Jump to content

Paul Brings Clarification To Psr And Tiers.


277 replies to this topic

#241 An Atlas

    Member

  • PipPipPip
  • 51 posts

Posted 25 September 2015 - 09:08 AM

View PostCaptRosha, on 25 September 2015 - 09:05 AM, said:

The point is it doesn't matter WHAT tier the other team was....


AGAIN.... is it just as easy to win against a Team of all Tier 1 as against a team of all Tier 5?

Of course not.

If you can't win against the Tier you're playing against..... you don't generally go up.....

Tier is LITERALLY the most basic requirement of the system.... BY REQUIRING YOU TO WIN AGAINST THEM.

#242 Johnny Z

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Knight Errant
  • 9,942 posts
  • LocationDueling on Solaris

Posted 25 September 2015 - 09:11 AM

View PostMechWarrior1086091, on 25 September 2015 - 09:08 AM, said:



AGAIN.... is it just as easy to win against a Team of all Tier 1 as against a team of all Tier 5?

Of course not.

If you can't win against the Tier you're playing against..... you don't generally go up.....

Tier is LITERALLY the most basic requirement of the system.... BY REQUIRING YOU TO WIN AGAINST THEM.


I heard its possible to go down on a win also. Also winning isnt a requirment just a good score. Although i agree with your sentiment 100%.

#243 Kira Onime

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Dragoon
  • The Dragoon
  • 2,486 posts
  • Twitch: Link
  • LocationMontréal, Québec.

Posted 25 September 2015 - 09:12 AM

View PostMechWarrior1086091, on 25 September 2015 - 08:58 AM, said:

-snip-



You seem unable to address anything that concerns group drops.

#244 PurpleNinja

    Member

  • PipPipPipPipPipPipPipPipPip
  • Liquid Metal
  • Liquid Metal
  • 2,097 posts
  • LocationMIA

Posted 25 September 2015 - 09:13 AM

View PostJohnny Z, on 25 September 2015 - 09:07 AM, said:

The regular queue matches have been alot better with this tiers in place. At least twice as good. You in the fixed mach group queue mostly?

Solo queue 99% of the time. I just haven´t noticed any change on matches after PSR.

#245 Ghogiel

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • CS 2021 Gold Champ
  • CS 2021 Gold Champ
  • 6,852 posts

Posted 25 September 2015 - 09:13 AM

View PostCaptRosha, on 25 September 2015 - 09:05 AM, said:

just that you won

you won't get a positive W/L in solo if you are actually bad at the game. But that's about as close as PSR gets to balancing the MM lol

#246 Gnume

    Member

  • PipPipPipPipPipPip
  • Bridesmaid
  • Bridesmaid
  • 279 posts
  • Facebook: Link
  • LocationPrattville, AL

Posted 25 September 2015 - 09:17 AM

View PostMechWarrior1086091, on 25 September 2015 - 09:05 AM, said:

The real problem here is people don't like that the PSR system judges you as a team player....b/c the results show how much they suck.

Everyone wants to it to just rate how much of a KDR ***** you are instead.

I want it to place me in matches that are a challenge for my "skill". I want it to calculate my skill based on how I perform in match. Count points for doing the things I do, Protected Light, ECM, Counter ECM, Savior Kills, and all those other things. Add in things like defending a capture point or base...if I did damage to a mech that was within one of those areas then I should get additional credit for defending it. All those TEAM BASED things I do as an individual should count whether my Team wins the match or not. My PSR should not go down because the others performed poorly and contributed to the loss.

I think the problem is that PGI wants a way for Tier's to fluctuate and for player's PSR scores to be able to fluctuate as well. The way they have decided to do that is via the W/L portion of the game. That is not a reflection on individual ability and not even a reflection as to whether an individual is a Team Player or not the way it is currently implemented. It's just an arbitrary measure for the PSR to fluctuate.

I think a better way to keep it fluctuating is to just take the average score of your most recent X number of matches. However, the issue may then become one where players purposefully tank some games to lower their current PSR and with it using a smaller sample size than your lifetime matches, then that player can tank it faster and raise it faster. They could combat that by only allowing a tier reduction of 1 (from the max tier you ever obtained) so that a T1 player could only ever drop to 2 and a 2 to a 3 and so on.

Edited by CaptRosha, 25 September 2015 - 09:21 AM.


#247 An Atlas

    Member

  • PipPipPip
  • 51 posts

Posted 25 September 2015 - 09:20 AM

View PostKira Onime, on 25 September 2015 - 09:12 AM, said:



You seem unable to address anything that concerns group drops.


Group drop doesn't use PSR to create matches.....nullifying it's relevence to this discussion completely.

PSR was designed to measure a single players ability to work with a team of random players to win against another team of random players that perform on their level as a team.

Now on that note, should PGI let any win / losses / match scores from group Q affect PSR if it's not going to be used to balance the games?

No, I think that's pretty silly of them.

View PostCaptRosha, on 25 September 2015 - 09:17 AM, said:

I want it to place me in matches that are a challenge for my "skill". I want it to calculate my skill based on how I perform in match. Count points for doing the things I do, Protected Light, ECM, Counter ECM, Savior Kills, and all those other things. Add in things like defending a capture point or base...


It currently does all of these things.

Generally I only do 400 ish damage, with a kill or two....

But I spot, capture, tag, narc, hold locks.... all that team player stuff.

And guess what.... It "ranks" me high for it.... b/c those things let my team win more easily... consistently.

Am I mad that the end of match screen doesn't say "YOU'RE THE BEST CAPTURE PLAYER EVER MATCH SCORE 1000 FOR SPOTTING"?

No... I'm happy that it says "VICTORY"

Edited by MechWarrior1086091, 25 September 2015 - 09:21 AM.


#248 Kira Onime

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Dragoon
  • The Dragoon
  • 2,486 posts
  • Twitch: Link
  • LocationMontréal, Québec.

Posted 25 September 2015 - 09:21 AM

View PostMechWarrior1086091, on 25 September 2015 - 09:18 AM, said:

Group drop doesn't use PSR to create matches.....nullifying it's relevence to this discussion completely.


Groups using PSR or not doesn't change the fact that it will influence your PSR anyways.
Play with a good group that can carry you and you're bound to go up by the nature of the system that you cannot go down in a win.

In turn, what would be a low tier player according to the system can be carried up to T1.

Also, from my general understanding, group queue does use PSR to match people but it takes the average PSR of the group to do so.

#249 kapusta11

    Member

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

Posted 25 September 2015 - 09:29 AM

Hm... so it's still based largely on w/l, some kind of Elo mutant.

#250 Gnume

    Member

  • PipPipPipPipPipPip
  • Bridesmaid
  • Bridesmaid
  • 279 posts
  • Facebook: Link
  • LocationPrattville, AL

Posted 25 September 2015 - 09:32 AM

View PostMechWarrior1086091, on 25 September 2015 - 09:20 AM, said:


Group drop doesn't use PSR to create matches.....nullifying it's relevence to this discussion completely.

PSR was designed to measure a single players ability to work with a team of random players to win against another team of random players that perform on their level as a team.

Now on that note, should PGI let any win / losses / match scores from group Q affect PSR if it's not going to be used to balance the games?

No, I think that's pretty silly of them.



It currently does all of these things.

Generally I only do 400 ish damage, with a kill or two....

But I spot, capture, tag, narc, hold locks.... all that team player stuff.

And guess what.... It "ranks" me high for it.... b/c those things let my team win more easily... consistently.

Am I mad that the end of match screen doesn't say "YOU'RE THE BEST CAPTURE PLAYER EVER MATCH SCORE 1000 FOR SPOTTING"?

No... I'm happy that it says "VICTORY"

It does, but the other factor involved in calculating that score is outside my control and skill ... whether my team won or lost. I could have done the most awesome job of team playing ever, racked up all sorts of points for doing team based things. and then we lose. I did earn a Exceptionally high match score because I did all those things, yet I am rewarded with a "Small Increase" in my PSR because the others on my team sucked. That's not rewarding...that's still a punishment in the form of an "admonishment" instead of a spanking :P

#251 Lightfoot

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 6,612 posts
  • LocationOlympus Mons

Posted 25 September 2015 - 09:32 AM

Not really false, but I used to run LRM Support mechs with TAG and BAP for the sole purpose of helping my team to win since I don't really like the role, but I do it well. It worked well up until the first patch of January 2015 at which time LRMs were Bugged or stealth-nerfed. After that it became like the LRMs were just bouncing off mechs. I recall one time there were 5 LRM mechs all firing on one lone Warhawk with no AMS or any friendly mech nearby and it just strolls across the open end of old River City taking minor damage to one arm from 1200 to 1500 LRM hits.

Then when the Territorial Challenge came I decided I would try to defeat the LRM nerf by making the best team support LRM mech I knew how. I got one match over 200 from this mech despite Tagging and damaging 7 to 9 mechs per match. So I switched to direct-fire and posted 300-600 scores in those matches.

LRMs don't get you Kills, they get you Assists, Component Damage, Flanking, Tagging, Scouting, even some Brawling. So it's not working. At least for MechWarrior's number one team support role which is LRMs. I am feeling that old Missile-Hate that is so common in MechWarrior games. Un-nerf the LRMs I guess because players will believe they will move their PSR up with good LRM team support when most likely they will just lose badly to direct-fire and get those double PSR drops.

#252 An Atlas

    Member

  • PipPipPip
  • 51 posts

Posted 25 September 2015 - 09:32 AM

View PostKira Onime, on 25 September 2015 - 09:21 AM, said:

Groups using PSR or not doesn't change the fact that it will influence your PSR anyways.

Play with a good group that can carry you and you're bound to go up by the nature of the system that you cannot go down in a win.


Wow you really don't ever read anything before responding do you.... FROM THE EXACT POST YOU QUOTED to argue this point about.

View PostMechWarrior1086091, on 25 September 2015 - 09:20 AM, said:

Now on that note, should PGI let any win / losses / match scores from group Q affect PSR if it's not going to be used to balance the games?

No, I think that's pretty silly of them.


As far as the "Group games balanced by PSR".... I feel like Sean mentioned it isn't use in CW.

But if it DOES average a groups Tier to balance group PUG matches then by definition it's working.

If a T1 and a T5 guy faced off against two T3 (their average), I'm betting it'd be a pretty close game....

Edit: Also... you mentioned "can't go down on a win" which doesn't really mean anything. If a T5 grouped with a bunch of T1s and went into a T1 match.... sure the team might win... but that T5 guy is going to have the floor wiped with him.

Will he go down? No.

Will he perform well enough to go up? I seriously doubt it.

Even if he DID go up....would he go right back down once the T1 guys were gone and he was being measured by his INDIVIDUAL skill as a team mate again? Yep.

That's what we want right? INDIVIDUAL measurements.

Edited by MechWarrior1086091, 25 September 2015 - 09:51 AM.


#253 Kira Onime

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Dragoon
  • The Dragoon
  • 2,486 posts
  • Twitch: Link
  • LocationMontréal, Québec.

Posted 25 September 2015 - 09:53 AM

View PostMechWarrior1086091, on 25 September 2015 - 09:32 AM, said:

As far as the "Group games balanced by PSR".... I feel like Sean mentioned it isn't use in CW.


CW doesn't have any MM balance what-so-ever. It finds 12 players for both teams and pitches them against each other.
12 pugs vs a 12man comp team? It doesn't care.



View PostMechWarrior1086091, on 25 September 2015 - 09:32 AM, said:

But if it DOES average a groups Tier to balance group PUG matches then by definition it's working.
If a T1 and a T5 guy faced off against two T3 (their average), I'm betting it'd be a pretty close game....


On paper the groups are balanced. In-game is a completely different story.

#254 Chuck Jager

    Member

  • PipPipPipPipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 2,031 posts

Posted 25 September 2015 - 09:54 AM

View PostMechWarrior1086091, on 25 September 2015 - 09:32 AM, said:

If a T1 and a T5 guy faced off against two T3 (their average), I'm betting it'd be a pretty close game....


I have noticed the group with the players closer to the average (all t3) do better than an average of a wide spread (t1 avg w t3). The playstyles and builds are closer together.

Edited by Chuck YeaGurr, 25 September 2015 - 10:21 AM.


#255 crashlogic

    Member

  • PipPipPipPipPipPip
  • Mercenary
  • 318 posts

Posted 25 September 2015 - 09:58 AM

Tames get kills, kills equal victory. The whole point of the game, if i am to believe PGI, is that teamwork equals victory (i agree with this too) If I am on a bad team, then my psr is going down, even if i do 500 damage. If I am on a good team my PSR is going up, even if I do 120 damage. Or more pertinentlay as a tier 3 player ona tier five team it does not matter what my inidividual skill level is, I am not going to improve. As a tier 3 player on a tier one team, as long as the team wins, I will go up a tier. Thus the team matters far more than individual production. Hence my fundamental flaw with PSR--its a team game, and PSR is an individual accomplishment. Unlesss the whole team is on an equal level, player improvement will not really help.
Or put it this way, its entirely possible to remain a tier 4 player doing 700 damge every time and never going up because your team can't get more the 4-5 kills

#256 Kira Onime

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Dragoon
  • The Dragoon
  • 2,486 posts
  • Twitch: Link
  • LocationMontréal, Québec.

Posted 25 September 2015 - 09:58 AM

View PostMechWarrior1086091, on 25 September 2015 - 09:32 AM, said:

a T5 guy faced off against two T3 (their average), I'm betting it'd be a pretty close game....

Edit: Also... you mentioned "can't go down on a win" which doesn't really mean anything. If a T5 grouped with a bunch of T1s and went into a T1 match.... sure the team might win... but that T5 guy is going to have the floor wiped with him.

Will he go down? No.

Will he perform well enough to go up? I seriously doubt it.




You'd be surprised what a little bit of shot calling can do to some people.


Also out of curiosity, why are you using an alt account to post?

#257 Tesunie

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The Seeker
  • The Seeker
  • 8,586 posts
  • LocationSeraphim HQ: Asuncion

Posted 25 September 2015 - 09:59 AM

View PostFenrisulvyn, on 24 September 2015 - 11:02 PM, said:

"Basing any kind of "skill" ranking system on "wins" in a "team" game, where you have 23 other players who determine your win/lose fates... Is not a good idea."

Yup. And I thought we already learned this during the recent kill/assist/win challenge. People were screaming at their team because they had to get a win to advance.

I'll be curious how this affects the Light Mech que. It seem obvious that the best way to ward off against red down arrows is to put high damage numbers on the board, which is tough in non-brawly lights. The dedicated capper role for Conquest is dead. The dedicated Narc light is dead. The scout role is now handicapped because movement to contact often means you stumble into a red team focus fire early in match and die with little points.

The information warfare scoring (scouting, spotting, tag/narc, counter ecm etc) need to be significantly increased.

I could be wrong - the light que is 8-11% right now. But that's a number I think PGI should keep an eye on.


Actually... I've been raising my Tier with light mechs... That and doing some things that I don't feel are exactly helpful to the team as much as other tactics. (Flank, snipe from afar, hide and run when someone shoots at me, etc.) Though I do my best to help the team still, because I am all about teamwork, a more selfish playing style is increasing my rank. (Preformed mostly as a test.)

The system currently doesn't support team play, and instead supports "play to get my own individual best performance, screw the team", especially on a win.

The win is too greatly focused on, as that's team cohesion sometimes or just blind luck other times.
Damage is too greatly focused on, as that does win games but doesn't really promote actual team working efforts. Damage seems to be half your match score on it's own, plus assists, kills, components destroyed, etc.
Team efforts and selfless play seems to not be rewarded enough. No one wants to be the point of a charge, and it's not rewarded anyway (kinda hard to reward this one, I'll admit). Capturing objectives in an objective game mode is also not really rewarded, as I can capture all the points, win the match by doing so, and have a low match score. Or even die trying to capture the points, lose, and have an even lower match score.


PSR I feel has a good base. It needs some improvements to reward other forms of play. It also really should be disassociated with winning and losing, as that isn't a "Player" skill ranking system. (And yes, I know this is a team game as I say most of this.)

View PostSarlic, on 25 September 2015 - 12:50 AM, said:

In my opinion this system has changed nothing at all and is just a XP bar and it's nonesense.(and not independence on teamwork, how is anything teamwork if you get rolled? The solo que is still brutally independence.)


It isn't an XP bar, as it can go down (and should once you are in the tier you should more or less be in). They have a good base, but things need to really change for it to work as intended. Read my above statements.


View PostDeRazer, on 25 September 2015 - 01:35 AM, said:

3. Set the threshold for Match Score going up and down by the Tier you are on.

Eg.
Tier 5 - MS to Go Up = 200, MS to Go Down = <100
Tier 4 - MS to Go Up = 250, MS to Go Down = <125
Tier 3 - MS to Go Up = 300, MS to Go Down = <150
Tier 2 - MS to Go Up = 400, MS to Go Down = <250
Tier 1 - MS to Go Up = 550, MS to Go Down = <400


I would like to point out a flaw in this... There is only so much damage a player can get in a match, which does limit the match score as well.

As you go up, in theory your opponents should get harder. This means that it should become harder to maintain the same match score. Match score for an increase to PSR should remain the same, as it should get harder to keep that level if you are "over tiered", and easier to get it if you are "under tiered".

If anything, the values of PSR numbers could shift based on tier, but I don't believe this would be good either. The concept with a PSR tier is skill level and matching someone up to their appropriate challengers/teammates. It isn't suppose to be a struggle to stay where you should be because it got nearly impossibly more difficult (with your proposed system), but neither should it be a breeze to get to where you probably don't belong either. Moving in and out of a tier should be fluid like, where you move as your skills indicate. Eventually, you should settle inside one tier over the long course of the game, and stay there for the most part.


Basically, the tier system shouldn't be a progression thing, and it shouldn't get increasingly more difficult to stay in a tier that is/may be appropriate for your actual skill. If anything, the system should work at trying to find an even tier for you, with ways to go up and down to keep you steady if you are appropriately placed. (I hope I'm making sense here.)



PS: Sorry, I couldn't read the 7+ pages between my responses... :blush:


Edit: I'm no longer watching this thread. It's exploded with activity. Quote me if you want a response from me. (Otherwise, I wont know.)

Edited by Tesunie, 25 September 2015 - 10:01 AM.


#258 An Atlas

    Member

  • PipPipPip
  • 51 posts

Posted 25 September 2015 - 10:06 AM

View PostKira Onime, on 25 September 2015 - 09:58 AM, said:



You'd be surprised what a little bit of shot calling can do to some people.


Also out of curiosity, why are you using an alt account to post?


To avoid trolls

#259 Widowmaker1981

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The Widow Maker
  • The Widow Maker
  • 5,031 posts
  • LocationAt the other end of the pretty lights.

Posted 25 September 2015 - 10:07 AM

View PostMechWarrior1086091, on 25 September 2015 - 09:05 AM, said:

The real problem here is people don't like that the PSR system judges you as a team player....b/c the results show how much they suck.

Everyone wants to it to just rate how much of a KDR ***** you are instead.


NO. I am maxed out Tier 1, according to this system im awesome. Im not complaining because its making me look bad or something, im compaining because ive used my brain to look at and understand how it works, and its completely stupid and only even vaguely useful for keeping new players away from vets, mainly because it is horribly polluted by group queue results, much like Elo was. When you can the control the variable 'other players on my team' your win/loss has FAR more to do with that than individual skill.

I would like a matchmaking system that functions to match players with similalrly skilled players. Due to its interaction with the group queue, this one does not.

The solution is simple. Remove the win/loss component of PSR change, at least for group queue matches.

Edited by Widowmaker1981, 25 September 2015 - 10:09 AM.


#260 Kira Onime

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Dragoon
  • The Dragoon
  • 2,486 posts
  • Twitch: Link
  • LocationMontréal, Québec.

Posted 25 September 2015 - 10:09 AM

View PostMechWarrior1086091, on 25 September 2015 - 10:06 AM, said:


To avoid trolls



Usually using alt accounts to post things generally paint you as what you're trying to avoid.





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users