48 replies to this topic

[Aeries]

Hello everyone. I have only been playing this game going on 5 or 6 months now and have always had my suspicion that MM in this game is non-existent. As a more or less average player that just got to T4 (only 10% in) I had a game yesterday that confirmed my suspicions.

Dropped into the game and sitting in lobby when someone posts their twitch stream. Don't know too many comp players, but I know who Sean Lang is, and yep there he is on my team. First impression was, hey that's cool, then it was, why am I even in a game with a comp player to begin with.

Now I know MM trys to make a similar tier match then if it can't in a reasonable amount of time it loosens its restrictions, but seriously comp players? Now I have always been of the opinion that facing better players makes you better. I want to fight against players that are better than me, its how you improve imo, but that is a bit much.

As a noob now I see why I was getting wiped endlessly and mercilessly till I learned the game a bit better. It also explains why we get the wipes that we still do with 0 - 12 steamrolls with 8 or 9 players on a side doing less than 100 damage, they probably got the luck of skilled players with literally MM only matching mech class.

Does this seriously work for most of you? I have read that there used to be a ELO system in place. Why was it done away with? Why not tier K/D ratio or KMDD/D ratio, would be better than the XP bar tier system we have now. Just trying to find the logic in this company's MM system.

[Lostdragon]

12-0 matches are a product of the game type as much as they are marchmaking. When you don't have respawns and you lose a teammate early or you have a disconnect that really hurts. Once you are down by three mechs the odds of coming back are extremely small in my experience. It is just a boils down to volume of fire and who effectively rotates the vanguard 99% of the time.

[STEF_]

Lostdragon, on 02 August 2017 - 04:31 AM, said:

12-0 matches are a product of the game type as much as they are marchmaking. When you don't have respawns and you lose a teammate early or you have a disconnect that really hurts. Once you are down by three mechs the odds of coming back are extremely small in my experience. It is just a boils down to volume of fire and who effectively rotates the vanguard 99% of the time.

12-0 can happen (and should happen).... but not 6-7 double digits.

[Lostdragon]

Stefka Kerensky, on 02 August 2017 - 04:40 AM, said:

12-0 can happen (and should happen).... but not 6-7 double digits.

Again, I think that is just a product of the no respawn mode as much as anything else and I don't think you will ever see it go away. Every player has a bad match now and then, just a few days ago I went up against one of thr best Locust players and I one shot him because he made a mistake. He did less than 10 damage, the last game I saw him in before that he did over 1k in his Locust.

All it takes is one or two guys out of position or making some kind of mistake against an aggressive team and things just snowball. MWO is not the type of game where you are going to see a lot of heroic comebacks.

[Aeries]

Lostdragon, on 02 August 2017 - 04:31 AM, said:

12-0 matches are a product of the game type as much as they are marchmaking. When you don't have respawns and you lose a teammate early or you have a disconnect that really hurts. Once you are down by three mechs the odds of coming back are extremely small in my experience. It is just a boils down to volume of fire and who effectively rotates the vanguard 99% of the time.

I definitely get that. You lose a couple of mechs early and it is hard to come back from and usually steamrolls from there. Been there, done that, got the t-shirt. Losses like that also happen if you lose the luck of the draw and get a team of new players that are still working things out vs guys that are posting up 3+ K/D averages on a regular basis.

I want to fight those guys, but spread them across the teams if the MM floodgates have to be opened. Still think a K/D or KMDD/D tier system and MM would work much better. Personally I like KMDD since you can core a guy 95% and still not get the kill. Could look something like this.

T5: 0 - 1 KMDD
T4: 1.1 - 1.9 KMDD
T3: 2 - 2.9 KMDD
T2: 3 - 3.9 KMDD
T1: 4+ KMDD

Try to match by tier then expand during low population times, but try to average the tier on each team.

Edit: Grammar

Edited by Aeries, 02 August 2017 - 04:54 AM.

[Lostdragon]

Aeries, on 02 August 2017 - 04:52 AM, said:

I definitely get that. You lose a couple of mechs early and it is hard to come back from and usually steamrolls from there. Been there, done that, got the t-shirt. Losses like that also happen if you lose the luck of the draw and get a team of new players that are still working things out vs guys that are posting up 3+ K/D averages on a regular basis.

I want to fight those guys, but spread them across the teams if the MM floodgates have to be opened. Still think a K/D or KMDD/D tier system and MM would work much better. Personally I like KMDD since you can core a guy 95% and still not get the kill. Could look something like this.

T5: 0 - 1 KMDD
T4: 1.1 - 1.9 KMDD
T3: 2 - 2.9 KMDD
T2: 3 - 3.9 KMDD
T1: 4+ KMDD

Try to match by tier then expand during low population times, but try to average the tier on each team.

Edit: Grammar

I just don't think there is a lot you are going to be able to accomplish with MM, it already basically works how you describe as far as matching tiers. The floodgates are open because populations are low. It doesn't matter if you change MM to work off a new formula, it won't solve that problem. You could have the best MM alogrithm in the world that accurately sorts players by skill with the highest precision possible, but if the number of top tier players is relatively low then you have to put them in matches with lower tier people to avoid wait times that make people want to quit.

It is kind of a no win situation for PGI. They could definitely improve the way tiers are assigned, but if there is not enough population of high tier players to segregate them from low tier players you wind up right back where we are now and they have wasted time and resources.

[STEF_]

Lostdragon, on 02 August 2017 - 04:51 AM, said:

Again, I think that is just a product of the no respawn mode as much as anything else and I don't think you will ever see it go away. Every player has a bad match now and then, just a few days ago I went up against one of thr best Locust players and I one shot him because he made a mistake. He did less than 10 damage, the last game I saw him in before that he did over 1k in his Locust.

All it takes is one or two guys out of position or making some kind of mistake against an aggressive team and things just snowball. MWO is not the type of game where you are going to see a lot of heroic comebacks.

Again, if in a team there are 5-6 double digits, then IT IS a MM issue.
Not one. Half a team shooting ONLY ONE TIME, it's simply reddicolous

[Inatu Elimor]

https://mwomercs.com...73#entry5814973

(goota is a typo )

(BTW Sean is a PGI spy)

(BTTWW Baradul recently posted on youtube showing some build and confessing loosing a couple of games in a row, not completely understanding why...it' s not you.)

Edited by Inatu Elimor, 02 August 2017 - 08:46 AM.

[Dread Render]

All I really want to hear about the MM is that PGI is on top of it and are doing the best they can.
...cause damn... this topic has come up many many times and a few times by me.

Stefka Kerensky, on 02 August 2017 - 08:31 AM, said:

Again, if in a team there are 5-6 double digits, then IT IS a MM issue.
Not one. Half a team shooting ONLY ONE TIME, it's simply reddicolous

there are 5-6 double digits, ...????

what are you talking about please?

[Lostdragon]

Stefka Kerensky, on 02 August 2017 - 08:31 AM, said:

Again, if in a team there are 5-6 double digits, then IT IS a MM issue.
Not one. Half a team shooting ONLY ONE TIME, it's simply reddicolous

MM can't fix that. If you have someone calling targets and people using FF and rotating the vanguard they will roll through a less organized team and cut them down. No mech can survive more than a few seconds with 3+ enemies shooting it and when people see that they try to retreat and it just snowballs into a lot of quick deaths in a 5 minute match.

MM can't predict who will get along with who and who will listen and who will not. Like I said before, even the best players get steamrolled sometimes, it is just the nature of the beast. The only way to stop stomps is to add respawns, a no respawn game mode is guaranteed to have matches that end that way even with the top 24 players in the game playing each other.

[fogsworth]

How did a T4 play a T1? I thought T1 could only fight T1-3?

[STEF_]

Lostdragon, on 02 August 2017 - 09:10 AM, said:

If you have someone calling targets and people using FF and rotating the vanguard they will roll through a less organized team and cut them down. No mech can survive more than a few seconds with 3+ enemies shooting it and when people see that they try to retreat and it just snowballs into a lot of quick deaths in a 5 minute match.

who will get along with who and who will listen and who will not. Like I said before, even the best players get steamrolled sometimes, it is just the nature of the beast. The only way to stop stomps is to add respawns, a no respawn game mode is guaranteed to have matches that end that way even with the top 24 players in the game playing each other.

AND you are talking about group que, while I see 6 double digits in SOLO que.

MM can't predict, MM can't fix that.
So, basicaly u are saying we don't have a MM, which is true.



Because a good MM with zero sum psr can actually fix a lot of $h1t.

[jonfett]

fogsworth, on 02 August 2017 - 09:16 AM, said:

How did a T4 play a T1? I thought T1 could only fight T1-3?

Hence, why MM is broken, and has been broken for over a year now. Also, why there is a post about it every week...

[Suko]

OP,

Was this in solo queue or group queue? If it's in Solo Q, then that really sucks. However, this is a very common thing in Group Q. I've even heard (though can't confirm) that MM doesn't really look at tier when in Group Q, so getting T1 vs T4 is definitely a possibility.

[Lostdragon]

Stefka Kerensky, on 02 August 2017 - 09:35 AM, said:

AND you are talking about group que, while I see 6 double digits in SOLO que.

MM can't predict, MM can't fix that.
So, basicaly u are saying we don't have a MM, which is true.



Because a good MM with zero sum psr can actually fix a lot of $h1t.

I never play group queue. Contrary to what people might believe from reading the forums people in QP are actually capable of coordination and teamwork. Sure, I see plenty of awful games, but I don't think it is as bad as many would have us believe, I often see people who listen to direction and work together. Sometimes you get that rare QP team where no effort is made to coordinate, everything just falls into place without a word said and your team makes all the right moves as if part of some collective conciousness. The next match you might be on the team with 50% of the same players and it all just falls apart and you get rolled.

Anecdotally, I would say in QP 25% of my matches my team wins handily, 50% are competitive and could go either way, and 25% my team gets wrecked. I often play several matches in a row with many of the same players, I see good pilots get and give stomps all the time. I think I will track some matches with EoR screencaps and see how accurate this is.

[Suko]

Lostdragon, on 02 August 2017 - 05:55 AM, said:

I just don't think there is a lot you are going to be able to accomplish with MM, it already basically works how you describe as far as matching tiers. The floodgates are open because populations are low. It doesn't matter if you change MM to work off a new formula, it won't solve that problem. You could have the best MM alogrithm in the world that accurately sorts players by skill with the highest precision possible, but if the number of top tier players is relatively low then you have to put them in matches with lower tier people to avoid wait times that make people want to quit.

It is kind of a no win situation for PGI. They could definitely improve the way tiers are assigned, but if there is not enough population of high tier players to segregate them from low tier players you wind up right back where we are now and they have wasted time and resources.

Another issue is that Tiers are just EXP bars. There are some players here in T1 that just shouldn't be. You can check their stats on the leaderboards and see it. Hell, I shouldn't be T1, I realistically should be a T2 player, maybe even upper T3. But, because I played enough pew pew games, I got to T1.

It also doesn't help that the tier system doesn't differentiate between Solo and Group games. Some of the T1 guys are there just because they played a lot of group matches and got dragged up to T1. Then they go play Solo Q once in a while and are just awful. Matchmaker thinks it put a T1 on your side, but the reality is they play like a T3 in Solo Q and your team suffers for it.

tldr= The tier system is broken. It's just an EXP bar and anyone can get to T1 be simply playing a lot of games. The tier rating needs to be less generous and should also degrade over time. I haven't played in over a month and if I hopped into a game right now, I sure as hell won't be playing like a T1 player until I get reacquainted with the game.

Edited by Suko, 02 August 2017 - 09:57 AM.

[Lostdragon]

Suko, on 02 August 2017 - 09:48 AM, said:

OP,

Was this in solo queue or group queue? If it's in Solo Q, then that really sucks. However, this is a very common thing in Group Q. I've even heard (though can't confirm) that MM doesn't really look at tier when in Group Q, so getting T1 vs T4 is definitely a possibility.

It has been established for a while that T5 and T1 are in QP matches together. I think this created some confirmation bias among people who are T1 and think every time they lose the rest of their team was T5 and the enemy was all T1. The MM is supposed to try to average out the tier level of each team, but we have no real way to prove conclusively how effective that is because tier is not shown on the EoR screen. My gut feeling is it is not nearly as bad as a lot of people make it out to be.

[Lostdragon]

Suko, on 02 August 2017 - 09:56 AM, said:

Another issue is that Tiers are just EXP bars. There are some players here in T1 that just shouldn't be. You can check their stats on the leaderboards and see it. Hell, I shouldn't be T1, I realistically should be a T2 player, maybe even upper T3. But, because I played enough pew pew games, I got to T1.

It also doesn't help that the tier system doesn't differentiate between Solo and Group games. Some of the T1 guys are there just because they played a lot of group matches and got dragged up to T1. Then they go play Solo Q once in a while and are just awful. Matchmaker thinks it put a T1 on your side, but the reality is they play like a T3 in Solo Q and your team suffers for it.

tldr= The tier system is broken. It's just an EXP bar and anyone can get to T1 be simply playing a lot of games. The tier rating needs to be less generous and should also degrade over time. I haven't played in over a month and if I hopped into a game right now, I sure as hell won't be playing like a T1 player until I get reacquainted with the game.

Yeah, it doesn't make sense to use one rating for two very different game types. Just for the record, I do think the MM and tier system can and should be improved, but I don't think it will ever be the silver bullet some people think it would/could be when it comes to generating more matches that are 11-12 or whatever threshold people think means the match was good.

[Brain Cancer]

The matchmaker has never worked, because tier numbers mean nothing and the resulting random roflstompage drove so many players away, most times it simply grabs whatever it can and stuffs it into matches regardless of tier levels. PSR system does nothing to sort by tier, after all.

If people could see actual tier rankings in their games and how much potato really grows in the fields, you could use the liquid sodium to cool a dozen city-sized nuclear reactors.

[Escef]

MatchMaker, I think, is working as it was designed to, the problem is the tier system, which was made to be used as criteria for the MM, is somewhat flawed.