80 replies to this topic

[Haipyng]

I usually find the solo MM to be far and away better than the group MM. I get odd streaks, like several wins or losses in a row. The occasional outright stomp.

Group MM with a small 2 to 3 man group results in near consistent losses with several 12-3 or worse stomps. Running in a 7+ man usually balances that out.

On Group MM; agreed it sucks, it hasn't gotten the attention it needs. On Solo MM...it's perhaps as close as they can make it using ELO. ELO is not a good system for team based games.

Just to throw in here before the Tryhards show up:

Haipyng, on 16 January 2015 - 08:36 AM, said:

This is a well worn subject. Everyone knows 2-4 man groups are borked in the group MM. There is at least a small chance you will get matched with another team of 2-4 man groups for a competitive game. That is the best there is for the foreseeable future.

You will find some very vocal people on here that won't be sympathetic, with stock non-answers that they play like old favorites, such as:

• Learn to play (otherwise known as L2P)
  
• Join big groups
  
• Don't play
  
• There is no problem with small groups
  
• Stop whining

...all sung in a Frank Sinatra voice.



They won't acknowledge the issue because they play in large groups, the problem doesn't affect them, they like the GGClose mechanic to farm with, or are just plain old sociopathic, unable to empathize with anyone else. Ideas (and I mean any ideas) that are put forth as ways to fix this won't work, are terribad, or will break the entire game including CW, the patcher and game launcher.

I totally understand where you are coming from, but we have been here before and we will be again. Just thought I would condense it down for the sake of brevity.

[Duke Nedo]

TELEFORCE, on 21 January 2015 - 03:27 PM, said:

In all seriousness, I'd be interested in seeing how the matchmaker would work if it were based the average of a player's match scores instead of wins and losses with Elo (it's the name of the scoring system's creator, hence the comment above).

Exactly this... I'd bet the average of the new rewards2.0 is a much better measure of player skill than Elo currently is. I am guessing the MM does the best it can matching Elo values, that is simple enough, but what good is that if the Elo values have little prediction value?

After all it's a 12-12 game, different compositions every time of skills, mechs, loadouts, ecm, map, map side, disconnects, etc etc. From that mess, it's supposed to fish out your individual contribution to drive a win. Good luck... and if that wasn't enough, you can build/destroy Elo in the group queue with a good/bad group and then use that Elo value when you drop in the solo queue. So... Elo is bound to be rather... random?

[Willard Phule]

Korak18, on 21 January 2015 - 01:54 PM, said:

I know it has been brought up before ,again and again, but I for one are really getting fed up with the MatchMaker in Pug's.
I don't mind a mix of players but it seems to be getting worse.
Last 5 -6 match's I had at least 7- 8 rookies..brutal to say the least.
Don't mind losing but we don't stand a chance from the start, waste of time.
I do try to help but we all know the story...no response ,no action...not enough experience etc
Yes I know all the reasons why it does not work (the mysterious ELO) and please don't tell me or anybody else to just play with my mates. They are not always on and sometimes I just prefer to go alone.
I have noticed a dearth of responses from PGI on this...may have missed some stuff but basically they have been pretty silent .
I know ROFL pugfests but come on.... this is beyond stupid...
It is frustrating enough for me to consider something else for my time and money...
I know the diehards will not care...go CW, but I am sure you have lost people because of this.
They quietly slip away.
It needs fixing to say the least.

/Rant off /

Keep in mind that you're OBVIOUSLY referring to the solo queue. The re-write of the matchmaker using team based Elo averages means that the better you are, the more you're expected to carry.

Egomane, on 22 January 2015 - 04:48 AM, said:

If you really want to know how the matchmaker works, instead of playing the blame game based on wild guesses, support my suggestion here: http://mwomercs.com/...sms-matchmaker/

You know, you put that thread up a LONG time ago and Karl STILL hasn't responded.

There is no comprehensive information with regard to how MM and Elo works other than random posts from over a year ago.

Until Karl speaks up, those of us that actually play the game are left with no other position than to say that in the solo queue, it truly feels like the better you are, the more new players you're expected to carry.

[Egomane]

Willard Phule, on 22 January 2015 - 06:13 AM, said:

You know, you put that thread up a LONG time ago and Karl STILL hasn't responded.

No, he didn't! But then again, the suggestion didn't seem to generate much support from the community either, so there is no pressure to respond. I'm not one who is going around in every thread and advertises his ideas every chance he gets. If I had done so, I'm sure I would have had posted that link at least a hundred times in different MM related threads, since I created the suggestion.

There are so many rants about the MM, but a suggestion to clear up all the misconception gets ignored, even by those who complain. Fine by me, but then why are you complaining, if you don't want an answer to your complaint?

Edited by Egomane, 22 January 2015 - 06:41 AM.

[LordSkippy]

TELEFORCE, on 21 January 2015 - 03:27 PM, said:

Mysterious ELO? I didn't know that Electronic Light Orchestra was that mysterious! Then again I've often wondered how they made their instruments talk

In all seriousness, I'd be interested in seeing how the matchmaker would work if it were based the average of a player's match scores instead of wins and losses with Elo (it's the name of the scoring system's creator, hence the comment above).

As soon as someone starts to assert how they believe "ELO" works, I start to ignore them. How much can they know if they can't even get the name right.

The biggest problem with Elo in MWO is that Elo was developed for a non-team competition, Chess. Basing player Elo ratings on a team based stat (win/loss of a team match) means there are too many factors outside of the player's skill that can affect the outcome and alter their rating. In other words, someone else having a bad game can negatively impact your Elo rating, as well as someone else having a great game can artificially raise your Elo rating.

Basing the MM off of something like match score would, in theory, be better, because it is closer to individual contribution. Teammate performance will still be a factor, but less so than the current system and ignoring the win bonus portion of the match score would help with minimizing other players impacts on your rating. Biggest issue is whether or not match score accurately describes a players contribution at the role they were performing. But that gets into whether or not role warfare is rewarded appropriately.

[mogs01gt]

What I hate worst about MM and how we level mechs is that we are stuck using a mech variant that we probably dont want. So we arent going to invest into upgrades. Now we have to go into MM and play against people with a stock mech that utterly sucks. I always feel like Im bringing the team down with these stock mechs....

[Trashhead]

Korak18, on 21 January 2015 - 01:54 PM, said:

(..)
I. do try to help but we all know the story...no response ,no action...not enough experience etc
(...)
/Rant off /

Off topic, but related question:
Are the chat channel currently all set to OFF by default ?
Or just the ALL chat channel ?

Anyone knows ?

[Joseph Mallan]

What I like most about Match Maker is in Open Que its broken and isn't worth the code its wrote on.

In CW it will be the salvation of teh Que.

Ain't Life grand?

[Willard Phule]

Egomane, on 22 January 2015 - 06:40 AM, said:

No, he didn't! But then again, the suggestion didn't seem to generate much support from the community either, so there is no pressure to respond. I'm not one who is going around in every thread and advertises his ideas every chance he gets. If I had done so, I'm sure I would have had posted that link at least a hundred times in different MM related threads, since I created the suggestion.

There are so many rants about the MM, but a suggestion to clear up all the misconception gets ignored, even by those who complain. Fine by me, but then why are you complaining, if you don't want an answer to your complaint?

You misunderstand. I'm not complaining. The solo queue is what it is. It's been this way for a long, long time. The vast majority of complaints with regard to the solo queue matchmaking seem to be coming from either new players that failed to take the all-inclusive tutorial...you know, the whole new player experience thing....or people coming over from the Group Queue that are astounded that the people in the Solo don't play like they do.

Me? I expect at least 30% new players and/or trial mechs per match, if not more. It is what it is, man.

[LordSkippy]

Joseph Mallan, on 22 January 2015 - 06:49 AM, said:

What I like most about Match Maker is in Open Que its broken and isn't worth the code its wrote on.

In CW it will be the salvation of teh Que.

Ain't Life grand?

I like how it's the worst possible match making system, yet still better than six months ago. When it was the worst possible match making system, yet still better than six months before that. When it was the worst possible match making system.

I'm making a new law; Skippy's laws: The current match making system is always the worst possible match making system.

[SgtKinCaiD]

@CocoaJin

We're not talking here about a slow landslide where a team is slowly taking the advantage over the other one.
We're talking about losing several teammates in the first 3 minutes of the match due to bad positionning / bad builds, in short a lack of basic knowledge of the game.

[Willard Phule]

SgtKinCaiD, on 22 January 2015 - 07:12 AM, said:

@CocoaJin

We're not talking here about a slow landslide where a team is slowly taking the advantage over the other one.
We're talking about losing several teammates in the first 3 minutes of the match due to bad positionning / bad builds, in short a lack of basic knowledge of the game.

See also: New players and Trial Mechs.

[Joseph Mallan]

SgtKinCaiD, on 22 January 2015 - 07:12 AM, said:

@CocoaJin

We're not talking here about a slow landslide where a team is slowly taking the advantage over the other one.
We're talking about losing several teammates in the first 3 minutes of the match due to bad positionning / bad builds, in short a lack of basic knowledge of the game.

I'm sorry but stupid/Unlucky players die quickly.(Reference: Me in Commandos)

It cannot determine if the build is bad and if its bad for you, is it bad for the user in question? For all you know, the last 3 matches, that PoS got 800 damage for the user!

Edited by Joseph Mallan, 22 January 2015 - 07:19 AM.

[Thorqemada]

The MM gives and takes - one match is Topguns vs Hotshots and next match is Distress vs Misery.

[Joseph Mallan]

Thorqemada, on 22 January 2015 - 07:18 AM, said:

The MM gives and takes - one match is Topguns vs Hotshots and next match is Distress vs Misery.

So working as intended. You win some you lose some...

[SgtKinCaiD]

Seriously Mallan i wish there was an ignore function on this forum just for you.

Edited by SgtKinCaiD, 22 January 2015 - 07:54 AM.

[Egomane]

SgtKinCaiD, on 22 January 2015 - 07:54 AM, said:

Seriously Mallan i wish there was an ignore function on this forum just for you.

There is (Forum Profile -> Edit my Profile -> Ignore Prefernces):
http://mwomercs.com/...ea=ignoredusers

[DaZur]

Here is where folks need to get their heads straight over their Elo and the MM... The MM is matching the composite Elo score of a team NOT the individual players.

Or in other words for the purpose of match-making the MM treats a composite team Elo score like a single player. As such Elo works perfectly well in terms of pairing "teams".

Once the MM pairs the teams then all bets are off... The result of the match outcome is wholly dependent upon functional math and compounded force strength and the functional skill of the involved players and their ability to play as a team.

We have to get over this MM boogieman crap and start thinking of ways to better quantify a players functional skill level and ability to "team-play"

Win/lose is a poor skill metric for an individual player because it's dependent upon a "team" and has little to do with quantifying their individual ability to contribute.

[Mister Blastman]

Deathlike, on 22 January 2015 - 01:27 AM, said:

At this point, I don't think the population and "Elo buckets" would support this anymore...

Had so many longstanding issues been addressed... we'd probably not be in such a rut.

Has it gotten worse in the last month or so?

[Joseph Mallan]

SgtKinCaiD, on 22 January 2015 - 07:54 AM, said:

Seriously Mallan i wish there was an ignore function on this forum just for you.

Egomane, on 22 January 2015 - 07:57 AM, said:

There is (Forum Profile -> Edit my Profile -> Ignore Prefernces):
http://mwomercs.com/...ea=ignoredusers

There ya go.