop players from sharing information an builds with the community.

I mean I guess you could remove winning as the primary goal and encourage playing to lose? There would be a metagame and best builds for that goal as well though.

That's what it means in mwo at least, in other games it often refers to playing an otherwise weak strategy to beat a specific predicted metagame in a tournament. Like main decking rest in peace to beat a metagame with dredge and Phoenix in mtg and so on.

Edited by Sjorpha, 18 April 2019 - 10:32 PM.

[Prototelis]

Variant1, on 18 April 2019 - 06:26 PM, said:

Meta builds should not be encouraged. The mechwarrior mechlab is all about building the mech to suit your playstyle not something the pros use. Otherwise whats the point of running any build if that one build will simply do better?

#remove jarls list

Everyone should play bad mechs on purpose so I can still lose to them!

[PhoenixFire55]

Bushrat, on 18 April 2019 - 12:50 AM, said:

Many nights I go without a win in 20 + matches, sometimes I'll get on a roll and win several ina row. I don't think my sole efforts have much to do with it.

Bushrat Overall 0.86 W/L
Bushrat Mar '19 0.95 W/L
So it would seem that for every 20+ matches loss streak you do get an about equally long 20+ match win streak. Since you W/L is actually about 0.9 and not 0.05.

Bushrat, on 18 April 2019 - 12:50 AM, said:

It can't be my fault every time on a 12-0, 12-1, or 12-2 match. Nor can it be my fault on astomp the other way

Stats displayed above clearly state that over the course of 2042 games you've played you and you alone were the reason why 7% of them were losses rather than wins.

Bushrat, on 18 April 2019 - 12:50 AM, said:

I think ELO is not a good way to track players skill,,, K/D or, Match score or some combo of those factors,, but ecpereince is not enogh to balance a good match. It make me very discouraged to play the game.

First, its Elo, not ELO. Second, current MM doesn't use Elo. It does in fact use match score, kills, assists and other things.

Bushrat, on 18 April 2019 - 12:50 AM, said:

RAC is OP....

No its not.

Bushrat, on 18 April 2019 - 12:50 AM, said:

Lights, Lights in Stealth are slightly OP. A good light pilot should be able to do great things, but they seem to be more effective in the hands of a good pilot than an assault is with an equally skilled pilot.

Assaults are statistically the best performing class. Lights are statistically the worst performing class.

Bushrat, on 18 April 2019 - 12:50 AM, said:

I've joined a great unit, I pay with groups, I try to contibute and play wisely. I try to build effective META builds and successful build for my play style. What else can I do?

Improve rather than rant?

Bushrat, on 18 April 2019 - 12:50 AM, said:

PGI please listen to me.... I have spent THOUSANDS of dollars on this game. Don't discourage me from playing, I'll be less likely to spend money on the game if these frustrations continue.

As much as I admire your attempt to threaten them with the only thing they care about, sadly they still don't read their own forum. Much less read random rants with nothing backing up a multitude of "this OP, that OP" claims in it.

crazytimes, on 18 April 2019 - 01:33 AM, said:

There are way too few people for any kind of meaningful MM. I am just happy we get matches at least. Late last year it was a 3-4 minute wait for solo QP. That was not cool.

This is a very popular misconception. Math, specifically LoLN clearly shows that the size of the playerbase required for a near-perfectly balanced 12v12 match between solo players is 24. Not 24k. Just 24.

[PhoenixFire55]

Xeno Phalcon, on 18 April 2019 - 04:36 AM, said:

Matchmaker is absolute crap but with such a small player pool it cant really be helped

Yes it can. Read just above.

[PhoenixFire55]

Bushrat, on 18 April 2019 - 07:05 AM, said:

losing most battles can get frustrating, please don't dismiss this.

It is exactly like with riding a bus regularly. On a day the bus is on time you don't even notice it, yet days it is late you fume about it and think about how "this stupid bus is late every single time". Again, your stats clearly display that you win about as many games as you lose.

Scout Derek, on 18 April 2019 - 07:50 AM, said:

Seriously though. Population is too small to actually warrant a good MM fix. Don't expect it.

And again. False.

Feral Clown, on 18 April 2019 - 10:39 AM, said:

You lost me at rac op....they are garbage and only way you should be worried about them is if you accidentally find yourself 100m away from them which 99% of the time is pilot error.

Why would you be worried about them at 100m range if at 100m range your 4xSRM6 alpha will actually all go into his face since all he can do is stare at you while you roll the damage between alphas?

Scout Derek, on 18 April 2019 - 10:42 AM, said:

Not at just 100M.... around 500M roughly. You're still going to get a face full of lead either way.

The only way you get a face full of lead is if you keep starting at him with your face all the time. And the only reason to do that is if you got RACs of your own.

Scout Derek, on 18 April 2019 - 11:22 AM, said:

What is amusing that with 3 RAC/2 supported by a bigger weapon such as a gauss rifle you're spitting out 2.7 damage per hit backed by a bigger single projectile. really works well tbh.

Oh wow ... and how many mechs in game can actually field 3xRAC2s and a Gauss rifle on top of that?

Scout Derek, on 18 April 2019 - 11:43 AM, said:

Again the issue is player population; there isn't enough people of the same skill to be matched properly at any point in time. Maybe 3-4 years ago, but now? It's kind of wavering truth be told.

It doesn't need to be same skill. It needs to me even distribution of good/bad between teams.