Jump to content

Psr Update And Changes - Jun 2020


490 replies to this topic

#41 D U N E

    Member

  • PipPipPipPipPip
  • Bad Company
  • Bad Company
  • 131 posts

Posted 04 June 2020 - 11:45 AM

Winning gives you more match score, either from the win, the kill assists from the 1 damage you did to an enemy etc. A player losing but doing 400ms can go up, though a player winning and getting 100ms needs to go down.

I'm all for a loss never giving an increase, though we also need to see decreases in match score for doing suboptimal. This being less than 250ms. It feels like I barely need to play to get 250ms,
Posted Image

The people in red should have went down, while the people in green carried their team.

#42 PrometheusTNO

    Member

  • PipPip
  • Star Captain
  • Star Captain
  • 43 posts
  • LocationUSA

Posted 04 June 2020 - 11:47 AM

View PostNearly Dead, on 04 June 2020 - 11:38 AM, said:

Teamwork = win

Who is guaranteed to be working as a team?



LOL you answered your own question wrong... Who will be working as a team? People that are concerned with winning.

#43 Joeseph Pierce

    Member

  • PipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 107 posts
  • LocationTerra

Posted 04 June 2020 - 11:49 AM

View PostDionnsai, on 04 June 2020 - 11:34 AM, said:


I think there is a slight EGO element in all of this. People are too worried about PSR gain/loss in individual matches. Across large sets of games, that just isn't going to matter.


Yeah, you're right on both counts. However, I think given how many people reacted negatively to the PSR reset in the first place (the "but I just reached tier 1" people), the PSR words and the little number next to the word "tier" does have a psychological effect on players, even if most of the top players know that in its present state it's meaningless. I know that I, as an average player, get some satisfaction seeing that I gained PSR on a loss, and I'm sure I would feel even worse about a very poor game if I lost PSR on a win.

As you mentioned, a fluke like an overall average player getting <100 or >400 MS isn't going to remotely affect the net PSR of the player in the long run, but it also provides at a glance an easy-to-decipher indication of how the player did. If the system will work as long as it forces good players up and bad players down in the long term, why not help those players figure out if they contributed or not by rewarding good play on a loss and punishing poor play on a win?

#44 PrometheusTNO

    Member

  • PipPip
  • Star Captain
  • Star Captain
  • 43 posts
  • LocationUSA

Posted 04 June 2020 - 11:54 AM

View PostD U N E, on 04 June 2020 - 11:45 AM, said:

The people in red should have went down, while the people in green carried their team.


Well... The winning team had an EXCELLENT score spread, indicating people were all contributing and working well together. Without knowing how the match actually went, it's possible that even the 102 score PIR played a key role in this (9 assists, probably quick back poking and harassing). It's not cut and dry here. The greens on the losing team would go up under both current and proposed systems. They would go up much more with the new system.

#45 Crashburn

    Member

  • PipPip
  • The Hitman
  • The Hitman
  • 39 posts

Posted 04 June 2020 - 11:54 AM

Going along w/ pretty much what's been posted so far. This new suggested PSR/rank format seems like it might skew ranks even more than previously. Like some have said if this is about getting more evenly based teams in matchmaker, I think this isn't going to solve it.

I agree with most of what Arjohan & Magic Pain Glove posted above. I was about to post a scale similar to his. Although if people think getting additional gains in PSR in a loss isn't the the best...perhaps making the scale more differentiated w/ additional levels could be a solution.

Loss
0 ~ 99 -5
100 ~ 199 -3
200 ~ 324 -1
325 ~ 399 0
400 ~ 450 +1
550 ~ infinite +2


I don't think you can discount high match score in a loss completely because...I 've been a series of games (been a long long while, but it has happened) where i've been on the rolled team 9 out of 10 matches. I'll use an extreme example to prove a point -> if you're a guy who gets 400 match score whilst the other 11 players get obliterated straight out...and you manage to stay alive to pull 400+ plus match score, I think that deserves a level up. I would say that players skill level is on a higher tier...but if that matchmaker always ends up throwing him on the inferior team match after match. That player will stay in the lower tiers...while a guy barely doing damage in a win starts leveling up to higher tiers. That's when the disparaties start to begin.

By the same token I would also change the scaling on winning. I don't really think it's that hard to get 250 match score in a win. Some matches....usually the stomps, the match finishes and I look at my dmg/match score and I'm surprised to see it as high, for not doing much. Usually it's because of a lopsided match where you're almost running full bore just to get shots on an enemy before the whole team evaporates. In cases such as that, I'm going to go up +3 for doing a whole lot of nothing IMHO. I'd expand that scale if possible and also give negative PSR for low match score in a win.

In the photo example of the match scores from another poster above. A Fafnir does 87 dmg in a gets a match score of 97. Had that been a win that would have bumped match score to 102 which would have given him a +1 PSR. I think everyone can agree on here if you're running a Fafnir and you do 82 dmg...I highly doubt you belong in a higher tier. There are always caveats (bad game, was afk, etc.)...and no system is perfect, but I'd say his Tier level is not on the same level as the guy pulling down 400/500+ match score in a loss.

Win

0 ~ 125 -1
125 ~ 199 0
200 ~ 324 +1
325 ~ 399 +3
400 ~ infinite +5

I think Arjohan scaling for wins is pretty close. Maybe adjust the -1 PSR scale slightly...and possibly add a 400 - 500 grouping etc. w/ smaller increases in a win. say +2/+3/+4/+5

That's my 2 cents.

Edited by Crashburn, 04 June 2020 - 12:02 PM.


#46 Mal Bolge

    Member

  • PipPipPipPipPip
  • 104 posts

Posted 04 June 2020 - 11:56 AM

View PostPaul Inouye, on 03 June 2020 - 12:09 PM, said:

New PSR values:
Player LOSES:
Match Score: 0-100 goes down in PSR by -5
Match Score: 101-250 goes down in PSR by -3
Match Score: 251-400 goes down in PSR by -1
Match Score: 401+ does not move.

Player WINS:
Match Score: 0-100 does not move.
Match Score: 101-250 goes up in PSR by +1
Match Score: 251-400 goes up in PSR by +3
Match Score: 401+ goes up in PSR by +5


This is not zero sum. This just helps bad players getting carried by their team, and punishes good players who carry a bad team. The fact that you use Player WINS instead of Team WINS almost implies that you don't think of this as a team game.

If you wanna go zero sum, try this:

Team LOSES:
Match Score: 0-100 goes down in PSR by -5
Match Score: 101-250 goes down in PSR by -2
Match Score: 251-400 does not move.
Match Score: 401-550 goes up in PSR by +2
Match Score: 551+ goes up in PSR by +5

Team WINS:
Match Score: 0-100 goes down in PSR by -5
Match Score: 101-250 goes down in PSR by -2
Match Score: 251-400 does not move.
Match Score: 401-550 goes up in PSR by +2
Match Score: 551+ goes up in PSR by +5

#47 PrometheusTNO

    Member

  • PipPip
  • Star Captain
  • Star Captain
  • 43 posts
  • LocationUSA

Posted 04 June 2020 - 11:56 AM

View PostJoeseph Pierce, on 04 June 2020 - 11:49 AM, said:

most of the top players know that in its present state it's meaningless


PSR considers me to be the equal of the top players in the world. I am offended at this on their behalf.

#48 OneTeamPlayer

    Member

  • PipPipPipPipPipPip
  • Ace Of Spades
  • 399 posts

Posted 04 June 2020 - 11:57 AM

I almost forgot to ask, if this change does not have the intended effect and a full 1 in 3 drops continues to be a stomp how do you have plans to adjust?

If not, how long do you think the playerbase will stick around for 1 out of 3 games as a stomp, especially the solo playerbase which makes a vast majority of the solo queue?

#49 Tahawus

    Member

  • PipPipPipPipPip
  • Little Helper
  • Little Helper
  • 189 posts

Posted 04 June 2020 - 11:59 AM

Wow, PGI, you're not inspiring confidence here.

Number 1. The proposed algorithm is not zero sum.

Number 2. My professors would have failed my *** if I said a move from 29.91% to 33.99% was a 5% increase. It's a 5percentage point increase. It is more correctly a 13.6% increase.

#50 PrometheusTNO

    Member

  • PipPip
  • Star Captain
  • Star Captain
  • 43 posts
  • LocationUSA

Posted 04 June 2020 - 12:02 PM

View PostMal Nilsum, on 04 June 2020 - 11:56 AM, said:

Match Score: 551+


551+ is rare enough that it should not be considered as a dedicated level. Continue to sub-divide the middle scores for a more impactful change.

#51 OneTeamPlayer

    Member

  • PipPipPipPipPipPip
  • Ace Of Spades
  • 399 posts

Posted 04 June 2020 - 12:05 PM

View PostD U N E, on 04 June 2020 - 11:45 AM, said:

Winning gives you more match score, either from the win, the kill assists from the 1 damage you did to an enemy etc. A player losing but doing 400ms can go up, though a player winning and getting 100ms needs to go down.

I'm all for a loss never giving an increase, though we also need to see decreases in match score for doing suboptimal. This being less than 250ms. It feels like I barely need to play to get 250ms,
Posted Image

The people in red should have went down, while the people in green carried their team.


Did you not notice that the difference between the two groups you highlighted largely comes down to light/mediums versus heavies/assaults?

Skipping the fact that Jarl's has the average match score on a bell curve peaking at 240ish (with the majority of players), by your metrics we should ensure that no one drops in any mech other than heavies or assaults as it is.

Lights have already been nerfed to hell and back, let's not take a bat to the kneecaps on the tail end, shall we?

#52 ThirdWorld

    Rookie

  • Survivor
  • 8 posts

Posted 04 June 2020 - 12:06 PM

I am excited as a narc raven to go down despite landing wins for my team.

#53 Dakkonn

    Member

  • PipPipPipPipPip
  • Big Brother
  • 120 posts
  • LocationMinnesota

Posted 04 June 2020 - 12:09 PM

I'd be interested in something like "top three" performers on the losing team still get PSR or at least don't lose it.

I'm a 300 average damage player. My score is often middle of the road or worse but I have certainly had games where I have 1000+ damage + 5 or more kills and score to reflect that but still be on the losing team.

Not to mention it shouldn't be a set number. PSR 5 match score requirements to advance should be lower than PSR 1 and score needed to advance from PS2 to PS1 should be the highest.

Also making it purely dependent on raw damage output heavily favors people just trying to deal max damage instead of teamwork.

Example, I am a light that managed to get around the enemy team unnoticed and capped their base despite the fact that we lost everyone else and I did no damage we won that match cause my team did their part and bought me the time needed to win. My match score is garbage but we won because of me. Why should I get punished afterward in the PSR rating?

Edited by Dakkonn, 04 June 2020 - 12:09 PM.


#54 MechWarrior414712

    Member

  • PipPipPipPipPipPip
  • 446 posts

Posted 04 June 2020 - 12:13 PM

I think only 300+ wins should add psr not anything below that, 300 match score is really easy to get for any decent player

#55 Taram

    Member

  • PipPipPipPipPip
  • Philanthropist
  • Philanthropist
  • 139 posts
  • LocationIn a Battlemech advancing on you.

Posted 04 June 2020 - 12:15 PM

I find this totally wrong. You should not penalize someone who played really well just because the team lost. The game overly penalizes good players when their team loses and over rewards bad players when their team wins.

Balance it out, make it zero sum but don't over reward someone just because their team happened to win despite them getting a ****** score. I also feel that doing really well, even in a loss, should give someone an increased match score.

Edited by Taram, 04 June 2020 - 12:29 PM.


#56 PrometheusTNO

    Member

  • PipPip
  • Star Captain
  • Star Captain
  • 43 posts
  • LocationUSA

Posted 04 June 2020 - 12:19 PM

View PostDakkonn, on 04 June 2020 - 12:09 PM, said:

My match score is garbage but we won because of me. Why should I get punished afterward in the PSR rating?


I don't think there's a nice way for me to put this...

PSR = Pilot Skill Rating

It didn't take skill to walk over to their base. You won that hypothetical match because other people on your team had the skills to fight well enough that you could stand in the square for a long time. THEY won that match for you. The enemy team just couldn't walk back fast enough when the fighting was over. And who knows... Maybe if you had been shooting some enemies in the back, you could have won AND scored points.

#57 D U N E

    Member

  • PipPipPipPipPip
  • Bad Company
  • Bad Company
  • 131 posts

Posted 04 June 2020 - 12:19 PM

View PostPrometheusTNO, on 04 June 2020 - 11:54 AM, said:


Well... The winning team had an EXCELLENT score spread, indicating people were all contributing and working well together. Without knowing how the match actually went, it's possible that even the 102 score PIR played a key role in this (9 assists, probably quick back poking and harassing). It's not cut and dry here. The greens on the losing team would go up under both current and proposed systems. They would go up much more with the new system.


But I was there, and 6 of the mechs died doing nothing, I remember the piranha running into their firing line, shooting all of them a little, and then died. Some of it's damage is from an arty strike I assume it placed, that or it seen an enemy light earlier and did some damage.

Annihilator sadly got ganked, TBF, not much he could do - he got ate as he tried with the rest of the team.

Pretty sure I remember the Adder, Kinatro, Hat, and BJ all try to fend off 8 mechs pushing from the choke point that the attackers usually nascar into the team.

We won cause the attackers split up after that, we had a few mechs spread around. Instead of breaking into groups, they kinda moved around in a confused pattern, and since we had mainly long range fire, we were able to sweep them up. Pretty sure they lost their cyclops in the push, possibly a stormcrow, only remember one running around for me to shoot.

With all that said, they did poor MS, next match they can do better and regain it. It's not like you never regain your MS once you lose it, if you have a bad match, you get penalized and next match you do better.

#58 OneTeamPlayer

    Member

  • PipPipPipPipPipPip
  • Ace Of Spades
  • 399 posts

Posted 04 June 2020 - 12:20 PM

People are still here suggesting that the top performers on the losing team go up in rank.

Here's the deal folks, PGI has said they're going to implement this system and these responses are essentially say "Yeah, but what if you tried using the old system, though?"

That's not how this is going to go down.

Here's a thought:

What if players:

1. Get off private discords and communicate with their team
2. Use the whole 9.99 it costs to get a functional mic OR:
3. Use the built in coordination tools that don't require a mike.
4. Stop using the mic to complain about a loss while having been silent all round, start using the mic to coordinate from the start
5. Accept that not all teams will want to coordinate but after a short time, people who won't work together will be pushed down while people who do will be pushed upwards.

Posted Image


This game has amazing potential and one of its limiting factors has always been that people talk about success in terms of individual performance instead of group success.

Well, PSR says that age is over so unless you're in one of the top tier groups that can literally drag a team to victory players are going to have to re-examine this "every mech is an island" concept that has been so ingrained.

Now it's "We all rise or we all Fall"- and we're going to have to drop accordingly.

Going to be a massive shift for most folk, but i can almost guarantee that the sea of "but what if you tried not making this based solely on team performance and allowed the losing team to individually win" are about to be real disappointed.

#59 Vidarion

    Member

  • PipPipPipPipPip
  • Philanthropist
  • Philanthropist
  • 102 posts

Posted 04 June 2020 - 12:20 PM

I'll just add my voice to this... the proposed PSR values are not zero-sum as they are independent of match score and purely rely on whether you were on the winning or losing team.

Mal's proposal is zero-sum and will allow good players to rise up regardless of the team's performance. And, while this might value player performance over team result, the fact that match score is already biased towards team performance means the PSR adjustment shouldn't place an additional penalty/bonus on the player.

View PostMal Nilsum, on 04 June 2020 - 11:56 AM, said:

If you wanna go zero sum, try this:
...
Match Score: 0-100 goes down in PSR by -5
Match Score: 101-250 goes down in PSR by -2
Match Score: 251-400 does not move.
Match Score: 401-550 goes up in PSR by +2
Match Score: 551+ goes up in PSR by +5


If the concern is that 551+ is "too rare", then lower it to 500+. The exceptionally good players will get that (just like they get that in events) anyway.

Another way to look at it is that your Tier should tend towards 6 - int(average MatchScore / 100). If you average 200-300 MS, you should be in Tier 3 or Tier 4. If you average below that, you should be in Tier 5. If you average 400+ you probably do belong in TIer 1 or 2.

#60 Anomalocaris

    Member

  • PipPipPipPipPipPipPip
  • Bad Company
  • 671 posts

Posted 04 June 2020 - 12:24 PM

To those stating that wins will ultimately even out - you are correct, in theory. The problem with this approach is it assumes that player distribution, particularly during the early ranking phase, will be random. It also assumes that players will be able to get enough games to balance out any matchmaking irregularities. This is generally assumed to be about 100 games, but truly we don't know.

In reality, prior to the last couple months, the average player in the game participated in fewer than 100 matches per month/season. Which means a substantial part of the player base will never achieve enough games to be ranked properly. Furthermore, matchmaking is not random because we have groups in play. This means that an external force is able to force a potentially unbalanced situation on the matchmaker.

This works in either direction. A really good solo player who never drops group is probably going to be undervalued by this new system. A mediocre player who plays a lot of group is going to be overvalued. Let's say they both end up at the same PSR rank and drop in a solo match. The mediocre player is considered a balance to the really good player and they are placed on opposite teams. The matchmaker has now failed to provide balance due to the group bias in PSR. Yes, if they play enough solo matches their rankings will move toward a more accurate position, but all that group player has to do is go group drop again to end up with an artificially boosted ranking.

This is why, in this game, at this time, you have to have the ability to move in the opposite direction of a win or loss if you significantly over or underperform. Until and unless there is a way to better assess player skill in groups vs. solos, looking at individual performance as an adjunct to WLR is all we have. And PGI have given no indications they have any idea on how to balance groups.





3 user(s) are reading this topic

0 members, 3 guests, 0 anonymous users