Jump to content

Psr Is Personal Skill Rating - It Shouldn't Depend On Team Performance


53 replies to this topic

#1 MrMadguy

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,288 posts

Posted 25 September 2015 - 09:19 AM

So, PSR means PERSONAL skill rating. The whole purpose of changing ELO into PSR was that new rating won't depend on team performance in any way, as in solo queue your and enemy's teams are nothing more, then just a bunches of randoms with random skills, so team's performance is just random unpredictable value too. And first of all, it should depend on your W/L, as W/L is team-based ratio - not personal one. But what we see now? We see this:
Posted Image
So your PSR doesn't depend only on your skill - it depends on whether your team wins a match or loses it. How could PSR be called PERSONAL skill rating, if it depends on W/L at the end? So, PGI, make PSR true PERSONAL skill rating, otherwise I will play stock variants only till I won't drop to Tier 5. And believe me, my team mates will be extremely happy to play with pilot in stock 'Mech. If you believe, that only new players should be in Tiers 4 and 5, then I'll pretend, that I'm newbie, lol. If you can't understand, that players' skill can't just magically improve after N matches played, then may be my team mates' cries will convince you to change your mind.

Edited by MrMadguy, 25 September 2015 - 09:20 AM.


#2 Ghogiel

    Member

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

Posted 25 September 2015 - 09:25 AM

Let me get this straight... So by plaing badly you think you can influence your w/l badly enough to drop to T5... but think there is no personal skill at possitively influencing the outcome of a match so as to get to T1?

#3 Christof Romulus

    Member

  • PipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 898 posts
  • LocationAS7-D(F), GRF-1N(P)

Posted 25 September 2015 - 09:28 AM

View PostGhogiel, on 25 September 2015 - 09:25 AM, said:

Let me get this straight... So by plaing badly you think you can influence your w/l badly enough to drop to T5... but think there is no personal skill at possitively influencing the outcome of a match so as to get to T1?

That's what I read, yup.

#4 Apnu

    Member

  • PipPipPipPipPipPipPipPipPip
  • Elite Founder
  • Elite Founder
  • 2,083 posts
  • LocationMidWest

Posted 25 September 2015 - 09:33 AM

View PostMrMadguy, on 25 September 2015 - 09:19 AM, said:

So, PSR means PERSONAL skill rating. The whole purpose of changing ELO into PSR was that new rating won't depend on team performance in any way, as in solo queue your and enemy's teams are nothing more, then just a bunches of randoms with random skills, so team's performance is just random unpredictable value too. And first of all, it should depend on your W/L, as W/L is team-based ratio - not personal one. But what we see now? We see this:
Posted Image
So your PSR doesn't depend only on your skill - it depends on whether your team wins a match or loses it. How could PSR be called PERSONAL skill rating, if it depends on W/L at the end? So, PGI, make PSR true PERSONAL skill rating, otherwise I will play stock variants only till I won't drop to Tier 5. And believe me, my team mates will be extremely happy to play with pilot in stock 'Mech. If you believe, that only new players should be in Tiers 4 and 5, then I'll pretend, that I'm newbie, lol. If you can't understand, that players' skill can't just magically improve after N matches played, then may be my team mates' cries will convince you to change your mind.



PSSSST: PSR is Elo. At least in terms of how PGI considered Elo. Meaning a win and it goes up, a loss and it goes down. The only change is they're showing you your rating, calling it "PSR" instead of Elo (marketing decision probably) and showing you how your score affects PSR changes.

What you've posted above is talking about how much up/down movement depending on the players personal contribution to the team's actions.

Last night I had two losses, but carried hard in a un-basic'd BK. I had 500 damage, and was in the top three of my team in those two games. My PSR rating showed "=" at the end of the round. Because we lost, but I put up good enough numbers to mitigate the down change I should have received.

That's it.

#5 Davegt27

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 7,041 posts
  • LocationCO

Posted 25 September 2015 - 09:35 AM

Quote

So, PSR means PERSONAL skill rating. The whole purpose of changing ELO into PSR was that new rating won't depend on team performance in any way


the reason for the change was people crying on the forums

hope the game makes it

Edited by Davegt27, 25 September 2015 - 09:35 AM.


#6 Death Proof

    Member

  • PipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 546 posts

Posted 25 September 2015 - 09:38 AM

PSR stands for PILOT SKILL RATING.

But yeah, I don't think it should be based so much on W/L like currently is. It'd should be based more heavily on match score.

Edited by Death Proof, 25 September 2015 - 09:49 AM.


#7 kapusta11

    Member

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

Posted 25 September 2015 - 09:38 AM

Isn't it player skill rating?

Edit: Oh great, now I look like a fool.

Edit 2: Wait, it's actually pilot skill rating.

Edited by kapusta11, 25 September 2015 - 09:40 AM.


#8 Bluttrunken

    Member

  • PipPipPipPipPipPipPip
  • The Patron Saint
  • The Patron Saint
  • 830 posts

Posted 25 September 2015 - 09:41 AM

Winning or losing is a random component. You can't take this into an equation to rate the player's skills nor can't you use it as an indicator of the applied teamwork of the player.

There are 23 other variables going into this and you would have to compare all individual performances(broken into specific team-related scores) to get a tendency of the player's liability of the match outcome.

#9 MrMadguy

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,288 posts

Posted 25 September 2015 - 09:43 AM

View PostApnu, on 25 September 2015 - 09:33 AM, said:



PSSSST: PSR is Elo. At least in terms of how PGI considered Elo. Meaning a win and it goes up, a loss and it goes down. The only change is they're showing you your rating, calling it "PSR" instead of Elo (marketing decision probably) and showing you how your score affects PSR changes.

What you've posted above is talking about how much up/down movement depending on the players personal contribution to the team's actions.

Last night I had two losses, but carried hard in a un-basic'd BK. I had 500 damage, and was in the top three of my team in those two games. My PSR rating showed "=" at the end of the round. Because we lost, but I put up good enough numbers to mitigate the down change I should have received.

That's it.

I feel that way too: they actually haven't changed anything - just it's name. It still depends on W/L, as ELO did. That's what I'm talking about - I WANT REAL IMPROVEMENTS, not just an illusion.

P.S. About stock variants - it isn't joke. I have about 10-15 'Mechs, I wasn't able to lvlup due to MM offering way to much stomps to me. I wanted to upgrade them first, but PGI doesn't give me another option, other then just play all this stock 'Mechs for ages. If I'm going to be stomped anyway, then what is the difference?

Edited by MrMadguy, 25 September 2015 - 09:44 AM.


#10 Felio

    Member

  • PipPipPipPipPipPipPipPip
  • Philanthropist
  • Philanthropist
  • 1,721 posts

Posted 25 September 2015 - 09:45 AM

Mostly what they did was stop your Elo from going up when your team wins in spite of you sucking eggs. Which is a worthwhile change, true.

They also let it go up or down by a variable amount, rather than everyone on the team rising or falling by exactly the same amount.

So they added some mobility to it, which I think was the main goal. Our Elos were basically static before, Russ seemed to suggest. Could they go farther? Should they? Yeah, they probably should.

#11 Raso

    Member

  • PipPipPipPipPipPipPipPip
  • The Sickle
  • The Sickle
  • 1,298 posts
  • LocationConnecticut

Posted 25 September 2015 - 09:45 AM

If I'm looking at the table correctly if you truly do play like you are god's gift to MWO and lose because you're stuck with filthy casual low tier trash you'll still go up in rank if you score high enough.

#12 Zergling

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Angel
  • The Angel
  • 2,439 posts

Posted 25 September 2015 - 09:52 AM

Being able to influence battles so your teams win more often than they lose requires skill, so there is nothing wrong with this system.

If you are unable to produce a positive winrate, then you will just have to accept that you aren't a good player by PGI's standards.

Edited by Zergling, 25 September 2015 - 09:54 AM.


#13 Bluttrunken

    Member

  • PipPipPipPipPipPipPip
  • The Patron Saint
  • The Patron Saint
  • 830 posts

Posted 25 September 2015 - 09:54 AM

View PostRaso, on 25 September 2015 - 09:45 AM, said:

If I'm looking at the table correctly if you truly do play like you are god's gift to MWO and lose because you're stuck with filthy casual low tier trash you'll still go up in rank if you score high enough.


Which doesn't make sense, doesn't it? If you play like god's gift to MWO you should rise up in tiers, and at a fast rate. As it is you only get the smallest increase for doing so. You can play terribly bad in the winning team and get the same increase. I got an increase in a winning team for ~240 damage, 1 kill, sub 200 match score.

#14 MrMadguy

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,288 posts

Posted 25 September 2015 - 09:55 AM

View Postkapusta11, on 25 September 2015 - 09:38 AM, said:

Isn't it player skill rating?

Edit: Oh great, now I look like a fool.

Edit 2: Wait, it's actually pilot skill rating.

Player, pilot, person... Don't this terms mean the same?

#15 Bluttrunken

    Member

  • PipPipPipPipPipPipPip
  • The Patron Saint
  • The Patron Saint
  • 830 posts

Posted 25 September 2015 - 09:58 AM

View PostZergling, on 25 September 2015 - 09:52 AM, said:

Being able to influence battles so your teams win more often than they lose requires skill, so there is nothing wrong with this system.

If you are unable to produce a positive winrate, then you will just have to accept that you aren't a good player by PGI's standards.

This was an really exceptional game for me but it shows the fallacy quite clearly. In this game at least the Victor and Mauler get the same increase as me, I think even the adder, while having a wildly different performance. I don't see the logic behind that. Im terribly sorry.

Posted Image

Edited by k05h3lk1n, 25 September 2015 - 10:03 AM.


#16 Zergling

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Angel
  • The Angel
  • 2,439 posts

Posted 25 September 2015 - 09:59 AM

FYI guys, the results of a single or small number of battles on PSR is irrelevant; what matters is the effects from a large number of battles over a long period of play.

Nobody that consistently plays like a god is going to keep on losing battles over the long term; their individual performance will influence enough battles to produce a positive winrate, and their rating will go up.

Nor will anybody that consistently plays like trash maintain a positive winrate; their rating won't be going up.

#17 Ghogiel

    Member

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

Posted 25 September 2015 - 10:05 AM

View Postk05h3lk1n, on 25 September 2015 - 09:41 AM, said:

Winning or losing is a random component.
We can just test it.


Just make an alt account and do sub 100 dmg every round. If w/l is random that account will have as much likelihood of having 50/50 as if you actually played hard and was doing +500 average dmg per match.

#18 Ghogiel

    Member

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

Posted 25 September 2015 - 10:09 AM

View Postk05h3lk1n, on 25 September 2015 - 09:58 AM, said:

This was an really exceptional game for me but it shows the fallacy quite clearly. In this game at least the Victor and Mauler get the same increase as me,


Show the player stats page! I would think you would have gone up in PSR, if you didn't I would be a bit surprised but I haven't had a 600+ match score loss since this patch to find out.

we know that only a 300 match score on a loss you do not go up PSR you get an = sign.

Edited by Ghogiel, 25 September 2015 - 10:10 AM.


#19 Bluttrunken

    Member

  • PipPipPipPipPipPipPip
  • The Patron Saint
  • The Patron Saint
  • 830 posts

Posted 25 September 2015 - 10:19 AM

View PostGhogiel, on 25 September 2015 - 10:05 AM, said:

We can just test it.


Just make an alt account and do sub 100 dmg every round. If w/l is random that account will have as much likelihood of having 50/50 as if you actually played hard and was doing +500 average dmg per match.


It's an interesting notion but I don't have time for it. I'd need to afk through at least 60+ matches to get a tendency and even then it's a bit unreliable. It should be more like 100+ matches to become somewhat reliable.

Edited by k05h3lk1n, 25 September 2015 - 10:20 AM.


#20 MrMadguy

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,288 posts

Posted 25 September 2015 - 10:51 PM

Also I want to say, that lack of personal skill based MM - is what leads to P2W. There is no real balance in this game, so there are Meta and FOTM 'Mechs here. And skill - is not always a personal experience. Builds used - are often may reach 50% of contribution into "skill". And Meta and FOTM - are what usually sold for $$$, even if it's just early access. It wouldn't be a problem in personal skill based game, cuz in this game Meta and FOTM players are matched against each other only. But when game developers say "you all will eventually end up in the same personal skill rating Tier" - it leads directly to P2W in a game, where there is no real balance between 'Mechs and builds. And I won't play P2W game, sorry.

Edited by MrMadguy, 25 September 2015 - 11:18 PM.






1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users