Jump to content

Psr Community Feedback - Round 1


357 replies to this topic

#81 ShogunKid

    Member

  • PipPip
  • The People's Hero
  • The People
  • 23 posts
  • LocationOklahoma

Posted 17 June 2020 - 02:51 PM

I vote 2A.


#82 Tarl Cabot

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Tai-sho
  • Tai-sho
  • 7,775 posts
  • LocationImperial City, Luthien - Draconis Combine

Posted 17 June 2020 - 02:55 PM

View PostTamerlin, on 17 June 2020 - 02:27 PM, said:

Does MWO allow for one team to have a group and not require the other team to have a like-size group? If so, it shouldn't.


IIRC, It is possible. Other issue is allowing a group to bring mechs of same weight class. There are the groups who will use the most out of the weight limit and make the other team pay for it, then there are the half-arse groups who bring snot-nose mechs to a drop. Mind you, both groups have been mostly Tier 1 players.

Truthfully, PGI could reduce group size to max 3 instead of 4 while restricting groups to 1 mech per weight class. It would reduce the lopsidedness of some of the tonnage differences, it will not fix those half-arse groups who should be in tier 3 or tier 4 instead of Tier 1. And the quicker PGI gets this change rolled out the better, though I really doubt PGI will review it 3 or months later. They never reversed nor made any alterations when it was changed from Elo to PSR/Tiers except to change newbies starting in Tier 5 instead of Tier 4.

#83 Jesslp

    Member

  • Pip
  • The Warden
  • The Warden
  • 17 posts
  • LocationEspaña

Posted 17 June 2020 - 02:56 PM

I vote Core 2B with JayZ's modifications

#84 BROARL

    Member

  • PipPipPipPipPipPip
  • General
  • General
  • 301 posts
  • Locationcommunity warfare

Posted 17 June 2020 - 03:13 PM

View PostNightbird, on 16 June 2020 - 05:21 PM, said:

No point crying over spilt milk so here's a lesson learned for MWO2.

The current MM creates a wide WLR range from <.5 to >2. The best and most important reason this has been bad for MWO and PGI is that player attrition rate is lowest when WLR=1. People quit when they win too little (too painful) and when they win too much (too easy).

Do a little calculation on overall pop and you can see the true monthly attrition rate is ~3% (not everyone that leaves for a season leaves forever).

If a Match Maker consolidated the WLR for all players into a 0.8-1.3 WLR region from the beginning of MWO, if you take the relative improvement from the graph above this would reduce the player attrition rate to 2%.

Such a change would result in a player base 3x the size today, and also generated for PGI 32% more lifetime revenue from MWO.

The MM I proposed on the last thread (with tweaks) would have done this. There are even better MM ideas (not presented due to difficulty) that would push the WLR for all players into a 0.95-1.05 region, and drive monthly attrition rate down to 1.5% and lifetime revenue up by 56%. (Yes I know this underestimates the benefits of a MM because it fails to take into account the influx of new players. Under-promise and all)

Amazing what a little MM can do right? Here's to hoping for success in the next multiplayer title.


^^^this^^^ in unreal4 (MW5) PvP campaign (in VR with melee) PLEASE!!!

#85 Xiphias

    Member

  • PipPipPipPipPipPipPip
  • Littlest Helper
  • Littlest Helper
  • 862 posts

Posted 17 June 2020 - 03:14 PM

View PostTermin8rSmurf, on 17 June 2020 - 02:00 PM, said:

While winning and losing are part of this game, placing a large percentage of the PSR rise/fall on that is a mistake.

All of PSR change should be based on winning/losing. Otherwise the system will diverge into two groups (top and bottom) instead of having five distinct tiers for the matchmaker.

Quote

A good pilot, solo dropping, outperforming his random pugs, and losing, should not be punished by his team under-performing. A good caller, making great calls, can still be ignored, can still be over-ruled, and can still have a loss, simply because he lost the pug lottery.

Meanwhile, someone with a terrible build, spending half his match AFK, returning in the last two minutes of a match, and doing very little to gain the win, will benefit greatly from not having died, from going to the bathroom, from being carried by his team.

Yes, some matches good players will go down. Some matches bad players will go up. The goal of the PSR system is to accurately assign a value to a player's skill, NOT to generate accurate adjustments every match. The latter is impossible without access to the global statistics, which PGI has stated aren't available to be used.

Stop focusing on individual matches. Will a good pilot solo dropping and consistently performing end up winning more than they lose? Yes. Will a bad pilot who spends half the match AFK lose more matches than they win? Yes. The individual matches don't matter as long as the trends are accurate.

Basing PSR movement by ranked matchscore will push players either to T1 or T5 depending on if they are over or under that community average match score. Based on your historic stats, that means you would probably end up in T5. Are you okay with that?

This is the difference between a type 1 system (orange) and a type 2 system (blue) after enough matches.
Spoiler


Quote

PSR should reflect the Skill of the Pilot.

When you go for your driving test, you will pass or fail, based solely upon your ability to drive the vehicle. It's not based on how Mrs Johnson was walking her pooch down the street, Nor how the postman was emptying the mailbox. Nor how nice the flowers look in the florist's windows. No. Only on the driver's skill.

PSR should be the same.

Even with drivers test there is a huge skill difference between drivers on the road. It's also not an option to have players take a solo test to get rated (if you want that play Solaris).

Insurance companies effectively assign a "skill" (risk) value to drivers to determine their rates. How do they do this? They look at the overall trends and the individual performance of their customers. If you're constantly in crashes, even if not at fault, your rates are going to go up because you are more likely to get in accidents and probably are doing something that's getting you into crashes.

Skilled players win more on average, the matchmaker works on averages. No one is getting punished by having the system based on wins/losses

#86 Zanotam

    Member

  • Pip
  • Legendary Founder
  • Legendary Founder
  • 16 posts

Posted 17 June 2020 - 03:49 PM

So a good point I'm seeing now is that players who think we should use type1 should just play Solaris.... and I kinda gotta agree. THis is a TEAM game. We can use a Type 1 system in theory MAYBE, but if we do we gotta do something like 0 kicker matchscore for losses and 200 matchscore kicker for wins to really differentiate the two and focus on what MOSTLY matters in a statistical sense: Winrate!

EDIT: The important thing though IMO is to lower damage's kicker for match score, give a small kicker for damage taken to reward tanking, and then reset everyone's ELO with whatever system they want to start with so we can start getting data and better matches ASAP. Anythign would be better than the currently fossilized "almost everyone is tier 1 so matchmaker can't even tell people apart in the first place" thing.

Edited by Zanotam, 17 June 2020 - 03:51 PM.


#87 Tarl Cabot

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Tai-sho
  • Tai-sho
  • 7,775 posts
  • LocationImperial City, Luthien - Draconis Combine

Posted 17 June 2020 - 03:51 PM

Quote

Basing PSR movement by ranked matchscore will push players either to T1 or T5 depending on if they are over or under that community average match score. Based on your historic stats, that means you would probably end up in T5. Are you okay with that?


Isn't that making an assumption that the player would almost always be facing players who are hitting 300+ MS? He had peaked for about 1500 games, exceeding 200 -231 MS for several months (likely tier 3 ) and is likely in Tier 1 now, where he is facing a much tougher crowd.

The real issue is that the current PSR thresholds is that losing a match did not really hurt a player, the values themselves were not even close to the Win values. I am sure I am not the only one who was surprised by the stark difference. And definitely helped those players with sub-200 MS to push themselves into Tier 1 simply through by playing 10s of thousands of games, WHILE losing more matches than won.

And for some who do not remember or did not see the values in Paul's original thread.

Player LOSES:
  • Match Score: 0-100 goes down in PSR by -2
  • Match Score: 101-250 goes down in PSR by -1
  • Match Score: 251-400 does not move.
  • Match Score: 401+ goes up in PSR by +1

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

Edited by Tarl Cabot, 17 June 2020 - 03:54 PM.


#88 Xaat Xuun

    Member

  • PipPipPipPipPipPipPip
  • The Defender
  • The Defender
  • 954 posts
  • LocationA hypervelocity planet

Posted 17 June 2020 - 04:10 PM

1B grabbed my Vote

2A&B, looks like it punishes a good player for being on a bad team

It's the kickers I'm more curious in how they'll affect the PSR

#89 Cluster Fox

    Member

  • PipPipPipPipPip
  • 104 posts
  • LocationStuck on a rock in Grim Plexus

Posted 17 June 2020 - 04:14 PM

Core 2B, backup 2A if we only have 4 choices...
JayZ's suggestions: Yes to season reset. Yes to the formula, No to going up on loss / down on win.

But:
- I'd rather see something based off average MatchScore and avgWLR.
- WLR needs to factor in heavily in the match maker, I think Nightbird prove this.
- A primarily WLR based MM would be great to reduce stomps.

However, reset or not this will diverge over time. JayZ's proposition to scale it is a bandaid, not a solution. Players playing more matches move quicker than casuals. Putting T2 players in T1 eventually and T4 players in T5. After enough games, Tier 3 is virtually empty.

Basically this: When do people stop moving down/up in tiers and stay where they really belong? - Never really.

Any Core1 solution needs a Win bonus which becomes a significant multiplier.
Like 150%, 200%even. Not a flat 50ms. Doesn't matter for Core2, which is good.

Again: I'd rather see something based off average MatchScore and avgWLR

Edited by Cluster Fox, 17 June 2020 - 08:34 PM.


#90 ESC 907

    Member

  • PipPipPipPipPipPip
  • The Slayer
  • The Slayer
  • 213 posts
  • Location'Murica

Posted 17 June 2020 - 04:23 PM

View PostZanotam, on 17 June 2020 - 01:55 PM, said:

If anyone has an idea for how to give a benefit to 'good ECM use' I would love to see that as well as I think damage taken when tanking properly, punishing losses/rewarding winners, and rewarding good ECM play need to be enhanced/increase if we're initially doing Core 1 systems (for which I vote for Core1a!)

Perhaps something that takes into consideration the "blanket" properties of ECM? Something that perhaps multiplies per friendly that is kept under your ECM bubble, and another for countering ECM too? Though I am fairly certain that there is at least something in-place already for countering ECM.

#91 Jay Z

    Member

  • PipPipPipPipPipPip
  • Deadset Legend
  • Deadset Legend
  • 436 posts
  • Twitch: Link
  • LocationAustralia

Posted 17 June 2020 - 04:38 PM

View PostCluster Fox, on 17 June 2020 - 04:14 PM, said:

However, reset or not this will diverge over time. JayZ's proposition to scale it is a bandaid, not a solution. Players playing more matches move quicker than casuals. Putting T2 players in T1 eventually and T4 players in T5. After enough games, Tier 3 is virtually empty.

Oh I know full well the 3 monthly PSR rescale is a bandaid, but it is one that seems reasonable to implement and works within the parameters. Every single option will of course cause uncontrolled divergence as there is no reference to global or stored stats.

View PostXiphias, on 17 June 2020 - 03:14 PM, said:

The goal of the PSR system is to accurately assign a value to a player's skill, NOT to generate accurate adjustments every match. The latter is impossible without access to the global statistics, which PGI has stated aren't available to be used.

100% the lack of access to global statistics hurts. I know I know I know WLR is a great metric but with 0 access to global or stored statistics it is outside of PGI's scope unless they change their scope, at which point hey WLR becomes a thing.

As far as I understand it, we are limited to individual matches with Matchscore as an input and Player PSR Shift as an output.

Edited by Jay Z, 17 June 2020 - 04:38 PM.


#92 Rkshz

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,866 posts
  • Twitch: Link
  • LocationOdesa, Ukraine

Posted 17 June 2020 - 05:39 PM

View PostPaul Inouye, on 16 June 2020 - 04:30 PM, said:

Core 1A: Global Compare

better than others


View PostNightbird, on 16 June 2020 - 05:21 PM, said:

No point crying over spilt milk so here's a lesson learned for MWO2.
Spoiler


it's interesting - totally agree

#93 OneTeamPlayer

    Member

  • PipPipPipPipPipPip
  • Ace Of Spades
  • 399 posts

Posted 17 June 2020 - 06:03 PM

As someone else said, they made a system to measure individual player skill- it's called Solaris and it's so much fun to gauge individual skill that the mode is depopulated almost 100% of the time. People who wanted their individual skill rating can't even be bothered to populate the mode that is defined by individual skill at its core. That should say volumes about basing the team mode on individual as opposed to team performance.


View PostESC 907, on 17 June 2020 - 12:35 PM, said:

IMHO, the people that are attempting to argue that players will "play the system" by intentionally tanking the scores of their teams are incredibly ignorant. What is the point of going up in PSR, if your W/L record gets tanked? What absolutely idiotic ******* would intentionally throw a match so that they themselves can go up? It will not be many. What makes you feel good in a game like MWO is winning and performing well, NOT being "high-level". When people boast about MWO, they post recordings or screenshots of their performance in a particular match, not their ranking.


If you've read the discussion on the changes you're well aware that many people think their tier is the most important metric of all.

Not everyone checks Jarl's, many people have never been there, and many more will never care about the stats there. Most people will never post a screenshot of this game, they will never post a recording, they will simply play the game and glance at whatever metric tells them how they've been doing over all the matches they've played. Some matches are good and some are bad and after a point the only thing that matters is the total progression to many players.

They will, however, chase that Tier 1 status like it's the holy grail.

Win Loss isn't displayed anywhere in-game so a majority of players don't give two shakes about win/loss other than the fact that currently winning improves players tier gain.

Look, i get that you guys play with normal competitive people, and that's great- I agree with you on a lot of points about how to play. I personally want close matches with active voice comms, a drop caller that people naturally listen to, and a fight to the death with everyone putting in their fair share because it's just more fun that way. MWO matches with a lot of communication and teamwork are what I remember as much more common during pre-Resize/Desync/Skill tree days and this game was a blast because of it.

Meanwhile 90% of the rest of this game isn't reading these forums and talks about Tier as if it's their newborn child because it's the only ingame metric that shows any sort of progress at all over a long period of time.

Once again designing systems for a game with a declining population around the (valid mind you) whims of a small group without considering the second order consequences to the majority of the playerbase which might have (and likely have) an entirely different view of the game than you do- to the point it looks alien from your vantage point.

It's shown on the front page, and it's something people stare at between every single match they play- of course they're going to think it's important over time.

Tell me I'm wrong if you want, tell me people don't chase the one simple metric they see most.

Cool, we'll see after the changes i guess.

Edited by OneTeamPlayer, 17 June 2020 - 06:07 PM.


#94 crazytimes

    Member

  • PipPipPipPipPipPipPipPip
  • Ace Of Spades
  • 1,366 posts

Posted 17 June 2020 - 06:04 PM

From my standpoint, I don't care what methodology you use. I am an about average player, I will end up around the middle of any PSR system.

All I really care about is "am I dropping in roughly balanced matches without waiting too long". As soon as the wait time pushes out much more than it currently is, then I won't bother playing. I doubt I'm alone in that.

What does this mean? However you sort people in to tiers- we all end up in the same matches anyhow, or there won't be enough people to form matches.

#95 Gormtrooper

    Rookie

  • The Angel
  • The Angel
  • 3 posts

Posted 17 June 2020 - 06:27 PM

I prefer the 2C from Jay Z setup, wins should count for more, but it should be possible for players to still gain if they performed well enough in a close match, or tried their best in a bad match.

#96 Cluster Fox

    Member

  • PipPipPipPipPip
  • 104 posts
  • LocationStuck on a rock in Grim Plexus

Posted 17 June 2020 - 07:06 PM

View PostJay Z, on 17 June 2020 - 04:38 PM, said:

Oh I know full well the 3 monthly PSR rescale is a bandaid, ...


No attack from me there JayZ, just wanted to highlight the point.
I do agree that while imperfect, if the redistribution is done well every 3 months, maybe flattening the bell curve along with it, it will help greatly with a viable system overall. I think this is the easiest solution I've heard for controlling the MS based PSR.

View PostGormtrooper, on 17 June 2020 - 06:27 PM, said:

... wins should count for more, but it should be possible for players to still gain if they performed well enough in a close match, or tried their best in a bad match.

Here I disagree. Nightbird research highlighted that WLR is very important to decrease stomps.
Especially at the matchmaker side, but if the MM is not touched, then we have to account for this in PSR.
Less stomps is good. This is not about punish / reward, it's about imperfect metrics to evaluate player's effect in a team. Over time! stop looking at a single match with punish/reward mindset...
In order to make the current MM work at least a bit like a WLR based matchmaker, you can't have any up movement on a loss or down movement on a win. This mitigates WLR and is against my principles for that reason.

Case 2A or 2B therefore make the most sense and Case 1A/B could work only with a Significant match score multiplier IMHO. If Wins/Losses blend in too much in PSR, stomps will increase.

Edited by Cluster Fox, 17 June 2020 - 08:20 PM.


#97 Raining Fire

    Member

  • PipPip
  • The Butcher
  • The Butcher
  • 28 posts
  • LocationCanada

Posted 17 June 2020 - 07:10 PM

I literally never post to the forums, but if I ever had one thing to say it would be this: The importance of win/loss should not be directly factored into the PSR calculation by going with 2a or 2b. W/L importance should be included in the effect it has on match score.

Ie. If you go with 1a and end up feeling like players from the losing team are rewarded too much, the only change should be an increase to the match score gained from winning, not a switch to the 2a system. Win/loss should not directly affect how the PSR is allocated. It should affect the match score of winning/losing players.

I vote for 1a. I believe the system should be true zero-sum. Given that I'm ruling out 2a and 2b for the reason above, 1a appears to be the most valid system. Simply implement the true zero-sum ranking by match score, then alter the match score kicker for winning until you feel as though the winning players are rewarded appropriately for winning (equivalently: the losing players are punished appropriately for losing).

Edited by Raining Fire, 17 June 2020 - 07:17 PM.


#98 D U N E

    Member

  • PipPipPipPipPip
  • Bad Company
  • Bad Company
  • 131 posts

Posted 17 June 2020 - 07:32 PM

I vote core 2B with Jays additions.

People who are looking at tiers like they are some goal or treasure - if someone consistently groups and achieves high MS and wins, regardless if you feel they should be there or not - by being in a higher tier they are more likely to be grouped with either: A the exceptionally good, or B, the horde of groups.
Which means, you will find less high skill groups stomping lower tier games.

Being the highest tier does not matter, I'd argue they should hide tier because it makes people feel insecure and act like it's something to be farmed. It's a match maker we are talking about, if someone consistently performs well in a group, that group should be raised into a higher tier so it has an increased chance of fighting against a team/group that isn't going to fold as easily.

Remember, this isn't the Guillotine award, this is a match maker, and the more you farm it to get higher, the more likely to fight against EmP, JGx or other elite teams that are enjoying farming in solo que, which as I have heard from some pugs, they don't like fighting against a 4 man of EmP 4 times in a row.

#99 MK7 GTI

    Rookie

  • The Terror
  • The Terror
  • 1 posts

Posted 17 June 2020 - 07:44 PM

My choice would be 1A.. if I had to chose between the 4

I play SOLO quick play. I am not a top 1% elite comp player. I never will be.

I try to win every game. With the soup queue, more often then not I end up the puppy being kicked by the 4 man groups. I don't mind it. The queues are faster now! I just think that if I get a great game.. I should be able to progress even though my team NSRs one at a time and lose in less then 3 minutes. And if I double d I should go down no matter what.

My idea would be the top 6 from winning team and the top 2 from the losing team go up. the bottom 8 match score go down winning team or not. the rest do not move.

Pretty easy there.

Thanks for reading.

#100 SuperMCDad

    Member

  • PipPipPipPipPip
  • Bridesmaid
  • Bridesmaid
  • 131 posts
  • Twitter: Link
  • LocationNew Zealand

Posted 17 June 2020 - 07:46 PM

My vote goes with JayZ's 2C (2B modified) if possible.





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users