31 replies to this topic

[MoonUnitBeta]

Chuck YeaGurr, on 30 September 2015 - 08:20 AM, said:

Learn to math
bad team but good performance PSR stays same or slightly drops
Good team good performance bigger jumps in psr than drop in losses (why people are upset with just playing to move up)

Over 1000 games you should draw close to equal team quality for wins and losses. Variance in random number is offset by bigger reward for win.

Only constant factor guess what - YOU. IF you constantly play well you move up. (note) there are ways to inflate numbers that help epeen, but they let the team fall short. Number padding by never being the person in front will result in consistently letting the team down. You will get more losses.

(total wins) adjusted by (consistent good play in losses) over (extended period of time) = movement up in psr

Your performance in the loss does affect how you move up even a negative arrow can still be less than a worse performance. Do not just look at the arrow and make assumptions.

This is a math problem with multiple parts

Yes math is learnt. Thanks for the outstanding lesson. But like I said, PSR is based on perfect world math. By your logic, (you being the constant variable) ELO is a perfectly fine rating system, because EVENTUALLY your 8.3% contribution to all the matches you played will EVENTUALLY make a difference. But ask yourself these questions: Who has that time? And,
Does PSR reflect the most relevant skill of the player? (think: learning curve! side questions: does our current PSR rating keep up with us, or is it only an accurate representation of our skill when we plateau and stay there?)

Here’s my answers: very few, and no.

The argument here is giving out the most accurate rating in the fastest possible fashion.

With ELO (as an example, please hang in there) you’ll get to the top, eventually, if you play THOUSANDS of matches for the game to place you in the correct area. I’ll keep saying this – you only make up 8.3% of your team. And yes, eventually with PSR you are placed into a tier that’s somewhat close to where you place. But worse yet, you make up 4.2% of the overall match players. The system is assuming that it’s placing you correctly with all those people. It takes a ridiculous amount of time for you to climb into the area where you’re supposed to be.

Like I said, in a perfect world, with every player who has hundreds of hours of free time, or plays the game straight for months and months, they’ll end up in the right tier. But that’s just the same with ELO.
As PGI mentioned, PSR only solves the problem with ELO that your team can carry you into higher tiers.

Are you saying that PSR is perfect? I don’t. This is what we can do to improve it. Can we please take Paul’s arbitrary rating numbers, and instead let the match and your team dictate whether you’re placed correctly or not?

Edited by MoonUnitBeta, 30 September 2015 - 09:11 AM.

[Almond Brown]

MoonUnitBeta, on 30 September 2015 - 09:08 AM, said:

Yes math is learnt. Thanks for the outstanding lesson. But like I said, PSR is based on perfect world math. By your logic, (you being the constant variable) ELO is a perfectly fine rating system, because EVENTUALLY your 8.3% contribution to all the matches you played will EVENTUALLY make a difference. But ask yourself these questions: Who has that time? And,
Does PSR reflect the most relevant skill of the player? (think: learning curve! side questions: does our current PSR rating keep up with us, or is it only an accurate representation of our skill when we plateau and stay there?)

Here’s my answers: very few, and no.

The argument here is giving out the most accurate rating in the fastest possible fashion.

With ELO (as an example, please hang in there) you’ll get to the top, eventually, if you play THOUSANDS of matches for the game to place you in the correct area. I’ll keep saying this – you only make up 8.3% of your team. And yes, eventually with PSR you are placed into a tier that’s somewhat close to where you place. But worse yet, you make up 4.2% of the overall match players. The system is assuming that it’s placing you correctly with all those people. It takes a ridiculous amount of time for you to climb into the area where you’re supposed to be.

Like I said, in a perfect world, with every player who has hundreds of hours of free time, or plays the game straight for months and months, they’ll end up in the right tier. But that’s just the same with ELO.
As PGI mentioned, PSR only solves the problem with ELO that your team can carry you into higher tiers.

Are you saying that PSR is perfect? I don’t. This is what we can do to improve it. Can we please take Paul’s arbitrary rating numbers, and instead let the match and your team dictate whether you’re placed correctly or not?

So your real quest here is to move UP Tiers Faster than the current system allows? Is that it? One player, that 8.3% can and do get poor Match scores at the end of the Match and yet they ARE the reason the Team Wins. Many scenarios support that conclusion as well.

So you want to punish those who did what was needed to WIN but in doing so sacrificed their Match score?

P.S. 1 Mech allowing 3 to chase them around for 2.5 or 3 minutes before dying, thus allowing the Others to wipe up the out numbered enemies left behind should get LESS credit for that WIN?

Seeing as no one can factor in the value of the "intangibles", the TEAM's results as a whole must be given precedence.

Edited by Almond Brown, 30 September 2015 - 09:28 AM.

[Lugh]

IF I outscore everyone else on the team even in a loss, I should have a small increase never an =.

There is still something off in the Damage calculations. If I am super crafty and get BEHIND my opponents and kill 3 before they can turn I very well may have only done 120 damage and gotten three kills ON FRESH MECHS. The damage calculation for kills needs to take the total amount of health in the mech and award it as Damage to anyone that participated in the damage on that mech.

This way the super accurate with the head shot guy gets credit for removal of the mech from the game at the same rate as the all over the place spray and prayers that do more overall damage but are less efficient at killing the things they are shooting.

It makes light pilots awarded less match score consistently because of their ability to out maneuver the enemy on a regular basis.

It makes accurate snipers get awarded less as well. I take out a head, that's what 35 damage? Yet 600 damage worth of Assault soak has been removed from the game. Match score needs each mech, by type, award a score that is consistent with its battlefield value regardless of how efficiently it got killed. Damage can then add less to the match score.

Additionally, an Assault mech, Heavy Mech, Medium mech, light mech, that does all the 'tanking' of damage for the team through good twisting etc would receive an award based on How much Damage it absorbed during the course of the match.

[kapusta11]

The true problem is how does one evaluate "good" performance? Is it how much damage you do per game? Hardly. Is it how many kills you score? Maybe, depends on who you've been facing a veteran or a beginner. Is it about that good flanking maneuver? Then how is it represented in any stat?

Edited by kapusta11, 30 September 2015 - 09:44 AM.

[MoonUnitBeta]

Almond Brown, on 30 September 2015 - 09:25 AM, said:

So your real quest here is to move UP Tiers Faster than the current system allows? Is that it? One player, that 8.3% can and do get poor Match scores at the end of the Match and yet they ARE the reason the Team Wins. Many scenarios support that conclusion as well.

So you want to punish those who did what was needed to WIN but in doing so sacrificed their Match score?

P.S. 1 Mech allowing 3 to chase them around for 2.5 or 3 minutes before dying, thus allowing the Others to wipe up the out numbered enemies left behind should get LESS credit for that WIN?

Seeing as no one can factor in the value of the "intangibles", the TEAM's results as a whole must be given precedence.

Up or down faster. With the ability to hold a player where they need to be if they are having good quality matches.

Listen - just because the player WINS THE MATCH, doesn’t mean that they’re a Tier 1 player. It means they won the match. So what? What matters more is if it was a GOOD match or NOT. If they did 50 match score, but the rest of their team did 200… 300. Etc, ( the point is he did terrible IN COMPARISON TO HIS TEAM, not in comparison to Paul's table of performance) then he is going to suffer a greater loss to his PSR rating because he’s obviously in an area that’s not suitable for him. We can also compare his team, to the other team, to gauge how the match went, and come to a round about conclusion about why his match score was so low.

When you’re talking about rating a player – why is bringing them down a PUNISHMENT – as if people in higher tiers are getting a c-bill multiplier…? As if being in a low tier is a BAD THING. The whole point is to get people into an area that’s appropriate for them – who cares what tier their in, as long as they’re getting quality matches?

kapusta11, on 30 September 2015 - 09:42 AM, said:

The true problem is how does one evaluate "good" performance? Is it how much damage you do per game? Hardly. Is it how many kills you score? Maybe, depends on who you've been facing a veteran or a beginner. Is it about that good flanking maneuver? Then how is it represented in any stat?

Yea, to some degree flanking is taken into consideration. There’s a flanking reward, protection rewards, and other rewards that have a match score rating. Ideally, you should be getting rewards that are appropriate to what you do in the match. Right now, match score achieves that to some degree, but PSR’s effectiveness at placing a player can still be improved upon. Paul’s fixed table is still used as a reference in the background as a guide to define good/bad, but that’s still very different than our current system of it literally tell you if you did good or not, a made up metric that paul placed his finger on.

You can define “good” performance anyway you want, whether it’s 4000 or 92, the number itself is arbitrary, and it’s only given definition when you have something to compare those numbers to. PSR achieves this with Paul’s table of fixed values, and players according to that table get placed accordingly, but eventually. And it takes a long time.

If we bring in overall match quality, and the team you were paired up with, it will become easier to define what was good, and what was a bad performance. The purpose of PSR and the tier system is to group like-skilled players together for quality matches. It’s rather difficult (and it’s the reason why it takes so long) to get those skills players together in a specific tier.

Edited by MoonUnitBeta, 30 September 2015 - 10:05 AM.

[Death Proof]

SuomiWarder, on 29 September 2015 - 08:11 PM, said:

But i would image that PGI thinks they are encouraging "team play" and that a system where you do best by being the highest scorer over your team mates is not "team play".

I'm 100% certain this is their reasoning.

Though, they could achieve the same thing without making PSR so dependent on the performance of the other 11 people on your team by rewarding the players more for winning and less for losing.

[MoonUnitBeta]

Almond Brown, on 30 September 2015 - 09:25 AM, said:

P.S. 1 Mech allowing 3 to chase them around for 2.5 or 3 minutes before dying, thus allowing the Others to wipe up the out numbered enemies left behind should get LESS credit for that WIN?

Seeing as no one can factor in the value of the "intangibles", the TEAM's results as a whole must be given precedence.

Sorry I was at work when I did my last response - I had to skim, so my earlier response didn't cover this. But now I can cover this in full, and I'll help clear up some misconceptions you have about PSR, match score, their meaning, and what the real reward should be for the player.

Here we go:
- - -

That's an incredibly specific scenario. In that particular situation, I'd be passing the blame onto the match score/rewards system for not being able to identify that kind of tactic, or team work. But, even though it may be an intangible, it doesn't mean that it goes unrewarded... (I'll touch on this in a minute)

You know this already, but it's worth repeating: PSR (the tier classification for a player) and match score (the outcome of your match performance) are two entities working together to define a player's overall skill rating. I kinda see PSR as a tub, and match score as the guy operating the faucet and drain. Maybe that a bad analogy - it's the best I got so far. lol

But that's it. It's a rating, and nothing more. They work together. Actually their sole existence is to simply ensure that matches, and the people you play with, are fair and play on the same level as you.

"1 Mech allowing 3 to chase them . . . should get LESS credit for that WIN? . . . the TEAM's results as a whole must be given precedence."

Lastly - If i was the mech kiting the 3 mechs around the real reward is not going up, or down, or staying at the same PSR level... The real reward is the c-bill bonus for winning. Not the boost or reduction in PSR. My reward for doing that, is being rewarded for the win. C-bills are more important to me than what tier I'm in. Good matches, are more important to me than what tier I'm in.

While I don't understand how you see getting reduced in PSR as a punishment, I'll just reiterate my understanding of the system: Tiers(PSR), they don't get you c-bills. It's not a rewards system. Tiers are just a tool to increase match quality. Further more, Match score is a tool that's used to help get you into the correct Tier. Whether you go up, or go down, it's neither punishment nor reward. It's just a rating, nothing more.


I hope this makes sense. I hope this helps you understand what the real importance of a win is. It's not a rating, it's actual rewards. That's the reward. That's why wins should mean very little for PSR. The literal closeness of a win is what should count the most, as an aid in how people should be getting rated.

Thanks for reading (again!)

Edited by MoonUnitBeta, 30 September 2015 - 07:02 PM.

[Simbacca]

PSR should definitely look at overall team performance as well. If players are constantly on the losing team, and with many of those losses being by a huge to a massive margin - then that needs to be taken into account as a form of check that the PSR rating given to the player is the correct one.

[MoonUnitBeta]

Simbacca, on 30 September 2015 - 07:44 PM, said:

PSR should definitely look at overall team performance as well. If players are constantly on the losing team, and with many of those losses being by a huge to a massive margin - then that needs to be taken into account as a form of check that the PSR rating given to the player is the correct one.

Yep! Right now, PSR does that currently for that particular scenario (by handing out large reductions for low performance on the losing side).

But when it comes to poor performance on the winning side? Or doing very good on the losing side? There's quite a few assumptions being made, and it makes it difficult for efficient and appropriate PSR reduction/increases to be delivered. This is why it takes so long to move through the tiers. The bias towards win/loss compounds these issues. Match quality on the other hand (a combination of match outcome, team composition, and pairing) should take precedent, and the bulk of win/loss should be saved for determining rewards, not ratings (see above your post).

By taking teammates and teams into consideration, poor performing people will still suffer large reductions. Win/loss still has it's place in the calculation, but it would be far less influential on how your PSR variance is calculated. Determining your PSR variance would be more about how you did in the match you were matched with to deliver a faster and more accurate(relevant) adjustment to your PSR.

Edited by MoonUnitBeta, 30 September 2015 - 08:50 PM.

[Chuck Jager]

It does take time to move up and it takes a large number of games with good results in both wins or losses. If folks get bent because the little arrow say down they have a lot to learn about team play. We all loose even when we play well. It is the ability to consistently and overcome these mistakes averaged over a period of time that is the sign of a top player in any field.

Tier 1 players have to have not only good results, along with a massive number of games. They can move faster but they will need fewer games with a massive number of 600 scores. That is the way it is. This is not grade school not everybody is equal and gets to be line leader.

PSR represents a type of player it is not an achievement.

[MoonUnitBeta]

Chuck YeaGurr, on 30 September 2015 - 09:19 PM, said:

It does take time to move up and it takes a large number of games with good results in both wins or losses. If folks get bent because the little arrow say down they have a lot to learn about team play. We all loose even when we play well. It is the ability to consistently and overcome these mistakes averaged over a period of time that is the sign of a top player in any field.

Tier 1 players have to have not only good results, along with a massive number of games. They can move faster but they will need fewer games with a massive number of 600 scores. That is the way it is. This is not grade school not everybody is equal and gets to be line leader.

PSR represents a type of player it is not an achievement.

I'm not sure where you're sitting. Is your response directed to my OP, or just to people in general?
I'm pretty bored so I'll just respond so you know where I'm sitting as well... My responses below:


It does take time to move up and it takes a large number of games with good results in both wins or losses.
Agreed. But it doesn't have to take so long, and it's inaccurate for a long time if your skill is outside of your PSR.

If folks get bent because the little arrow say down they have a lot to learn about team play.
Agreed. But not everyone plays for T1. some play just for casual fun. If those types of people are getting upset by a down arrow, they don't understand what the rating system is for.

We all loose even when we play well.
Agreed. Why shouldn't the match maker acknowledge these efforts in a relevant fashion?

It is the ability to consistently and overcome these mistakes averaged over a period of time that is the sign of a top player in any field.
Agreed to some degree. Being a good player means you overcome and improve, but being a good player doesn't mean you outlasted a match maker.

Tier 1 players have to have not only good results, along with a massive number of games. They can move faster but they will need fewer games with a massive number of 600 scores. That is the way it is.
Agreed. And that is the way it would still be.

This is not grade school not everybody is equal and gets to be line leader.
Agreed. Who said differently? I'm not advocating equality, I'm advocating efficient and accurate ratings, and a way to ensure well matched games.

PSR represents a type of player...
Agreed. So why is wanting a speedy and accurate rating system so difficult for people to accept?
it is not an achievement...
Agreed. I've not once mentioned achievements.


---
I've just now realized that you and Almond Brown are completely different people! lol. I was responding to almond thinking it was you this whole time. lol. Not major, but I might have to re-read my responses because I think I accused almond about things that you were talking about lol.

Edited by MoonUnitBeta, 30 September 2015 - 11:32 PM.

[MoonUnitBeta]

Still would like to hear more people's thoughts about this.

A lot of people are finding that there's issues with PSR.
But there's also issues with Match Score too.
If we add relevancy to PSR, we can get grouped more efficiently.

I've been thinking and having atleast 2-3 more tier levels will help.
Right now you have T5 people mixing in with people much more skilled than them, which can result in wacky matches.