Jump to content

Solo Queue Match Maker Tightened Up.


254 replies to this topic

#201 SFC174

    Member

  • PipPipPipPipPipPipPip
  • The Pharaoh
  • The Pharaoh
  • 695 posts

Posted 10 September 2017 - 08:38 AM

TL:DR - stomps are too common, the MM tightening hurt me, PSR is not representative of my skill (overrates me), would like to see improvements.


My stats have been hit particularly hard since the MM tightening. Previous couple of seasons I was averaging a match score of around 315-320 with W/L of 1.2 and KDR of 1.5 showing a gradual trend of improvement. I hit Tier 1 mid-season 13 and am about 10% of the bar away from maxxing Tier 1 now.

At the start of this season (pre-MM change) I was at 330 avg score with W/L of 1.1 and KDR of 1.6 (or about what you'd predict from my 3-season avg, especially given the small number of games I'd played at that point). Then the MM change hit and I rapidly plummeted.

At my lowest I was at a match score of 270, W/L of 0.7 and KDR of 1.0. While I bemoan the MM changes (or more accurately, how poorly representative tier is of skill - despite my crappy record this season, my PSR has still gone up!!), I also tried to adapt, so I went more try-hard/meta, and started only playing the mechs I tend to do really well with - HBR laser vomit or ERPPC sniper, HBK IICA laser vomit, MAD-IIC laser vomit, etc. In the 20 or so matches since I did that, I've managed to climb back up to avg match of 290, W/L of 0.9 and KDR of 1.1. It will be tough to get back to my previous season avgs, but at least I've mitigated the pain.

I started screenshotting most of my matches after the MM announcement as well. At first I was going through and looking at avg match scores for all the players, but honestly, its a game and I don't have time to do a proper statistical analysis on that _and_ play the game. I did pull out the W-L stats for the last 15 games I played though and the number of players doing less than 200 dmg on each team. That seems to the level where, if you fall below, you're just not contributing enough to win (exceptions noted like capping lights, AMS boats). And 200 dmg seems like a level most mechs should be able to hit without too much effort (note, these matches are all since I shifted to tryhard mode so I was not in the under 200 club in any of them) The numbers were as follows:

(W/L -- Players below 200 dmg)
3-12 -- 6-1
12-3 -- 0-3
12-2 -- 2-7
9-7 -- 6-4 cap
12-7 -- 2-2
1-12 -- 8-3
12-2 -- 1-8
1-12 -- 10-1
4-12 -- 4-0
5-12 -- 5-4
12-10 -- 1-1
12-3 -- 5-6
7-12 -- 4-4
6-12 -- 2-3
12-4 -- 3-5

Now, I don't know if there is a commonly accepted range for what is a stomp or not. I'd say 12-3 or worse. Some might say 12-2 or 12-4. At 12-3 there were 7 stomps out of 15 matches. Generally in the stomps the losing team had a lot more "low achievers" than the winning team as you would expect.

MM vagaries aside, half the matches resulting in stomps doesn't seem like a good way to retain players or generate interest. Yes, winning a stomp feels ok, but after awhile, not so much. And losing in a stomp sucks. Nothing like getting into position, taking a few shots and looking at the status screen to find you're down 0-5 2 min after the engagement began.

Edited by SFC174, 10 September 2017 - 10:44 AM.


#202 MischiefSC

    Member

  • PipPipPipPipPipPipPipPipPipPipPipPip
  • The Benefactor
  • The Benefactor
  • 16,697 posts

Posted 10 September 2017 - 09:16 AM

Except those numbers will always be like that. Comp games between teams who are a razors edge of skill apart are 8-1 stomps all the time.

There is nothing you will ever do with the matchmaker that will change that, it's a product of mech durability, higher than average TTK for a FPS and the benefits of focused fire.

As soon as one team loses a few mechs they are at a significant disadvantage and the match snowballs. That is irrelevant of player skill.

The MM doesn't overrated you - just that most players are absolutely terrible. If you're not in danger of choking on your own tongue, you don't play with a joystick and you understand that some weapons and builds are better than others you're miles ahead of most players.

#203 SFC174

    Member

  • PipPipPipPipPipPipPip
  • The Pharaoh
  • The Pharaoh
  • 695 posts

Posted 10 September 2017 - 09:31 AM

I think it does overrate me. After all, with the previous MM, I was consistently winning more than I lost. With the new one (and no real change in the mechs I play, plus more experience on my part) I was suddenly losing more than I won - the W/L ratio basically flipped. I had to go try hard and even then I'm still about even. That means that based on my PSR/Tier/whatever, I'm expected to carry more than I'm capable of at this point. In mainly Tier1 only matches, I'm a loser (baby, why don't you kill me). Yet my PSR continues to rise. Should I be Tier1?

Maybe that's the way things should be, I don't know. But I'm sensing (and feeling) a lot of discontent with the matchups since PGI tightened the MM. Not having as much fun to be sure.

#204 Tomo Sukesada

    Member

  • PipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 97 posts

Posted 10 September 2017 - 09:37 AM

PGI has the data, all they would need to do is measure the percentage of stomps pre-change vs post-change. This fairly straight forward measurement would tell us. But I am certainly not seeing more close matches over the last few days...

#205 Exilyth

    Member

  • PipPipPipPipPipPipPipPipPip
  • Bridesmaid
  • 2,100 posts
  • LocationTerra

Posted 10 September 2017 - 12:17 PM

Meh.

I'm getting a mixed bag of games, some good, some horrible. One noticable change was that when stomps happen, they tend to be more onesided.

I seem to be rising in tier faster too, just got T3. So far, doesn't play much different from T4 except that some matches seem to require piloting with a smaller margin of error.

#206 TWIAFU

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Pest
  • The Pest
  • 4,011 posts
  • LocationBell's Brewery, MI

Posted 10 September 2017 - 12:37 PM

View PostKodiakGW, on 08 September 2017 - 09:50 PM, said:

Yep, matchmaker certainly looks tightened up to me....

Posted Image

My one, and only, game for the weekend. Only waited about 30 seconds for this game. So this is all T1 and 2.

Looks like your "fix" isn't worth the effort you put in. Try again.



Just wait for 8v8 to "fix" QP.

#207 Tarl Cabot

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Tai-sho
  • Tai-sho
  • 7,828 posts
  • LocationImperial City, Luthien - Draconis Combine

Posted 10 September 2017 - 01:22 PM

View PostKodiakGW, on 08 September 2017 - 09:50 PM, said:

Yep, matchmaker certainly looks tightened up to me....

Posted Image

My one, and only, game for the weekend. Only waited about 30 seconds for this game. So this is all T1 and 2.

Looks like your "fix" isn't worth the effort you put in. Try again.

It was really never a fix, it simply removed the disgruntled MM is broken I am getting dropped with T4/5 folks!! The MM is working with what it has been provided per the PSR. PSR has not been reset and seeded with last 3-6 seasons. There are people who are high tier simply due to playing enough and doing average damage, imho this includes me.

Even if the PSR is ever sorted out, stomps like that will still happen due to lack of communication, bad positioning, bad luck and too many being passive (due to lack of communiction/no plan) vs the other side of the coin. There is no respawn, no hail marys (unless conquest, etc).

It is possible PGI first made this change to the MM to see HOW things shake out before doing something about the PSR itself /shrugs. I am likely giving them too much credit though... Or they did not wish to actually announce it which then people would have attempted to play the system prior the reset... but I do not see how it would have mattered unless PGI has no intention on realigning the PSR every quarter/etc.

Edited by Tarl Cabot, 10 September 2017 - 01:24 PM.


#208 Bigbacon

    Member

  • PipPipPipPipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 3,108 posts

Posted 10 September 2017 - 02:13 PM

View PostExilyth, on 10 September 2017 - 12:17 PM, said:

Meh.

I'm getting a mixed bag of games, some good, some horrible. One noticable change was that when stomps happen, they tend to be more onesided.

I seem to be rising in tier faster too, just got T3. So far, doesn't play much different from T4 except that some matches seem to require piloting with a smaller margin of error.


you aren't rising faster it just damn near impossible to actually go down.

#209 Christophe Ivanov

    Member

  • PipPipPipPipPipPip
  • Big Daddy
  • Big Daddy
  • 385 posts
  • LocationSeattle area

Posted 10 September 2017 - 02:28 PM

like I always say, Tier means absolutely nothing to me. Your good one match, terrible the next. All depends on the circumstances at the time you play a match. Get rid of the Tier effects on QP to keep it simple folks. Just saying because I can.

#210 mistlynx4life

    Member

  • PipPipPipPipPipPip
  • 351 posts

Posted 10 September 2017 - 02:48 PM

View PostMischiefSC, on 10 September 2017 - 06:50 AM, said:

If someone wants to win more, they need to get better at the game.


I respectfully disagree. You need to define what getting better actually means, and if it's dealing damage then why bother to have anything other than optimized builds for damage or modes that have win conditions that aren't about destroying the enemy? Official word is that raising your PSR has to do with playing as a team but in a drop where you don't know your own team composition or even the map/mode ahead of time, if you're trying to hedge your PSR bets, you need to design a mech that can operate completely on its own... which is the opposite of teamwork.

And this is why PSR always inspires such lively discussion - because it's an unclear ranking system towards an unclear goal representing an unclear quantity about a player. It's like arguing about how fast a car can go without knowing how much fuel is in the tank or if there's even an engine under the hood... or if there's even a race going on at all.

#211 jss78

    Member

  • PipPipPipPipPipPipPipPipPip
  • Philanthropist
  • Philanthropist
  • 2,575 posts
  • LocationHelsinki

Posted 10 September 2017 - 02:52 PM

View PostMischiefSC, on 10 September 2017 - 09:16 AM, said:

Except those numbers will always be like that. Comp games between teams who are a razors edge of skill apart are 8-1 stomps all the time.

There is nothing you will ever do with the matchmaker that will change that, it's a product of mech durability, higher than average TTK for a FPS and the benefits of focused fire.

As soon as one team loses a few mechs they are at a significant disadvantage and the match snowballs. That is irrelevant of player skill.

The MM doesn't overrated you - just that most players are absolutely terrible. If you're not in danger of choking on your own tongue, you don't play with a joystick and you understand that some weapons and builds are better than others you're miles ahead of most players.


Agreed, I was never quite convinced the stomps are produced by imbalanced MM. But it's psychologically tempting to go "PGI plz fix!" instead of accepting that sometimes stuff just happens...

If anything, in pure high-tier games with exclusively good players (for the sake of argument let's assume the current Tier 1-2 guys are "good"), I'd expect the tighter MM to make stomps more frequent. Skilled players play more cohesively and make better use of their numbers advantage. Low-skill players might still feed themselves to you one at a time, giving easy kills and a chance to recover the game.

#212 Brain Cancer

    Member

  • PipPipPipPipPipPipPipPipPip
  • The 1 Percent
  • 3,851 posts

Posted 10 September 2017 - 04:52 PM

The system as it stands doesn't even let you go "Tier 1 good, tier 2 less good" etc. etc.

It's basically "You played minimally in X number of games, at least." Even with the new MM, there's plenty of target practice to be had. I just watched an entire lance walk up to the top of a hill, one at a time, and get Gaussed to death while standing in place, because they saw no red doritos.

These had to be T3 players, minimum. An average player SHOULD know better, but the PSR system floats mediocre or worse players into the top tiers regardless. Only the truly abysmal remain in the T5 underhive.

#213 adamts01

    Member

  • PipPipPipPipPipPipPipPipPip
  • Death Star
  • 3,417 posts
  • LocationPhilippines

Posted 10 September 2017 - 09:49 PM

View PostMischiefSC, on 10 September 2017 - 09:16 AM, said:

Except those numbers will always be like that. Comp games between teams who are a razors edge of skill apart are 8-1 stomps all the time.

There is nothing you will ever do with the matchmaker that will change that, it's a product of mech durability, higher than average TTK for a FPS and the benefits of focused fire.

As soon as one team loses a few mechs they are at a significant disadvantage and the match snowballs. That is irrelevant of player skill.

The MM doesn't overrated you - just that most players are absolutely terrible. If you're not in danger of choking on your own tongue, you don't play with a joystick and you understand that some weapons and builds are better than others you're miles ahead of most players.
Everything about your post is a contradiction. Comparing stomps in comp games to stomps in quick play is apples and oranges. Half the stomps that happen in quick play are because half the players stuck on one team are "absolutely terrible" as you put it. What MM can do is filter out lore builds and joystick users. Nothing wrong with those, but two of them on the same team is a death sentence. Get a LRM Jenner and a LRM Atlas on the same team and you've lost before the match has even started, not because of the same snowball effect you see in comp games, but because of the potato lottery. It's an important difference to distinguish.

#214 adamts01

    Member

  • PipPipPipPipPipPipPipPipPip
  • Death Star
  • 3,417 posts
  • LocationPhilippines

Posted 10 September 2017 - 09:53 PM

View PostChristophe Ivanov, on 10 September 2017 - 02:28 PM, said:

like I always say, Tier means absolutely nothing to me. Your good one match, terrible the next. All depends on the circumstances at the time you play a match. Get rid of the Tier effects on QP to keep it simple folks. Just saying because I can.

This isn't true at all. Average players are good one match and terrible the next. Great players are good no matter what. I'm somewhere in the middle of those, and a perfect MM with a perfect playerbase would keep me in tier 2. I don't care if he brings a LRM build on a brawling map, a good player will find a way to make his build work, regardless of his team.

#215 Thorqemada

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • 6,396 posts

Posted 10 September 2017 - 11:14 PM

View PostMischiefSC, on 10 September 2017 - 06:50 AM, said:

Except stomps always happen, will always happen. The fact that teams saddled with terribads will always be functionally gimped from doing anything coherent or useful isn't a good thing.

It's a 12 v 12 team game. Or, hopefully, 8 v 8. You're always on a team. Trying to gimp teamwork by design is a truly magical kind of bad game design.

If someone wants to win more, they need to get better at the game. Not try to design the game to increase the odds of playing against worse opponents.

Elo-hell was a myth people used to justify why they lost matches. The biggest problem with Elo as it was implemented was a sort of high-low balance, goods and bads on a team to make an average when human skill doesn't work that way, the bads are significantly more of a negative than the goods are a positive. If you're in a tighter band that's way less of an issue and your improvement is more impactful on the teams overall win odds - so if you're rated a 1400, but you work hard, git gud and are playing like a 1600, you'll move up because you're playing with/against 1400s. If you're on a team of 600s-2000, averaging 1400 then your performance change can get washed down to minimal value and take a lot of matches to settle.

Also we need a mech Elo score. Especially with the last 3 years of balance changes the value of the mech and loadout you take can be massively impactful. For bad players the value of the mechs performance can be as much of a factor as their skill. For good players where a win/loss flips on a tiny advantage it can still be critical.

What we need is player and mech Elo and tighter band matchmaking.


Sorry this is only half the picture.
Elo-Hell forced a Player to carry a Team way more and wait way longer then PSR bcs of the discrepancy of Elos that could occur in a Team and the struggle to build a Team from the Elo-Pool first which pretty much prevented some Players from playing QP.

And you did not read - Stomps happen - more Stomps happen the bigger the skill coherencey of hte Teams is bcs there is no Potato that stops the Snowball-Effect for taking no action or doing wrong like in incoherent Team Setups.

#216 Thorqemada

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • 6,396 posts

Posted 10 September 2017 - 11:31 PM

View PostMischiefSC, on 10 September 2017 - 09:16 AM, said:

Except those numbers will always be like that. Comp games between teams who are a razors edge of skill apart are 8-1 stomps all the time.

There is nothing you will ever do with the matchmaker that will change that, it's a product of mech durability, higher than average TTK for a FPS and the benefits of focused fire.

As soon as one team loses a few mechs they are at a significant disadvantage and the match snowballs. That is irrelevant of player skill.

The MM doesn't overrated you - just that most players are absolutely terrible. If you're not in danger of choking on your own tongue, you don't play with a joystick and you understand that some weapons and builds are better than others you're miles ahead of most players.


A game without fun is a worthless game, Fun comes from the gameplay and/or the interaction (which PGI does not promote), success brings Satisfaction, Prestige, Self-Esteem - which one you lack you strive for.

In my eyes a game not played for fun is a plight, sort of a boring Job that inflates a meaningless statistic.

Now dont get that wrong, i like to have success and i like good statistics but the overarching goal is a fun gameplay and then get the **** (most success/achievements possible) out of it!
Not to have the most success and biggest stats with boring exploiting the game mechanics and then autosuggest myself into believeng i had fun while i feel totally bored in truth...


When you have real fun and real Entertainment - then you are Miles ahead!

Edited by Thorqemada, 10 September 2017 - 11:32 PM.


#217 Roadbuster

    Member

  • PipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 1,437 posts
  • LocationAustria

Posted 11 September 2017 - 12:19 AM

It's funny how people complain about longer wait times.

I've been playing this weekend without knowing about the change to matchmaking.
Guess what. I didn't notice any change in wait times.
Sometimes I had to wait 1-2 minutes, other times matches were instant.
Overall, wait times felt like they always did.

Match quality was better though.
I didn't think too much about this, because depending on time of day, match quality always varied.
But overall, the matches were closer more often, less stomps and, now that I think of it, there were very few players who had no clue what they were doing.

Now that I know what PGI did, my experience from this weekend makes sense.
Seems like a positive change.

Posted Image

#218 Nema Nabojiv

    Member

  • PipPipPipPipPipPipPipPip
  • 1,783 posts
  • LocationUA

Posted 11 September 2017 - 01:22 AM

View PostSFC174, on 10 September 2017 - 09:31 AM, said:

I think it does overrate me. After all, with the previous MM, I was consistently winning more than I lost. With the new one (and no real change in the mechs I play, plus more experience on my part) I was suddenly losing more than I won - the W/L ratio basically flipped. I had to go try hard and even then I'm still about even. That means that based on my PSR/Tier/whatever, I'm expected to carry more than I'm capable of at this point. In mainly Tier1 only matches, I'm a loser (baby, why don't you kill me). Yet my PSR continues to rise. Should I be Tier1?

Maybe that's the way things should be, I don't know. But I'm sensing (and feeling) a lot of discontent with the matchups since PGI tightened the MM. Not having as much fun to be sure.

I have pretty much the same impression. Up to the point I started playing less now.

#219 QuantumButler

    Member

  • PipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 4,534 posts
  • LocationTaiwan, One True China

Posted 11 September 2017 - 01:40 AM

View PostRoadbuster, on 11 September 2017 - 12:19 AM, said:

It's funny how people complain about longer wait times.

I've been playing this weekend without knowing about the change to matchmaking.
Guess what. I didn't notice any change in wait times.
Sometimes I had to wait 1-2 minutes, other times matches were instant.
Overall, wait times felt like they always did.

Match quality was better though.
I didn't think too much about this, because depending on time of day, match quality always varied.
But overall, the matches were closer more often, less stomps and, now that I think of it, there were very few players who had no clue what they were doing.

Now that I know what PGI did, my experience from this weekend makes sense.
Seems like a positive change.

Posted Image

timezones exist friendo, not everyone plays at prime Canadian or USA or Euro times.

Off peak wait times for me have most definitely gotten far, far worse.

#220 Brain Cancer

    Member

  • PipPipPipPipPipPipPipPipPip
  • The 1 Percent
  • 3,851 posts

Posted 11 September 2017 - 02:52 AM

Quote

This isn't true at all. Average players are good one match and terrible the next. Great players are good no matter what. I'm somewhere in the middle of those, and a perfect MM with a perfect playerbase would keep me in tier 2. I don't care if he brings a LRM build on a brawling map, a good player will find a way to make his build work, regardless of his team.


Oh, trust me. Put a "great" player in QP and they will have their bad games as well.

If anything, reputation can make life harder as people go out of their way to burn the "good" guy out before he does too much damage. Last game I was in with B33f, he didn't cross 200 damage, the team did horrible, and I think I topped the charts with 2 kills and like 350 damage on the loss.

Everyone has their bad games. Everyone. Now, you DO see those well-polished stats at times, but that's much more a benefit of group play than solo play. A good player solo can't always carry, but a solid group is much more difficult to break.

Edited by Brain Cancer, 11 September 2017 - 02:53 AM.






1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users