Jump to content

Easy Fix To Make Psr Fair


75 replies to this topic

#21 Kuritaclan

    Member

  • PipPipPipPipPipPipPipPip
  • Ace Of Spades
  • 1,838 posts
  • LocationGermany

Posted 02 February 2016 - 07:09 AM

View PostMrMadguy, on 02 February 2016 - 06:54 AM, said:

Do you realize, that your Tiers have different stabilization avg Match Score levels? Do you understand, that stabilization level - is avg Match Score, you are allowed to have without further advancement? Different stabilization levels - players will be allowed to have different avg Match Scores in different Tiers. Tier, you will settle in - depends on your skill. Is it fair, that players with different skill levels will be allowed to have different avg Match Scores? I don't think so.

The Matchscore only depends on your skill - not on your Tier you are settled in. Someone who is above average of his tier he currently is in will have a better matchscore than the surrounding players no matter what. He only will gain an equal matchsocre on average, when he actually is surrounded by evenly skilled players. This correction towards a equal matchscore is only achivable with a deminishing return on matchscore, not an open end matchscore, where everybody will reach the same pool over time.

View PostLugh, on 02 February 2016 - 07:04 AM, said:

Loose applies to ties, shoelaces, rope knots and people of questionable morals.

Sry - didn't saw this mistake out of itself (no native english speaker). Sure it should be lose.

Edited by Kuritaclan, 02 February 2016 - 07:11 AM.


#22 MrMadguy

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,221 posts

Posted 02 February 2016 - 07:20 AM

View PostMystere, on 02 February 2016 - 07:06 AM, said:

You want to reduce winning as a major part of PSR? Do tell then how we should rate people who vanguard, TAG, spot, NARC, bait, and distract in lieu of killing to help their team win? Should make their PSR drop to the floor?

How do you accurately measure someone playing vanguard? What accurate method do you use to detect someone baiting, distracting, or delaying the enemy?

Isn't it much simpler to just use winning as an approximation?

While winning has nothing to do with having fun, performance and rewards - it can't be used as approximation of someones' usefulness for his team. He was baiting, distracting, delaying enemies. We won. Ok. My Match Score is still <100. So what's the difference? +10K CB for win? Pfffff.

Edited by MrMadguy, 02 February 2016 - 07:27 AM.


#23 Kuritaclan

    Member

  • PipPipPipPipPipPipPipPip
  • Ace Of Spades
  • 1,838 posts
  • LocationGermany

Posted 02 February 2016 - 07:25 AM

View PostMystere, on 02 February 2016 - 07:06 AM, said:

You want to reduce winning as a major part of PSR? Do tell then how we should rate people who vanguard, TAG, spot, NARC, bait, and distract in lieu of killing to help their team win? Should make their PSR drop to the floor?

How do you accurately measure someone playing vanguard? What accurate method do you use to detect someone baiting, distracting, or delaying the enemy?

Isn't it much simpler to just use winning as an approximation?

Winning is over all the best count. However it doesn't say something about the quality how you did interfere in the match. And at least within the match score are such figures as flanking, lance information, scouting etc.. I rather get into my pool of same skillers by a faster progress because my skill gives me a higher score, than w8 long so that the luck gives me over time this pleasure because of 11 others who influence the outcome a way more than me.

Edited by Kuritaclan, 02 February 2016 - 07:26 AM.


#24 Mystere

    Member

  • PipPipPipPipPipPipPipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 22,783 posts
  • LocationClassified

Posted 02 February 2016 - 07:27 AM

View PostMrMadguy, on 02 February 2016 - 07:20 AM, said:

While winning has nothing to do with having fun, performance and rewards - it can't be used as approximation of someones' usefulness for his team. He was baiting, distracting, delaying enemies. We won. Ok. My Match Score is still <100. So what's the difference? +10K CB for win? Pfffff.


Once again, how do you accurately measure someone playing vanguard? What accurate method do you use to detect someone baiting, distracting, or delaying the enemy?

And winning has nothing to do with having fun and rewards? Really? What world do you live in?

Edited by Mystere, 02 February 2016 - 07:27 AM.


#25 Hotthedd

    Member

  • PipPipPipPipPipPipPipPipPip
  • The People's Hero
  • 3,213 posts
  • LocationDixie

Posted 02 February 2016 - 07:28 AM

View PostMrMadguy, on 02 February 2016 - 07:20 AM, said:

While winning has nothing to do with having fun, performance and rewards - it can't be used as approximation of someones' usefulness for his team.

Wait, WHAT?

Over the course of many matches, someone who wins more than the lose is doing SOMETHING to be an asset to their team. Conversely, there are builds that can EASILY do a lot of damage, but not necessarily help the team to victory. Just as there needs to be more than win/loss as a metric, there needs to be more than the current iteration of match score.

#26 pyrocomp

    Member

  • PipPipPipPipPipPipPipPip
  • Philanthropist
  • 1,036 posts

Posted 02 February 2016 - 07:31 AM

First. Any talk about average score with surrounding players since this game is about teams (not your unaffected single handed performance against AI with a strict set of rules and no random element) may refer only to the players within the same match and strictly within same team. I've been on a 4-12 loss where no one on our team dealt less than 270 dmg and we severely outdamaged oppenent team. There were no 1k dmg. Just plain fact. We severely outdamaged them and lost. So who are the best players? If only damage was meaningful then we did better. But we lost the match. So W/L should be factored no matter what.
Second. All talks about only high score matters always forget about squirrels and puhsers that are sometimes vital to the team result. Sometimes they are spoiled by a cowardy or just not very aware team. Just what mechs and how do you play? Are you in a last line brawler assault? Or are you a sniper ecm light?

PS: OP, it's the Nth 'I hate this game' thread for the last few weeks (two or three about polar, a pair on LRMs, who-knows-how-many on 'broken' MM and now PSR). Why do you even bother to play this game?

Edited: typos

Edited by pyrocomp, 02 February 2016 - 07:40 AM.


#27 Mystere

    Member

  • PipPipPipPipPipPipPipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 22,783 posts
  • LocationClassified

Posted 02 February 2016 - 07:32 AM

View PostKuritaclan, on 02 February 2016 - 07:25 AM, said:

Winning is over all the best count. However it doesn't say something about the quality how you did interfere in the match. And at least within the match score are such figures as flanking, lance information, scouting etc.. I rather get into my pool of same skillers by a faster progress because my skill gives me a higher score, than w8 long so that the luck gives me over time this pleasure because of 11 others who influence the outcome a way more than me.


Let us assume we have a player who is very good at hanging on the coattails of the people he drops with. I'd say he's a lucky mascot and as such have no problem having him on the team. His mere presence is good if he wins a lot. Posted Image

#28 pyrocomp

    Member

  • PipPipPipPipPipPipPipPip
  • Philanthropist
  • 1,036 posts

Posted 02 February 2016 - 07:33 AM

View PostMystere, on 02 February 2016 - 07:27 AM, said:


Once again, how do you accurately measure someone playing vanguard? What accurate method do you use to detect someone baiting, distracting, or delaying the enemy?

And winning has nothing to do with having fun and rewards? Really? What world do you live in?

Is it groundhog day already? We've been through exactly this set of arguments a week ago. Against same guy with same set of arguments.

#29 Mystere

    Member

  • PipPipPipPipPipPipPipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 22,783 posts
  • LocationClassified

Posted 02 February 2016 - 07:36 AM

View Postpyrocomp, on 02 February 2016 - 07:33 AM, said:

Is it groundhog day already? We've been through exactly this set of arguments a week ago. Against same guy with same set of arguments.


Actually, we had the same thing just 3 days ago.

#30 pyrocomp

    Member

  • PipPipPipPipPipPipPipPip
  • Philanthropist
  • 1,036 posts

Posted 02 February 2016 - 07:37 AM

View PostMystere, on 02 February 2016 - 07:36 AM, said:


Actually, we had the same thing just 3 days ago.

Felt like a week. Need a break.

#31 Kuritaclan

    Member

  • PipPipPipPipPipPipPipPip
  • Ace Of Spades
  • 1,838 posts
  • LocationGermany

Posted 02 February 2016 - 07:37 AM

View PostMystere, on 02 February 2016 - 07:32 AM, said:


Let us assume we have a player who is very good at hanging on the coattails of the people he drops with. I'd say he's a lucky mascot and as such have no problem having him on the team. His mere presence is good if he wins a lot. Posted Image

Posted Image Ok. but this is groupe queue anyway. So you wanna have him with you no matter what. In the lets **** 12 players on a team anyway queue you may not wanna have such mascot players surrounding you for everrrrr.......... . Posted Image

#32 MrMadguy

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,221 posts

Posted 02 February 2016 - 07:39 AM

View PostKuritaclan, on 02 February 2016 - 07:09 AM, said:

The Matchscore only depends on your skill - not on your Tier you are settled in. Someone who is above average of his tier he currently is in will have a better matchscore than the surrounding players no matter what. He only will gain an equal matchsocre on average, when he actually is surrounded by evenly skilled players. This correction towards a equal matchscore is only achivable with a deminishing return on matchscore, not an open end matchscore, where everybody will reach the same pool over time.

According to your table stabilization levels for Tiers are the following ones:
Tier 1 => ~125MS
Tier 2 => ~150MS
Tier 3 => ~175MS
Tier 4 => ~200MS
Tier 5 => ~225MS
Tier 6 => ~250MS
Tier 7 => ~275MS
Tier 8 => ~300MS
Tier 9 => ~325MS

I settled in Tier 1. You settled in Tier 9. My average Match Score will be 125, your average Match Score will be 325. Is it fair? No.

There is target Match Score, every player should have. Now it's 175. If your average Match Score in higher, then 175 - you will advance and play with/against more skilled players, so your MS will drop, till it won't become 175 or you won't reach 100% Tier 1. If your average Match Score is lower - you will drop and play with/against less skilled players, so your MS will rise, till, again, it won't become 175 or you won't reach 0% TIer 5. When all players will settle - they all, except some minority of top and bottom players, will have exactly the same average Match Score. All are in equal conditions - it's fair.

With your system players will settle at different average Match Scores. And this is unfair.

Edited by MrMadguy, 02 February 2016 - 07:42 AM.


#33 Kuritaclan

    Member

  • PipPipPipPipPipPipPipPip
  • Ace Of Spades
  • 1,838 posts
  • LocationGermany

Posted 02 February 2016 - 07:50 AM

View PostMrMadguy, on 02 February 2016 - 07:39 AM, said:

I settled in Tier 1. You settled in Tier 9. My average Match Score will be 125, your average Match Score will be 325. Is it fair? No.

It is totally fair with 125 matchscore you are just bad so you play with others who on average gain a matchscore of around 125. If you wann play with/against someone of T6 you have to get better or a luck that waiting times are so high that the t6 has to w8 6 minutes and get therefore some of t1 mixed in.

View PostMrMadguy, on 02 February 2016 - 07:39 AM, said:

With your system players will settle at different average Match Scores. And this is unfair.

It is totally fair to do it that way. Everybody is liable to his match score. And everybody can with more practice/training/knowledge about the game have a raise in its average match score. Skill is trainable. Muscle memory for example.

Who starts to play the game may not have the knowledge to reach a score of your "175" - However everybody who isn't drugged and or otherwise disabled will have a avg. Match Score far above 175. So in the end EVERYBODY will slowly rise into the maximum tier, what makes this tier obsolete and we could also match everybody against everybody beside some players who are just plain bad.

The purpose of a rating system is to match evenly skilled players against each other - not everybody who has stepped over a certain hurdle.

Edited by Kuritaclan, 02 February 2016 - 07:57 AM.


#34 pyrocomp

    Member

  • PipPipPipPipPipPipPipPip
  • Philanthropist
  • 1,036 posts

Posted 02 February 2016 - 07:59 AM

View PostKuritaclan, on 02 February 2016 - 07:50 AM, said:

It is totally fair to do it that way. Everybody is liable to his match score. And everybody can with with more practice/training/knowledge about the game have a raise in its average match score. Skill is trainable. Muscle memory for example.

Match score depend on damage dealt (and win/loss is at present a twice acting factor as there is also a bonus to the match score for the win). Damage dealt depend on average weight of the oponnent team (which is a factor of what was the weight of mechs just released and just released for CBs, what is on sale and of what current playerbase mood is) and average oponnent pilot's damage spread skill. So the match score depend not only on your skill but on the opponnet piloting skill.

And, if you missed previous incrnations of this thread, PSR shoud be renamed (as it in fact is) to PV as Pilots Value. As it reflects piloting skills only partially and requires commanding and awareness skills also (plus some other that we will ommit for clarity). And skill bars (if added) should be a less XP bars and more rely on match averages (like average number of KMDD, brawling, scouting and so on).

#35 Kuritaclan

    Member

  • PipPipPipPipPipPipPipPip
  • Ace Of Spades
  • 1,838 posts
  • LocationGermany

Posted 02 February 2016 - 08:12 AM

View Postpyrocomp, on 02 February 2016 - 07:59 AM, said:

Match score depend on damage dealt (and win/loss is at present a twice acting factor as there is also a bonus to the match score for the win). Damage dealt depend on average weight of the oponnent team (which is a factor of what was the weight of mechs just released and just released for CBs, what is on sale and of what current playerbase mood is) and average oponnent pilot's damage spread skill. So the match score depend not only on your skill but on the opponnet piloting skill.

Clearly the match score is a result of this variables - however there are also in this your and only those wich are on your liability variables like: KMDD, Assists+Kills, UAV Detection/Kills, Counter ECM, Scouting, Savior Kills, Scouting, Flanking, Hit and Run, Brawling etc.. Those are all counted into the matchscore.

When the new mech X is released it needs on average more damage to kill all enemies - however this also applies to all who are on the same tier who play at that time. So it is totally equal. Over time it balance itself. So in the end even if you make in this period of time the mech is favored a higher match score you only get as high as you are supposed to get, since if you rise in match score enemies will get better and you are more likely to get killed faster, if you not belong in that tier.

Last but not least. Sure match score depends heavy on the opponent piloting skill - which isn't that much away from yours if you are matched into the same tier as long as there are enough players you can matched with. If you play on a european server at 4 a.m. you will probably have better players in the deep night. And if you normally play on the evening where everybody is online you may drop in your tier if you go up against those night players to often. However if you are as good as them you stay on your tier in the deep night as on the evening. If you can not handle the try hards - don't play at this time.

Edited by Kuritaclan, 02 February 2016 - 08:20 AM.


#36 MrMadguy

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,221 posts

Posted 02 February 2016 - 08:25 AM

View PostKuritaclan, on 02 February 2016 - 07:50 AM, said:

It is totally fair with 125 matchscore you are just bad so you play with others who on average gain a matchscore of around 125. If you wann play with/against someone of T6 you have to get better or a luck that waiting times are so high that the t6 has to w8 6 minutes and get therefore some of t1 mixed in.

It is totally fair to do it that way. Everybody is liable to his match score. And everybody can with more practice/training/knowledge about the game have a raise in its average match score. Skill is trainable. Muscle memory for example.

Match Score isn't absolute - it's relative to other players, you are in one match with. Good player in match with 23 bad players will have higher match score and bad player in one match with 23 good players will have lower match score, then the same player in match with same skill players. Balance - is not when all players at the same Tier have the same Match Scores. It's when all players in entire game have the same Match Scores. Equal conditions for all players - that is the real purpose of having matchmaker.

It's not just my ephemeral idea - I experienced it by myself many times. Buy AWS-8R, equip 4xLRM15 + 2K missiles. Play some games, gain 50% of Tier for example. Then switch to your ordinal 'Mech and enjoy your enless stomp streak with <100 MS matches. Your "skill" won't help you earn higher Match Scores. And it will be happening, till you won't drop back to your original PSR rating. If you don't know this - then you don't know the real purpose of having matchmaker in any game. Matchmaker is here to prevent situation, where pro skilled player stomps complete noobs, deals 100500dmg and earns 12 kills, while his team is doing nothing and earns 0 dmg and kills.

View PostKuritaclan, on 02 February 2016 - 07:50 AM, said:

Who starts to play the game may not have the knowledge to reach a score of your "175" - However everybody who isn't drugged and or otherwise disabled will have a avg. Match Score far above 175. So in the end EVERYBODY will slowly rise into the maximum tier, what makes this tier obsolete and we could also match everybody against everybody beside some players who are just plain bad.

BINGO!!! That's why current PSR system is so terrible - it's way too biased towards winning and has way too low stabilization level. That's why it needs to be changed.

Edited by MrMadguy, 02 February 2016 - 08:39 AM.


#37 Scar Glamour

    Member

  • PipPipPipPipPipPip
  • The Undertaker
  • The Undertaker
  • 267 posts

Posted 02 February 2016 - 08:26 AM

You don't need any global changes to fix PSR. In fact you can fix it right now and all by yourself using a secret method.

You should go into game Settings and turn off showing PSR on front end. Problem solved, have fun.l!

#38 pyrocomp

    Member

  • PipPipPipPipPipPipPipPip
  • Philanthropist
  • 1,036 posts

Posted 02 February 2016 - 08:31 AM

View PostKuritaclan, on 02 February 2016 - 08:12 AM, said:

Clearly the match score is a result of this variables - however there are also in this your and only those wich are on your liability variables like: KMDD, Assists+Kills, UAV Detection/Kills, Counter ECM, Scouting, Savior Kills, Scouting, Flanking, Hit and Run, Brawling etc.. Those are all counted into the matchscore.

When the new mech X is released it needs on average more damage to kill all enemies - however this also applies to all who are on the same tier who play at that time. So it is totally equal. Over time it balance itself. So in the end even if you make in this period of time the mech is favored a higher match score you only get as high as you are supposed to get, since if you rise in match score enemies will get better and you are more likely to get killed faster, if you not belong in that tier.

Last but not least. Sure match score depends heavy on the opponent piloting skill - which isn't that much away from yours if you are matched into the same tier as long as there are enough players you can matched with. If you play on a european server at 4 a.m. you will probably have better players in the deep night. And if you normally play on the evening where everybody is online you may drop in your tier if you go up against those night players to often. However if you are as good as them you stay on your tier in the deep night as on the evening. If you can not handle the try hards - don't play at this time.

When the new mech is released you get extreme drop deck. Like 6 MAL/MAD/WHM on a single team. You get heavier (or a lighter drop with luck) on average for quite a time. But disparity between matches grow and you get more stomps as people level and get used to their new mechs. And those mechs usualy do low damage. Ask people about their experience leveling Black Knight.

Edited by pyrocomp, 02 February 2016 - 08:40 AM.


#39 Mawai

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 3,495 posts

Posted 02 February 2016 - 08:35 AM

View PostMrMadguy, on 02 February 2016 - 06:54 AM, said:

Do you realize, that your Tiers have different stabilization avg Match Score levels? Do you understand, that stabilization level - is avg Match Score, you are allowed to have without further advancement? Different stabilization levels - players will be allowed to have different avg Match Scores in different Tiers. Tier, you will settle in - depends on your skill. Is it fair, that players with different skill levels will be allowed to have different avg Match Scores? I don't think so.


Do you realize that your scheme is completely broken?

If you play in balanced matches which is the GOAL of this process correct? i.e. The matchmaker selects players of the same skill level in order to create a match in which either team has exactly a 50% chance of winning. That is what balanced means.

However, if a player does not average 325 match score between both a win and a loss they will drop. Depending on how you implement the change it would be 250 to 400 ... though if the amount of change is linear then you need to hit the 325 average to stay even.

In losses, the usual outcome is 12:2,3,4 mucjh of the time because the game snowballs. If someone is out of position or the other team pushes at the right time then you will lose. In losses, there is often FAR less opportunity to do damage and build up match score (unless the two teams agree to snipe at each other for a while to pad their damage stats). This means that match scores on a loss tend to be much lower.

Now consider the following, on a team, when folks work together, ALL of the match scores are lower since the damage is usually spread among all the players. I've seen great games where the entire team had a match score under 300 since the damage was well focused and distributed. In your example, after this great performance, the players score would at best stay even and the folks under 250 would drop.

Finally, the goal of the game is WINNING. It isn't doing damage. It isn't cheering for your team. IT IS WINNING. If you don't win then you FAILED no matter what damage you did in the match. I've seen team mates rack up damage and kills as the last one alive ... they got a high match score but were totally incompetent in terms of actually helping their team when it was needed and winning the game. They had NO SKILL and yet the suggestion you have made would consider them the best player and reward them with a higher tier. (To be honest, I'd be glad to foist folks like that off on another group of players).

Similarly, I have seen pilots drop a critical UAV that high lights the entire opposing force and results in their team defeating a planned ambush. The team won due to that UAV but unfortunately, the pilot who dropped it didn't get away. In your model, he would be greatly penalized as an under performing pilot ... as it was he still didn't get any increase in PSR but at least it didn't drop. Similarly, the mechs that lead a push are often at high risk but that push at the right time will win the match.

There are dozens more situations in which the actions of a player will change the odds of WINNING at the expense of their personal match score. These are the reason why WINNING is a factor in PSR calculations ... it is impossible to try to add up all the situational elements that contribute to a win (even saying the right thing in chat or on coms can often make a difference) ... having someone organize is even better. Team work is OP and there is NO way to award a match score to the guy who died during the push but organized the plan that won the game.

For all these reasons, your perfectly even win/loss stabilization chart is totally meaningless and has no real way to separate players by "skill".

P.S. Add to the fact that certain weapons like LRMs are very good at padding damage and match score and you have a situation where the LRM boaters or the ones driving OP mechs will be considered the "skilled" players ... LOL.

#40 KHETTI

    Member

  • PipPipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 1,328 posts
  • LocationIn transit to 1 of 4 possible planets

Posted 02 February 2016 - 08:35 AM

PGI need to ditch their match maker and the absolute joke PSR has turned out to be.
There's more low tier players at tier 2/1 than players that actually deserve to be in that tier.
Some of us are continually forced to play with the Downs Syndrome Elite squads, which guarantees you have less than 5% chance of winning!, and the best thing is i have to wait up to 10 minutes for BS matches.

MW4 style lobbies please, atleast this way i can filter out the rejects of the human race being in my games, Yes you ruin the game for me!.





2 user(s) are reading this topic

0 members, 2 guests, 0 anonymous users