Jump to content

Solo Queue - Mm Update (30/05/2018)


159 replies to this topic

#61 Nekromanteia

    Rookie

  • Ace Of Spades
  • Ace Of Spades
  • 4 posts
  • LocationSingapore

Posted 01 June 2018 - 09:14 AM

As a new player to the MWO scene, I think that this issue being picked up on by a staff is good news for everyone. It's not going to be perfect; but then again nothing is and you can't please everyone.

If anything, it's a step in the right direction regardless of it being delivered properly or lack thereof. Cheers!

#62 justcallme A S H

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • CS 2020 Referee
  • CS 2020 Referee
  • 8,987 posts
  • LocationMelbourne, AU

Posted 01 June 2018 - 10:16 AM

View Postkapusta11, on 01 June 2018 - 04:46 AM, said:


We have just one tier for people who can pull their weight between 1 and 5+ times and 4 tiers for those who can't do it even once, that's ridiculous. When top 20% have so much impact on the outcome of the game they are the ones that should be split into tiers depending on their performance, not the homogenous 80%.


I believe I said the top 3-5%... Which is what needs to happen. I would wait another 90-120s for better games. As would most.

View Postkapusta11, on 01 June 2018 - 04:46 AM, said:

No I'm not facing the right people. If there are 6 tier 1 players in queue: proton, bowser, heimdelight and 3 people like me (or even worse) the game can put the former 3 on one team and the latter 3 on the other.


Tarogato addressed this major flaw on Page - 1.

I addressed the other 50% of the issue on Page - 2.

Between them both, that is most of the MM issues, covered.

Granted it wont be perfect but it will be far better than what we had have for over 2 years

#63 Brain Cancer

    Member

  • PipPipPipPipPipPipPipPipPip
  • The 1 Percent
  • 3,851 posts

Posted 01 June 2018 - 11:02 AM

View Postkapusta11, on 01 June 2018 - 01:22 AM, said:

It's not matching tier 1 players with tier 4 or 3 ones that is the problem, it's tiers themselves.

I have 1.5 and 2.0 W/L and K/D ratios respectively yet the game makes no distinction between me and someone who has 5+ W/L and K/D ratios, the likes of Proton and Bows3r and there are tons of people like me who don't belong in tier 1. Basically, all people who can pull their weight are lumped together and we have 4 tiers for those who don't, it's bananas, it should be the other way around.


You're in the top 5%, almost 4% of the game's players.

No, really. Check for yourself on Jarl's. Even someone like me who's spent the past few seasons playing too many horrid, unskilled things for events is still pushing the high 80s for those seasons with a comparative 1.16 and 1.19. It can't parse much further than that or you'd literally never see that top 5% get a game.

Edited by Brain Cancer, 01 June 2018 - 11:03 AM.


#64 kapusta11

    Member

  • PipPipPipPipPipPipPipPipPip
  • Little Helper
  • Little Helper
  • 3,854 posts

Posted 01 June 2018 - 12:20 PM

Forget about me, I'm talking about tier 1 as a group, the one where skill difference is much broader than in the next 4 tiers combined. The impact someone with 5 K/D ratio makes on the outcome of the game compared to someone with 2 K/D ratio is much greater than a potato with 1 K/D ratio vs another potato with 0.7 K/D ratio, yet the former two are put in one tier and the game makes no distinction between them and the latter two could be 2 tiers apart.

My point is this: put all potatoes in one, maybe two tiers and then arrange the rest (5% 10% 20% I dunno how many people are in tier 1) in 3-4 tiers. Lock tiers, make it impossible to progress. Every 6-12 months reset PSR and assign new tier for everyone. Skill doesn't change much, not in a short period of time. Right now it's just an exp bar for you to stroke and should not be used in matchmaking at all.

And if we're going to use match score (which can be infuenced by things that don't require much skill btw) like Tarogato suggested why keep fundamentally flawed tier based matchmaking to begin with? To waste everyone's time? Current matchmaker uses tiers that don't mean anything. I'd say fix it instead of making things even more convoluted.

Edited by kapusta11, 01 June 2018 - 12:21 PM.


#65 Doomich

    Member

  • PipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 68 posts
  • LocationRussia

Posted 01 June 2018 - 01:00 PM

View Postkapusta11, on 01 June 2018 - 12:20 PM, said:

And if we're going to use match score (which can be infuenced by things that don't require much skill btw) like Tarogato suggested why keep fundamentally flawed tier based matchmaking to begin with? To waste everyone's time? Current matchmaker uses tiers that don't mean anything. I'd say fix it instead of making things even more convoluted.

This.

#66 Panzerkuh

    Member

  • PipPip
  • Elite Founder
  • Elite Founder
  • 22 posts

Posted 01 June 2018 - 02:00 PM

Sigh RIP synch drops for us streamers, can't get any games with community now

#67 justcallme A S H

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • CS 2020 Referee
  • CS 2020 Referee
  • 8,987 posts
  • LocationMelbourne, AU

Posted 01 June 2018 - 04:25 PM

I had no issues with that last night...
It should not change anything actually.

#68 SFC174

    Member

  • PipPipPipPipPipPipPip
  • The Pharaoh
  • The Pharaoh
  • 695 posts

Posted 01 June 2018 - 06:27 PM

View Postkapusta11, on 01 June 2018 - 12:20 PM, said:

Forget about me, I'm talking about tier 1 as a group, the one where skill difference is much broader than in the next 4 tiers combined. The impact someone with 5 K/D ratio makes on the outcome of the game compared to someone with 2 K/D ratio is much greater than a potato with 1 K/D ratio vs another potato with 0.7 K/D ratio, yet the former two are put in one tier and the game makes no distinction between them and the latter two could be 2 tiers apart.

My point is this: put all potatoes in one, maybe two tiers and then arrange the rest (5% 10% 20% I dunno how many people are in tier 1) in 3-4 tiers. Lock tiers, make it impossible to progress. Every 6-12 months reset PSR and assign new tier for everyone. Skill doesn't change much, not in a short period of time. Right now it's just an exp bar for you to stroke and should not be used in matchmaking at all.

And if we're going to use match score (which can be infuenced by things that don't require much skill btw) like Tarogato suggested why keep fundamentally flawed tier based matchmaking to begin with? To waste everyone's time? Current matchmaker uses tiers that don't mean anything. I'd say fix it instead of making things even more convoluted.


Can't like this enough. I had the pleasure of dropping against proton 4 times today in 2 different 1 hr play sessions (lost 12-3, 12-2, 12-2 and 12-3). I was quite salty by the end of it, you'd figure I'd get him on my team at least once. In 2 of those matches there were 1 or 2 other high ranked players on the opposing team while my team was composed of people like myself (avg Tier 1 players who should probably be Tier 2 or something). I have tremendous respect for Proton's skills, but if I continually have to face god tier players without one on my team to balance, it's time to log off and try again later (which I did, and still got ganked again :) )

If we're not going to rework Tiers in an effective manner, then a match score balancing is probably the next best thing.

p.s. - its strange but I always seem to get the biggest matchmaking issues at the beginning of a season. I don't know if something in the matchmaker gets reset each month, but the first few days are usually outliers. I either go on monster win streaks against inferior opponents or get stuck with matches like I did today.

#69 SFC174

    Member

  • PipPipPipPipPipPipPip
  • The Pharaoh
  • The Pharaoh
  • 695 posts

Posted 01 June 2018 - 06:32 PM

View PostBrain Cancer, on 01 June 2018 - 11:02 AM, said:


You're in the top 5%, almost 4% of the game's players.

No, really. Check for yourself on Jarl's. Even someone like me who's spent the past few seasons playing too many horrid, unskilled things for events is still pushing the high 80s for those seasons with a comparative 1.16 and 1.19. It can't parse much further than that or you'd literally never see that top 5% get a game.


I think you just have to balance them with another similar skill player on the opposing team. I have no issues with mixing diverse skill levels as long as there is near 1:1 equivalency on each team. Or, you need to balance with enough potatoes (by match score totals perhaps?), if you can't find a similar player. The latter is less preferable because, well, how many potatoes does it take to go all Harrison Bergeron on a guy like Proton or Nova or Hammer Coleman? I don't know and it won't be easy to figure out.

#70 Tier5 Kerensky

    Member

  • PipPipPipPipPipPipPipPipPip
  • Bridesmaid
  • 2,049 posts

Posted 01 June 2018 - 08:04 PM

I believe these changes are a good step forward. Few days ago I had the most unusual game I've ever been, 9 assults per side. We got totally wacked but during the about 15k games I've played I've never seen so many assults in SQ. Didn't suspect it would be because of MM changes.

I don't know how these will change games, less stomps, more stomps, but it should be overall more fail and equal.

Edited by Teer Kerensky, 01 June 2018 - 08:04 PM.


#71 Humpday

    Member

  • PipPipPipPipPipPipPipPip
  • The Pharaoh
  • The Pharaoh
  • 1,462 posts

Posted 01 June 2018 - 08:21 PM

I had to make some serious gameplay adjustments. Namely backing the eff off. It would see most people are too scared too engage aggressively and would rather hang back.

But the annoying thing is, is when you're leading the way, all they do is jump in front of you trying to scavenge all your work.

I've now switched to longer range due to this.

This change doesn't help anything. Everyone bubbles up to t1-3, so if you're crap, you still are crap.

Not to mention if you quit the game 2 years ago and came back at t1, you're still poo. System is broken.

#72 SFC174

    Member

  • PipPipPipPipPipPipPip
  • The Pharaoh
  • The Pharaoh
  • 695 posts

Posted 01 June 2018 - 11:06 PM

View PostTeer Kerensky, on 01 June 2018 - 08:04 PM, said:

I believe these changes are a good step forward. Few days ago I had the most unusual game I've ever been, 9 assults per side. We got totally wacked but during the about 15k games I've played I've never seen so many assults in SQ. Didn't suspect it would be because of MM changes.

I don't know how these will change games, less stomps, more stomps, but it should be overall more fail and equal.


Outside of games where the matchmaker gets unbalanced with god tier players, I've been finding the matches are more competitive. I don't really mind if there are more assaults and heavies playing. It actually makes playing some lights more fun (target rich environment).

#73 N3tRunn3r

    Member

  • PipPip
  • Philanthropist
  • 28 posts
  • LocationTemple of Nod

Posted 02 June 2018 - 01:56 AM

View PostPaul Inouye, on 30 May 2018 - 02:10 PM, said:


...

[color=orange]Blah blah blah.. numbers number numbers... what does this mean to me? Here's the TL;DR[/color]

* Wait times have been reduced between 40-50%.
* MAXIMUM average wait times have dropped from 150 seconds to 80 seconds.
* Matches being played are a lot more challenging.
* Average PSR ranking differences between two teams has gone from 950 to 200.

I've monitored this data over the last 3 days watching all 3 prime times (NA/EU/OC) and non-prime times and the data is consistent.

...

So that's it for now. I'll be keeping an eye on the wait times and I've made my own copy of the keys to the server config tool. Posted Image


Thank you very much by taking some time and effort into looking to this issue we had. I will test the Solo QP MM soon.

Posted Image

#74 Alreech

    Member

  • PipPipPipPipPipPipPipPip
  • Little Helper
  • Little Helper
  • 1,649 posts

Posted 02 June 2018 - 07:18 AM

View PostPaul Inouye, on 30 May 2018 - 02:10 PM, said:

1) Stop allowing Tier 1 players to play against Tier 4 and 5 players completely. Hard line.. no ifs ands or buts no matter how low the player count is in off peak times. To do this, the value for Tier separation has been set to 2. The biggest skill gap allowed now is Tier 1 to Tier 3. Tier 2 to Tier 4. Tier 3 to Tier 5.

2 & 2a) The Match Maker is going to wait no longer than 20 seconds to come up with a 3/3/3/3 build. After that 20 seconds is up, it will INSTANTLY allow 6/6/6/6 weight class distribution and eventually widen up to 12/12/12/12 over the next 30 seconds. The Match Maker will also weight class match on both teams. If there are 6 assaults on one team, there's guaranteed to be 6 assaults on the other. This also helps SIGNIFICANTLY when the marketing team runs an event using a single 'Mech as it's reward giving element.

[color=orange]Blah blah blah.. numbers number numbers... what does this mean to me? Here's the TL;DR[/color]

* Wait times have been reduced between 40-50%.
* MAXIMUM average wait times have dropped from 150 seconds to 80 seconds.
* Matches being played are a lot more challenging.
* Average PSR ranking differences between two teams has gone from 950 to 200.

I've monitored this data over the last 3 days watching all 3 prime times (NA/EU/OC) and non-prime times and the data is consistent.

Thanks for making effords to fix the Solo Queue in a game that is build around teamwork of 3 lances of 4 players on each side.

But to be honest: Playing Solo in a game what wants to be about teamwork isn't the best idea.
Even with the new matchmaking settings the team is created new for each match and disbanded after the match.
In the short time before the match is started almost no player cares to take company or even lance command, and even if the does he knows nothing about the other players loadouts...
During the match building a murderball and playing skirmish (regardless of the gamemode) is the best tactic - because it doesn't need much teamwork.

Many Stomps in MWO are the result of lack of teamwork. Some players don't join the murderball (because the want do do silly stuff like capturing CPs or the enemy base or scouting for the enemy) and weak their team.
Other reasons are the stuck with the wrong mech on the wrong map syndrome (adding dropdecks to quickplay would be helpfull) and the lack of respawns (don't take a risk if you can't respawn).
Respawns would also make the match last longer, improve the chance that the team improves the teamwork during the match.

But IMHO the best solution would be the improvment of the "Looking for Group" feature.
- VOIP for Groups in the Frontend so you don't have to join an external Teamspeak/Discord
- showing the region and language of the players in the group
- selection of a group description with the playstyle the group wants (like "Light Mechs" or "Brawling")
- fixing the group size to 4 / 8 and 12 players with the option to upgrade if the group is full
- game modes for groups of 4 players / 8 players / 12 players

#75 Scyther

    Member

  • PipPipPipPipPipPipPipPip
  • 1,271 posts
  • LocationOntario, Canada

Posted 02 June 2018 - 07:22 AM

I think an MM that gives us 'somewhat equivalent' matches, that is flexible enough to apply across QP/GP/Solaris and even FP (seriously PGI, putting FP noobs up against veteran teams was worst idea ever), would be good enough to provide interesting matches with some variety of team makeups.

A 'good match' should include more than just one factor - PSR, tiers, weight classes, or individual player skill. Something that combines aspects of all of those while being a simple/fast calculation could help a lot when matches need to be made with less-than-ideal pilot composition in the queues.

A good match is one (IMO) in which player skills, quality of mech builds, and weight/speed of the mechs is distributed in such a way as to provide an interesting fight - as opposed to a steamroll. Roflstomps will always happen but increasing the ratio of good fights to stomps across all modes would go a long way towards creating greater player satisfaction.

Just gonna toss this snippet from another thread in here.

"Match Value is based on the 'Mech and modified by pilot stats (PSR or performance). A simplified version would be:

Base Mech value: the average Match Score of that mech for that pilot, defaults to players average Match Score if < 10 matches
Mechlab modifier: (Alpha strike * DPS) * (Top speed/100) ... (of the saved mech build)

Pilot modifier: Tier 5 - 0.7 Tier 4 - 0.9 Tier 3 - 1.1 Tier 2 - 1.3 Tier 1 - 1.5

Match Value = (Base Mech Value + Mechlab Modifier) * Pilot Modifier

This involves very minimal calculation. The 'Mechlab' value is saved with the mech when you 'save' in the Mechlab. The average match score is also calculated and saved on the mech when any match ends.

Now you can eliminate Solaris divisions completely (or break it down to 2-3), use any weight class, base a Match Value on saved Drop Decks, toss in the Pilot Modifier on the fly (which could include more than current tier but probably doesn't need to). Pull any number of people from the waiting queue and simply sort them so the Match Value on both sides is roughly equal. It doesn't have to be too fussy - up to a 56/44 unequal weighting would still give better matches than we currently get, and would solve a pile of FP and Solaris issues at the same time.

FP should also add a modifier for group sizes. The Match Value of the drop group would be added together and multiplied by 1.1 (3-5 man ), 1.25 (6-8 man) or 1.4 (9-12 man)."


Primary goal: fewer Solaris divisions, a useable MM for FP, and the ability to construct a potentially interesting/semi-'fair' match even when queues hold a minimal selection of pilots. Not saying it's an ideal solution but something along these lines would at least contribute to more consistently entertaining matches.

Edited by MadBadger, 02 June 2018 - 08:44 AM.


#76 Jman5

    Member

  • PipPipPipPipPipPipPipPipPip
  • Littlest Helper
  • Littlest Helper
  • 4,914 posts

Posted 02 June 2018 - 10:13 AM

View Postkapusta11, on 01 June 2018 - 01:22 AM, said:

It's not matching tier 1 players with tier 4 or 3 ones that is the problem, it's tiers themselves.

I have 1.5 and 2.0 W/L and K/D ratios respectively yet the game makes no distinction between me and someone who has 5+ W/L and K/D ratios, the likes of Proton and Bows3r and there are tons of people like me who don't belong in tier 1. Basically, all people who can pull their weight are lumped together and we have 4 tiers for those who don't, it's bananas, it should be the other way around.


I agree with your broader point about how we need a more granular matchmaker to differentiate between skills at the top end. One thing I would point out though is that WLR and KDR is exaggerated at the top end by group queue. The numbers they are getting are just not sustainable in the solo queue on decently populated servers. This is not meant to diminish their ability, but all you have to do is look at the stats of various EMP players. Proton has one of the lower WLR on EMP, but he's clearly one of their better players. Why? Because he does a lot of solo queue.

If there was a solo queue leaderboard a lot of these guys with 5+ WLR would suddenly look mortal.

#77 N3tRunn3r

    Member

  • PipPip
  • Philanthropist
  • 28 posts
  • LocationTemple of Nod

Posted 02 June 2018 - 02:34 PM

View PostN3tRunn3r, on 02 June 2018 - 01:56 AM, said:


View PostPaul Inouye, on 30 May 2018 - 02:10 PM, said:


...

[color=orange]Blah blah blah.. numbers number numbers... what does this mean to me? Here's the TL;DR[/color]

* Wait times have been reduced between 40-50%.
* MAXIMUM average wait times have dropped from 150 seconds to 80 seconds.
* Matches being played are a lot more challenging.
* Average PSR ranking differences between two teams has gone from 950 to 200.

I've monitored this data over the last 3 days watching all 3 prime times (NA/EU/OC) and non-prime times and the data is consistent.

...

So that's it for now. I'll be keeping an eye on the wait times and I've made my own copy of the keys to the server config tool. Posted Image


Thank you very much by taking some time and effort into looking to this issue we had. I will test the Solo QP MM soon.

Posted Image


Game is still unplayable!!
Posted Image

#78 Humpday

    Member

  • PipPipPipPipPipPipPipPip
  • The Pharaoh
  • The Pharaoh
  • 1,462 posts

Posted 02 June 2018 - 03:53 PM

Nevermind I take it back, I like this better.

At first I was put put off because I was having a hard time pulling high numbers, but now...i realize most games played are mostly equally distributed amongst the best of the team.

What does throw me off though is that, people are far less aggressive and prefer to hang back as opposed get up in da face.

#79 Doomich

    Member

  • PipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 68 posts
  • LocationRussia

Posted 03 June 2018 - 03:12 AM

well, so .. IMO
for 3 days, every day I have 1-2 losestreaks (~6-8 games) with a score of 12 - 3/2/1
only 1 or 2 of those loses was realy intensive with high skill

I believe that Tier system doesn't work
because Tier 1 contains a lot of players with low skills who don't deserve to be there (solid statement huh Posted Image).
They are there only because they play a lot, and this system gives more than it takes.

Just make Elo based on player performance in the match, or something like that.
Anything will be better than the current Tier system. Now it looks like the system doesnt exist

In any case, I just altf4 ragequit after another one losestreak
and so far im not launch the game again for a while Posted Image
... *tinkering around in mechlab* Posted Image

Edited by Doomich, 03 June 2018 - 03:29 AM.


#80 Brain Cancer

    Member

  • PipPipPipPipPipPipPipPipPip
  • The 1 Percent
  • 3,851 posts

Posted 03 June 2018 - 08:40 AM

That's because the illusion of skill division died a horrible, slow death as more and more people capable of getting positive PSR crawled up into T3, T2, and then T1.

Nothing really will push them out of those Tier rankings. For all intents and purposes, we've almost eliminated any division by skill in MWO, which in turn leads to more team frustration and negative effects on player population.

I mean, you could also tighten up PSR +/-/nochange requirements as Tier goes up and cut down the massive positive bias to allow people to shake out more, but given MWO's apparent attempt to sacrifice skill division for "biggest pool of players possible", it's unlikely even if a positive in the long run. Done right, you might actually see truly unskilled players slide back down into T4 and hence out of the T1-3 shark tank.





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users


  • Facebook