40 replies to this topic

[XxXAbsolutZeroXxX]

It seems there are 2 main aspects of the current PSR system people take issue with.

-Current PSR tiers are currently bound to matchmaker win/loss. The concern here is someone could be the best pilot in the game and they would lose every match if the matchmaker gave them a bad enough team. How good or bad a team the matchmaker allocates is a prime factor determining whether PSR rises or falls. Individual skill and team contributions are distant seconds.

-Damage is the 2nd mitigating factor which complicates matters due to damage being a relative rather than static variable. Depending on game modes and other factors there will be some games where each player scores an average of 400-500 damage on a win/loss. Other times each player will have an average of 200-300 damage at end game. That can make matters difficult if a hard coded static number is used to determine PSR increase/decrease. If damage and match score are relative rather than static variables, It might be more fair to grade on a curve.

The simplest (and possibly dumbest) solution might therefore be curve grading.

Top 4 match scores - PSR increases.
Middle 4 match scores - PSR stays the same.
Bottom 4 match scores - PSR decreases.

This would mean that as long as someone's match score was in the top 4 for their respective team their PSR would increase, whether they win or lose the game.

[Chuck Jager]

It is a ton easier to get high damage than it is to win or lose.

Most losses come from many team members being more concerned about their damage over helping a team.

The PSR calculations were based on numbers that were developed over a 2-4 years and do not take into account the sudden movement of many very new players into tiers that used to be separated by longer periods of time.

IF you think that gaining a tier level is a reward or a recognition of your skill especially if you do not actively participate in team games, this is the wrong game.

If you do play a ton in the group Q you should know how easy it is to move up by getting tons of wins just by being on a decent team that has many good players.

[El Bandito]

I Zeratul I, on 26 March 2016 - 07:11 PM, said:

It seems there are 2 main aspects of the current PSR system people take issue with.

-Current PSR tiers are currently bound to matchmaker win/loss. The concern here is someone could be the best pilot in the game and they would lose every match if the matchmaker gave them a bad enough team. How good or bad a team the matchmaker allocates is a prime factor determining whether PSR rises or falls. Individual skill and team contributions are distant seconds.

-Damage is the 2nd mitigating factor which complicates matters due to damage being a relative rather than static variable. Depending on game modes and other factors there will be some games where each player scores an average of 400-500 damage on a win/loss. Other times each player will have an average of 200-300 damage at end game. That can make matters difficult if a hard coded static number is used to determine PSR increase/decrease. If damage and match score are relative rather than static variables, It might be more fair to grade on a curve.

The simplest (and possibly dumbest) solution might therefore be curve grading.

Top 4 match scores - PSR increases.
Middle 4 match scores - PSR stays the same.
Bottom 4 match scores - PSR decreases.

This would mean that as long as someone's match score was in the top 4 for their respective team their PSR would increase, whether they win or lose the game.

Nope. Your concern about "someone could be the best pilot in the game and they would lose every match if the matchmaker gave them a bad enough team", is simply not valid. I have not seen in solo-q a known good player having less than 1:1 WLR, let alone the best. Now in group queue, that can be the case, but group queue and solo-q PSR by all means should be separate in the first place.

PSR right now needs less value on damage and slightly more penalty when losing. That's about it.

Edited by El Bandito, 26 March 2016 - 08:27 PM.

[XxXAbsolutZeroXxX]

El Bandito, on 26 March 2016 - 08:24 PM, said:

Nope. Your concern about "someone could be the best pilot in the game and they would lose every match if the matchmaker gave them a bad enough team", is simply not valid. I have not seen in solo-q a known good player having less than 1:1 WLR, let alone the best. Now in group queue, that can be the case, but group queue and solo-q PSR by all means should be separate in the first place.

PSR right now needs less value on damage and slightly more penalty when losing. That's about it.

Think of how many times you've seen people talk about their 20 game losing streaks.

In some cases, those losing streaks aren't due to individual skill level or lack of being a good team player. They're a result of the matchmaker repetitvely saddling them with teams that make winning impossible.

El Bandito if I made 11 clones of the worst player in this game and you played 20 games with 11 of those clones on your team, you would lose all 20 games. PSR isn't determined as much by skill or being a good team player as it is by the hand of cards the matchmaker dealt you.

If the matchmaker gives you 20 games where you have good teammates you'll probably win those 20 games and your PSR will increase.

On the flip side if the matchmaker gives you 20 games with bad teammates you'll probably lose all 20 games and your PSR will decrease.

How good or bad you are, whether you spot targets and tag for LRM's, scout, how much damage you do all those issues are secondary concerns. The matchmaker is the main variable that will determine how well you do, and its the main variable that determines what your overall PSR will be.

That would seem to be the main shortcoming with PSR.

PSR tier rankings aren't an indication of individual skill, they're moreso an indication of how good or bad the teams the matchmaker dealt you were in the long run.

This is pointless to discuss, I doubt it matters whether what I said was accurate or not. Accuracy doesn't seem like its important these days. I'm just throwing this out there to offer a different (?) perspective half expecting someone to have said this repeatedly in the 6 months I was gone from the game. I'm surprised no one seems to have brought this up before?

Edited by I Zeratul I, 26 March 2016 - 08:55 PM.

[El Bandito]

I Zeratul I, on 26 March 2016 - 08:49 PM, said:

Think of how many times you've seen people talk about their 20 game losing streaks.

In some cases, those losing streaks aren't due to individual skill level or lack of being a good team player. They're a result of the matchmaker repetitvely saddling them with teams that make winning impossible.

El Bandito if I made 11 clones of the worst player in this game and you played 20 games with 11 of those clones on your team, you would lose all 20 games. PSR isn't determined as much by skill or being a good team player as it is by the hand of cards the matchmaker dealt you.

If the matchmaker gives you 20 games where you have good teammates you'll probably win those 20 games and your PSR will increase.

On the flip side if the matchmaker gives you 20 games with bad teammates you'll probably lose all 20 games and your PSR will decrease.

How good or bad you are, whether you spot targets and tag for LRM's, scout, how much damage you do all those issues are secondary concerns. The matchmaker is the main variable that will determine how well you do, and its the main variable that determines what your overall PSR will be.

That would seem to be the main shortcoming with PSR.

PSR tier rankings aren't an indication of individual skill, they're moreso an indication of how good or bad the teams the matchmaker dealt you were in the long run.

This is pointless to discuss, I doubt it matters whether what I said was accurate or not. Accuracy doesn't seem like its important these days. I'm just throwing this out there to offer a different (?) perspective half expecting someone to have said this repeatedly in the 6 months I was gone from the game. I'm surprised no one seems to have brought this up before?

There were a lot of threads asking for personal score only PSR, ever since September. And I disagreed with all of them. This is a team based game, like all other team based games. One should not fault PSR if the chief culprit is the MM giving you 11 trash players every single match. Tell PGI to fix the MM formula (tightening tier difference, or take account of ECM mechs per team) or find a way to attract more players to the game.

And those players who have 20 losing streak in a row? If they are actually good players, their WLR will still stay above 1:1.

Edited by El Bandito, 26 March 2016 - 09:08 PM.

[sycocys]

My tier level is pretty well based on my ability to consistently shell out 500+ damage in solo queue.

~50% w/l ratio, but I climb steadily because when I drop in solo my PSR ranking almost always goes upwards or stays even.

Group queue my numbers drop, but I also play with mechs that aren't built for group queue play.

[Davers]

I find it somewhat unbelievable that a player who is consistently doing 'good work' would lose 20 games in a row.

[XxXAbsolutZeroXxX]

El Bandito, on 26 March 2016 - 09:07 PM, said:

There were a lot of threads asking for personal score only PSR, ever since September. And I disagreed with all of them. This is a team based game, like all other team based games. One should not fault PSR if the chief culprit is the MM giving you 11 trash players every single match. Tell PGI to fix the MM formula (tightening tier difference, or take account of ECM mechs per team) or find a way to attract more players to the game.

And those players who have 20 losing streak in a row? If they are actually good players, their WLR will still stay above 1:1.

Let's say we agree the matchmaker is flawed.

Why not eliminate the MM from the PSR equation?

At the end of game top 4 highest match scores (on a winning or losing team) have their PSR increase. Middle 4 match scores stay the same. Bottom 4 match scores decrease.

This way if everyone on a team does 20 damage and one person does 400 damage their PSR can still increase and people won't be penalized for limited matchmaker selection pool and assorted errors.

Davers, on 26 March 2016 - 09:14 PM, said:

I find it somewhat unbelievable that a player who is consistently doing 'good work' would lose 20 games in a row.

It happened to me once when I was on a lot.

Its almost like if you're on 6-8 hours a day or more there's a switch that kicks in guaranteeing you'll lose every game to boot you off the server.

[El Bandito]

I Zeratul I, on 26 March 2016 - 09:21 PM, said:

Let's say we agree the matchmaker is flawed.

Why not eliminate the MM from the PSR equation?

At the end of game top 4 highest match scores (on a winning or losing team) have their PSR increase. Middle 4 match scores stay the same. Bottom 4 match scores decrease.

This way if everyone on a team does 20 damage and one person does 400 damage their PSR can still increase and people won't be penalized for limited matchmaker selection pool and assorted errors.

I told the others, and I will tell it to you: Score only PSR change will turn this game into "Hide behind teammates and farm points warrior: Online". There will be no reason to be the first one to brave the enemy fire, no reason to lead the charge, no reason to act as a bait/distraction for the enemy, and even less reason to play Lights and Mediums. No reason to care about victory, as long as you get enough points to be the top 4 in the scoreboard. That's very flawed.

And removing MM will quite literally chase away the few thousand Steam players PGI managed to retain. I don't want to play against complete newbies, or even T3 players, and I am sure most of them do not wish to face me and get obliterated without doing much in return. All popular online games have some sort of matchmaking in them, and for good reason.

Edited by El Bandito, 26 March 2016 - 09:35 PM.

[Chuck Jager]

I Zeratul I, on 26 March 2016 - 09:21 PM, said:

Let's say we agree the matchmaker is flawed.

Why not eliminate the MM from the PSR equation?

At the end of game top 4 highest match scores (on a winning or losing team) have their PSR increase. Middle 4 match scores stay the same. Bottom 4 match scores decrease.

This way if everyone on a team does 20 damage and one person does 400 damage their PSR can still increase and people won't be penalized for limited matchmaker selection pool and assorted errors.



It happened to me once when I was on a lot.

Its almost like if you're on 6-8 hours a day or more there's a switch that kicks in guaranteeing you'll lose every game to boot you off the server.

Many game in the last 2 weeks I am that man and my PSR usually stays the same. What this does is make me want to make a new alt account and play the first 30 matches in the worst trial mech I can find. It appears that PGI (completely tinfoil hat theory) does have a simple algorithm to identify better new accounts and move them up faster if they hit certain markers.

Right now the PSR system and MM seems to reward lower tier players better. My wait times are usually less than 1 min in a heavy, but I am pretty sure I am being used to balance my team. The sad fact is I may sometimes get to play on teams with comp players, but I am not one of them. An even sadder fact is that when I see a comp player on the other side with no other recognizable META players it is usually a landslide in our favor. (note solo pug Q only)

I would happily wait longer for only a 2 tier spread.

[XxXAbsolutZeroXxX]

El Bandito, on 26 March 2016 - 09:33 PM, said:

I told the others, and I will tell it to you: Score only PSR change will turn this game into "Hide behind teammates and farm points warrior: Online". There will be no reason to be the first one to brave the enemy fire, no reason to lead the charge, no reason to act as a bait/distraction for the enemy, and even less reason to play Lights and Mediums. No reason to care about victory, as long as you get enough points to be the top 4 in the scoreboard. That's very flawed.

Good point and a slippery slope to boot.

But is there any real incentive to do any of those things you mentioned now? Over the past few days I've seen a lot of people completely give up on winning the game and build their tactics around raising their KDR instead. It seems common for some to wait till their entire team is gone before they enter the fight hoping to pick off a few easy kills like vultures.

Damage is the main component of match score and it already seems common for some to use teammates as human shields and play selfishly to rack up e-brownie points. I don't think that anything would change people wanting to win the game and most focusing on that goal. But it is truly difficult not to underestimate what people are capable of and how impossible it is to create systems that are idiot proofed against exploitation and abuse. So maybe I'd have to plead the fifth on that.

Edited by I Zeratul I, 26 March 2016 - 09:48 PM.

[El Bandito]

I Zeratul I, on 26 March 2016 - 09:46 PM, said:

Good point and a slippery slope to boot.

But is there any real incentive to do any of those things you mentioned now? Over the past few days I've seen a lot of people completely give up on winning the game and build their tactics around raising their KDR instead. It seems common for some to wait till their entire team is gone before they enter the fight hoping to pick off a few easy kills like vultures.

Damage is the main component of match score and it already seems common for some to use teammates as human shields and play selfishly to rack up e-brownie points. I don't think that anything would change people wanting to win the game and most focusing on that goal.

Over the past few days there is this Archer Leaderboard tournament going on. A lot of Archer players want kills to boost their scores on board, so they are playing very selfishly. The phenomenon might subside, once the event is over. And I already mentioned in my first post in this thread that damage should be valued less in PSR calculation.

Edited by El Bandito, 26 March 2016 - 09:50 PM.

[KodiakGW]

Different people keep making new posts about the PSR system. Posting up screen shots of teams failing miserably while they are one of the top damage players.

The white knights roll in and say "Nothing to see here folks, move along."

People are frustrated because they are not seeing good games. They are tired of 12 mans where it seems that they have to play godlike to make sure the team gets a win. If they don't, the team will will end up folding and causing yet another loss.

When people get frustrated with a game they do two things. They don't spend money on it, and eventually they stop playing.

So, just continue staying the course, thousand points of light. Then we'll all see what happens. Either that, fix PSR so it's not an XP bar, or introduce new modes where personal skill is a deciding factor.

[Chuck Jager]

I am tired of seeing 1 red dorito come up over 600m away and watching 3 of my team mates not even fire once and run 1-2 squares away on my t2 account.

If someone has less than 100 games in group Q they should never make t3 even if they are an alt account seal clubbing. I would rather see the alts terrify the seals than have the stupid seals be on my team. IE taste of their own medicine.(or PGI could add a few more line to their MM algorithms)

[Navid A1]

I don't have any problem with PSR being based on wins and losses.

However,

- PSR MUST be per weight class and ideally, per chassis.
- Game score MUST give rewards to things other than doing damage... I love to NARC. and PGI likes to show me the finger when it comes to rewards..


Also... PGI should stop catapulting new players into T3 as soon as they win their first couple of matches.

The amount of brain-deads in solo queue is off the charts... you can actually tell the result of a match before the match even starts. Who are these people?... How did they even get to T3? And why do they love to use VOIP to suggest horrible battle strategy?
I'm explaining things like ecm to some of them and how its not a hack... during a match.

Edited by Navid A1, 27 March 2016 - 12:55 AM.

[Brizna]

It's strange becuase if anything I would remove completely the influence of that super flawed score formula influence on any kind of PSR calculation.

I've been a great asset for my team with under 100 damage many time, believe it or not there's so much more to wining a game than DPS alone. People who think damage is paramount usually only drive heavies and assaults and don't see the battles fromt he skirmishers perspective. A good light job is the difference between pushing against 5 mechs waiting for you or a bunch of stressed mechwarriors looking each to a different place.

[XxXAbsolutZeroXxX]

KodiakGW, on 26 March 2016 - 10:24 PM, said:

People are frustrated because they are not seeing good games. They are tired of 12 mans where it seems that they have to play godlike to make sure the team gets a win. If they don't, the team will will end up folding and causing yet another loss.

Exacta. If I were to summarize the worst aspect of this game it would go something like: who wants to spend hours being molested by the random mood swings of MWO's matchmaking software, which decides they might lose 20 straight games because PSR rank doesn't measure individual skill. PSR skill rating is based on whether someone was in the right place, or on the right team, at the right time to get into the 'W' bracket. That measures the collective probability of them having good teammates rather than their intrinsic individual skill. Does that make sense? I'm not saying that I have it all figured out or that I know the answer or probable cause but matchmaker dependability could still one of the biggest and toughest to define issues which many strangely turn a blind eye to.

In worst case scenarios its like playing 20 hands of poker and being dealt a crap hand everytime. Or flipping a coin 20 times and the outcome being heads or tails on every flip. The matchmakers tendency to defy the laws of probability are difficult to explain.

Generally people want to win or lose based on whether their choices & actions meritted it.

[maniacos]

sycocys, on 26 March 2016 - 09:12 PM, said:

My tier level is pretty well based on my ability to consistently shell out 500+ damage in solo queue.

~50% w/l ratio, but I climb steadily because when I drop in solo my PSR ranking almost always goes upwards or stays even.

Group queue my numbers drop, but I also play with mechs that aren't built for group queue play.

You can not do that with every mech. You won't shell out 500+ damage in a low-alpha light in every losing game, especially when you do the role a light mech was meant for.

[PurpleNinja]

Myabe, just maybe, you should work to win the game.
Problem solved.

[maniacos]

PurpleNinja, on 27 March 2016 - 03:46 AM, said:

Myabe, just maybe, you should work to win the game.
Problem solved.

Another one who has not gotten the point.