Jump to content

Psr: Less Emphasis On Winning Matches


76 replies to this topic

#41 Sjorpha

    Member

  • PipPipPipPipPipPipPipPipPip
  • Philanthropist
  • Philanthropist
  • 4,477 posts
  • LocationSweden

Posted 02 March 2016 - 11:09 PM

View PostSandpit, on 02 March 2016 - 10:46 PM, said:

...
I almost resisted...

really?
you know darn well what everyone is talking about in regards to W/L,or at least you should because of the context of their posts, the thread, and a little common sense...?Posted Image


Since there are so many posts where the use of "W/L" only leads to the posters conclusion if it refers to the ratio, and others where it could refer to single match results (bu the conclusions are still usually a little strange), I think there are examples of both interpretations and that this leads to errors of category in the discussion.

#42 MrMadguy

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,221 posts

Posted 02 March 2016 - 11:28 PM

It's absolutely right attitude, because win doesn't mean anything, when your performance is poor. Yeah, W/L correlates with your skill, but it's NOT EXACT REPRESENTATION of it - it's may be true only for 1vs1 and static player base situation. When you can solely win match - you are at too low rating, yeah, but at higher rating you start to be carried - your contribution towards victory starts to be lower, lower and lower - right to a point, where your team can win match even without you. W/L no longer represents your skill. At this point your rating should be tweaked according to your performance, i.e. match score.
Proof, that W/L doesn't mean anything:
Posted Image
Posted Image
Posted Image
Posted Image

Edited by MrMadguy, 02 March 2016 - 11:29 PM.


#43 Prof RJ Gumby

    Member

  • PipPipPipPipPipPipPipPip
  • The People's Hero
  • 1,061 posts

Posted 03 March 2016 - 12:17 AM

Russ's plan is just a reaction to our expectations. It's because many people think"W/L does actually change anything in the long run.

W/L actually barely influences anything. It's most often almost 1 to 1 if not counted on a very short time sample. If it isn't - it means you statistically don't "have your moments" often enough. If it is - you can carry or get a good score even on a lost match.

What actually influences PSR in general, not in small occassional fluctuations, is whether you can get a good score on a win, and - more importantly - whether you can get a good score on a loose. You get good scores on a loose - no PSR drops to cancel out those inevitable rises on wins.

This is in fact a good indicator. If you can still get decent score even while your teams is collapsing around you like a deck of cards and you're basically on you own vs a bunch of enemies - that means you are actually too good for your tier, because you're basically better than anyone in that match, including members of the winning team.

Whether scouting/harassing and other important actions that not give you lots of dmg are represented enough - hard for me to say. Maybe this part could use an improvement.

EDIT:
Food for thought: if certain actions don't give you much dmg/kills/points, but really help your team win, doesn't "focusing more" on points and "less on wins" actually screw people doing this? PSR can't measure skill of a good caller, charismatic commander or long range scout who keeps the team informed even when the enemy is under ECM/too far away. On Polar Highlands a good scout can decide on whose terms the teams will clash and who will be better positioned and ready. How to measure that if not by W/L ratio alone?

Edited by Prof RJ Gumby, 03 March 2016 - 12:25 AM.


#44 Intrepid

    Member

  • PipPipPipPipPipPip
  • The Blade
  • The Blade
  • 265 posts

Posted 03 March 2016 - 01:50 AM

View PostStefka Kerensky, on 02 March 2016 - 10:21 AM, said:

ELO was taking in consideration w/l only. And it was a really bad system


Wins are the only thing that matter. ELO was a much better system and is why it is used.

#45 nehebkau

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,386 posts
  • LocationIn a water-rights dispute with a Beaver

Posted 03 March 2016 - 06:15 AM

View PostMystere, on 02 March 2016 - 08:05 PM, said:

I'd say a winner is much better than someone who is constantly losing whenever he is the biggest scorer.


True, however, as we have seen with PSR, the small population and the large tolerances in different player skills matched together, wins are, more often than naught at the mercy of who got the biggest (most) scrub(s). In a larger population with tighter restrictions of what skill players get put together I would agree -- in MWO I don't.

#46 Lugh

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Widow Maker
  • The Widow Maker
  • 3,910 posts

Posted 03 March 2016 - 06:34 AM

View PostIntrepid, on 03 March 2016 - 01:50 AM, said:


Wins are the only thing that matter. ELO was a much better system and is why it is used.

In single player match ups. Sure.

In a team game where I can only do so much about 11 other yahoos on my team. Not so much. Accounting for my personal score is where it's at.

They need to increase score rewards for protected 'asslt,hvy,med,lt'. They need to increase score rewards for scouting. Institute rewards to score to AMS (yes it is valuable to your team to have that added protection from the rain and help the fight in the shade).

ECM protection should also provide a score reward as well.

They COULD add a score reward for leadership items like Pinging the map at C1, and then any mechs that die in C1 on the enemy team give a little 'leadership bonus' to the guy that pinged the map...(within say 1-2minutes of that ping)

Edited by Lugh, 03 March 2016 - 06:34 AM.


#47 Screech

    Member

  • PipPipPipPipPipPipPipPipPip
  • Knight Errant
  • 2,290 posts

Posted 03 March 2016 - 06:53 AM

Worst thing they could do in the solo queue is lessen the importance of winning in matches. If you can't consistently win in matches currently then you are in the right tier or perhaps need to be placed in a higher tier. If you believe the reason you can't consistently win is your teammates then this is doubly so.

Last thing the solo queue needs is to have it behave like event play 24/7 with folks trying to max out match scores.

#48 RAM

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Resolute
  • The Resolute
  • 2,019 posts
  • Google+: Link
  • LocationVancouver, BC

Posted 03 March 2016 - 07:15 AM

View PostLugh, on 03 March 2016 - 06:34 AM, said:

In a team game where I can only do so much about 11 other yahoos on my team. Not so much. Accounting for my personal score is where it's at.

No. The math is absolutely solid on this. Individual/solo, formed team or random group makes no difference when only wins matter.

ELO not only works, it works VERY well. Whether you comprehend the math behind ELO or not is irrelevant.


RAM
ELH

#49 Torgun

    Member

  • PipPipPipPipPipPipPipPip
  • 1,598 posts

Posted 03 March 2016 - 07:24 AM

View PostRAM, on 03 March 2016 - 07:15 AM, said:

No. The math is absolutely solid on this. Individual/solo, formed team or random group makes no difference when only wins matter.

ELO not only works, it works VERY well. Whether you comprehend the math behind ELO or not is irrelevant.


RAM
ELH


I think you missed the part that Elo works for teams when you play in the same team that makes up the entire team every time. Slapping together 12 individuals on each side and then rank them as a teams with Elo is epic fail. PSR does have serious issues too, but the basis for the rankings is way better then Elo.

#50 Mystere

    Member

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

Posted 03 March 2016 - 07:39 AM

View PostBoogie138, on 02 March 2016 - 07:36 PM, said:

Kinda self-serving to me.

The more team work = better personal performance (usually)

So, whatevs..

View PostMrMadguy, on 02 March 2016 - 11:28 PM, said:

It's absolutely right attitude, because win doesn't mean anything, when your performance is poor. Yeah, W/L correlates with your skill, but it's NOT EXACT REPRESENTATION of it - it's may be true only for 1vs1 and static player base situation. When you can solely win match - you are at too low rating, yeah, but at higher rating you start to be carried - your contribution towards victory starts to be lower, lower and lower - right to a point, where your team can win match even without you. W/L no longer represents your skill. At this point your rating should be tweaked according to your performance, i.e. match score.
Proof, that W/L doesn't mean anything:


Other than winning, how does MWO reward someone who specializes in and is very successful at commanding rather than direct combat?

Edited by Mystere, 03 March 2016 - 07:41 AM.


#51 Sandpit

    Member

  • PipPipPipPipPipPipPipPipPipPipPipPip
  • Veteran Founder
  • Veteran Founder
  • 17,419 posts
  • Facebook: Link
  • Twitter: Link
  • LocationArkansas

Posted 03 March 2016 - 08:44 AM

View PostSjorpha, on 02 March 2016 - 11:09 PM, said:


Since there are so many posts where the use of "W/L" only leads to the posters conclusion if it refers to the ratio, and others where it could refer to single match results (bu the conclusions are still usually a little strange), I think there are examples of both interpretations and that this leads to errors of category in the discussion.

You directly quoted me, you knew what I was talking about though
anyhow, moving on...

Again, I don't see anywhere that anyone is asking for that scenario

View PostMystere, on 03 March 2016 - 07:39 AM, said:


Other than winning, how does MWO reward someone who specializes in and is very successful at commanding rather than direct combat?

That's the other side of it.

It's very easy to sit back and say "reward me for this", but how do you quantify that? How do you put an actual integer and value on some of the things being discussed?

Players should be rewarded for leading their team to a victory.
Ok, how? How does the game decide who the "leader" was? How does PGI know who the drop lead was?

There's just no way to quantify that in this type of game that I can think of.

#52 Mister Blastman

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Survivor
  • Survivor
  • 8,444 posts
  • LocationIn my Mech (Atlanta, GA)

Posted 03 March 2016 - 08:47 AM

We don't need rewards, we need segregation. We need a simple system...

IF Player does less than [poor player threshold]

THEN Player is segregated to PUGisland™ for [timelimit]

I don't code anymore, but that should be easy enough to understand, where:

[poor player threshold] = damage value decided by PGI. I recommend we set it to between 100 and 150 pts of damage. This value is also dependent on player scoring this way three to four games in a row.

[timelimit] = 1 day first stay, 7 days second, 1 month third.

You may earn your way off PUGisland™ by achieving 550 damage four drops in a row.

Otherwise, you serve your time with other PUGislanders™


WHY BLASTMAN?

Because people are fed up with bads on their team AND bads are fed up with getting rolled.

EVERYONE WINS!

Bads play with bads, everyone else plays with MWOcompetents™

Problem solved.

This is not a complicated system. Penalties are needed for PSR, not benefits.

#53 Mystere

    Member

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

Posted 03 March 2016 - 08:53 AM

View PostSandpit, on 03 March 2016 - 08:44 AM, said:

That's the other side of it.

It's very easy to sit back and say "reward me for this", but how do you quantify that? How do you put an actual integer and value on some of the things being discussed?

Players should be rewarded for leading their team to a victory.
Ok, how? How does the game decide who the "leader" was? How does PGI know who the drop lead was?

There's just no way to quantify that in this type of game that I can think of.


Precisely. This is why I think a player's capacity to win should be the major metric because in the long term the player is the only constant.

#54 Sandpit

    Member

  • PipPipPipPipPipPipPipPipPipPipPipPip
  • Veteran Founder
  • Veteran Founder
  • 17,419 posts
  • Facebook: Link
  • Twitter: Link
  • LocationArkansas

Posted 03 March 2016 - 09:02 AM

View PostMystere, on 03 March 2016 - 08:53 AM, said:


Precisely. This is why I think a player's capacity to win should be the major metric because in the long term the player is the only constant.

I don't think it should the the "major" or "main" metric used, but it should be in the formula. I think match score is still the best indicator of how the individual player did on the team.
You could set it up on a sliding scale based on a curve for the entire team.

Got highest match score on your team?
You obviously did well based on your entire team's performance.

Most times, for me anyhow, when my team wins and it was a decent match (meaning at least 3-6 kills on both sides) and not a complete roflstomp, most of the players (if not all) on the winning team are going to have 200+ damage with a few exceptions here and there.

Ont he losing side, not so much a lot of times. If you're on a team where 5 players didn't break 200 damage with a few even under 100 damage, but you turned in 300 damage and got a kill and a couple of assists, you did everything you possibly could to help that team win. You deserve to have that taken into consideration when it comes to calculating your individual PSR.

That's the other thing, yes, this is a team based game.

BUT

PSR is an individual score. Your skill ranking according to the game and the opponents you face regardless of what your team looks like is based on your individual PSR.

#55 L A V A

    Member

  • PipPipPipPipPipPip
  • The Grizzly
  • The Grizzly
  • 308 posts
  • LocationOn the beach!

Posted 03 March 2016 - 09:20 AM

View PostMister Blastman, on 03 March 2016 - 08:47 AM, said:

damage
damage
damage


Exactly.

Like I said before, it all comes down to damage, boating massive alphas and using your team mates as meat shields.

#56 Mister Blastman

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Survivor
  • Survivor
  • 8,444 posts
  • LocationIn my Mech (Atlanta, GA)

Posted 03 March 2016 - 09:27 AM

View PostL A V A, on 03 March 2016 - 09:20 AM, said:


Exactly.

Like I said before, it all comes down to damage, boating massive alphas and using your team mates as meat shields.


Well... good damage doesn't require big alphas. And in the PUG queue, scoring well usually requires willingness to sacrifice your own robot's health for a better position.

In other words, good damage requires tenacity and willingness to soak damage for your team--not using them to soak it for you. The trick is staying alive while dishing it out until the end of the game.

If you soak damage for them and put out damage while not dying, you're going to help them win far more often than not.

But that's too hard a system to implement--one that measures damage output plus damage taken. An ideal one would penalize players who do low damage but die fast while giving huge rewards to players who do big damage but also soak big damage--much bigger than players who do big damage but take little damage.

#57 Lily from animove

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • The Devoted
  • The Devoted
  • 13,891 posts
  • LocationOn a dropship to Terra

Posted 03 March 2016 - 09:30 AM

View PostSandpit, on 02 March 2016 - 10:12 AM, said:

I don't get things like this. Russ specifically stated they're looking at making a minor adjustment to give a little less weight to W/L and a little more to personal match score.

It's not like they said they're tossing out the formula in favor of individual match scores. It's not going to increase "selfish" behavior, it's going to reward the players who get stuck on team derp, do really well individually, but have zero chance to win the match because 3 guys on their team decided to rambo off and die 30 seconds into the match.

THAT'S selfish behavior.


yeha this, and funfact is the stupid rambos on the winning team becaue they had 4 skilled guys carrying them will still get an up. But this is the issue this is 6 derps in a game and all 6 should by PSR be downvoted, while the current system just upvotes 3 of them and downvotes the other 3 of them.

yet a good system would up all 4 good carriers.

But scoring has it's part in this bad judgement, because actualyl the game is not able to identify how much of an individual players performance added to a win. Make a capture win and you have a lousy score. While in fact those people capturing were the ones repsosible for the win.

One of the first things is to give every gammode an OWN score system so that skirmish rates damage and kills normal, while in assault and conquest damage and kills gte rated a LOT lower in score.

View PostMister Blastman, on 03 March 2016 - 09:27 AM, said:

But that's too hard a system to implement--one that measures damage output plus damage taken. An ideal one would penalize players who do low damage but die fast while giving huge rewards to players who do big damage but also soak big damage--much bigger than players who do big damage but take little damage.


The danger here is that people in the end go and let themself get damage just to game the system. But I guess there won't ever be a perfect system. Yet we should surely find a better one than the current.

Edited by Lily from animove, 03 March 2016 - 09:35 AM.


#58 Mister Blastman

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Survivor
  • Survivor
  • 8,444 posts
  • LocationIn my Mech (Atlanta, GA)

Posted 03 March 2016 - 09:37 AM

View PostLily from animove, on 03 March 2016 - 09:30 AM, said:

The danger here is that people in the end go and let themself get damage just to game the system. But I guess there won't ever be a perfect system. Yet we should surely find a better one than the current.


If there is a system it can be gamed. Humans are resourceful creatures.

#59 L A V A

    Member

  • PipPipPipPipPipPip
  • The Grizzly
  • The Grizzly
  • 308 posts
  • LocationOn the beach!

Posted 03 March 2016 - 09:39 AM

View PostMister Blastman, on 03 March 2016 - 09:27 AM, said:


And in the PUG queue, scoring well usually requires willingness to sacrifice your own robot's health for a better position.


Oh come on.

Certainly you realize that many folks, especially high tiered ones, do everything they can to avoid damage.

If they have even a whiff of a possible loss coming, they will hide in the back and spend the whole game doing whatever it takes to stay alive so that they can accumulate enough damage that the resulting loss will not effect their PSR.

You think folks get to Tier 1 by sacrifice their robot's health in Pug games? Posted Image

#60 Lily from animove

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • The Devoted
  • The Devoted
  • 13,891 posts
  • LocationOn a dropship to Terra

Posted 03 March 2016 - 09:42 AM

View PostMister Blastman, on 03 March 2016 - 09:37 AM, said:


If there is a system it can be gamed. Humans are resourceful creatures.


No stone, paper, sciccors can't be gamed. because good systems cannot be gamed sicne they are fair or balanced, bad ones however. they can.

View PostL A V A, on 03 March 2016 - 09:39 AM, said:


Oh come on.

Certainly you realize that many folks, especially high tiered ones, do everything they can to avoid damage.

If they have even a whiff of a possible loss coming, they will hide in the back and spend the whole game doing whatever it takes to stay alive so that they can accumulate enough damage that the resulting loss will not effect their PSR.

You think folks get to Tier 1 by sacrifice their robot's health in Pug games? Posted Image


avoiding damage, YES but not hiding. because hiding is when you miss the opportuntaties to get in better positions or doing damage. Good gameplay even in pug relies on aggressive gameplay with a good portion of risk. (And seismic)

Edited by Lily from animove, 03 March 2016 - 09:43 AM.






3 user(s) are reading this topic

0 members, 3 guests, 0 anonymous users