Jump to content

Psr Tier Metrics - Round 1


544 replies to this topic

#21 Bud Crue

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Rage
  • Rage
  • 9,242 posts
  • LocationOn the farm in central Minnesota

Posted 15 July 2020 - 05:11 PM

View PostOneTeamPlayer, on 15 July 2020 - 04:03 PM, said:

How does matchmaker place mixed tier groups?

Your example of three T1s and their cadet buddy is one issue but what about:
  • A T1, T2, T3, T4
  • 2 T2, T3, T4
  • T1, 2 T3, T5
and all other manner of mismatched tier ratings? Will the tier system "average"? Will it use the highest tier rating available? Some other metric?



Right...we are dropping this evening as a 4-man of: T1,T2, T2, T4. Most matches have had the 4 min wait time then we get a drop. How does that work and why?

#22 Dakkalistic

    Member

  • PipPipPipPipPipPipPip
  • 670 posts

Posted 15 July 2020 - 05:14 PM

Why do ppl always make it about stomps? Stomps are lore:

https://www.sarna.ne..._Loss_Groupings

Final match score says very little how good the match actually was imho.

#23 Paladin IIC

    Member

  • PipPipPip
  • The Carnivore
  • The Carnivore
  • 94 posts

Posted 15 July 2020 - 05:18 PM

RIP MWO.

#24 Nightbird

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The God of Death
  • The God of Death
  • 5,301 posts

Posted 15 July 2020 - 05:21 PM

View PostDakkalistic, on 15 July 2020 - 05:14 PM, said:

Why do ppl always make it about stomps? Stomps are lore:

https://www.sarna.ne..._Loss_Groupings

Final match score says very little how good the match actually was imho.


PGI made it about stomps, this entire PSR update is to reduce stomps

https://mwomercs.com...anges-jun-2020/

#25 Dakkalistic

    Member

  • PipPipPipPipPipPipPip
  • 670 posts

Posted 15 July 2020 - 05:25 PM

View PostNightbird, on 15 July 2020 - 05:21 PM, said:

PGI made it about stomps, this entire PSR update is to reduce stomps

https://mwomercs.com...anges-jun-2020/

Yeah ofc they're going to adress it, but tying match quality to stomp or not is a crass oversimplification, and has nothing to do with psr distribution right now.

#26 Nightbird

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The God of Death
  • The God of Death
  • 5,301 posts

Posted 15 July 2020 - 05:27 PM

View PostDakkalistic, on 15 July 2020 - 05:25 PM, said:

Yeah ofc they're going to adress it, but tying match quality to stomp or not is a crass oversimplification, and has nothing to do with psr distribution right now.


https://mwomercs.com...cy-with-graphs/

Proven here with math, stomp and match quality go hand in hand

#27 Dakkalistic

    Member

  • PipPipPipPipPipPipPip
  • 670 posts

Posted 15 July 2020 - 05:34 PM

View PostPaul Inouye, on 15 July 2020 - 02:11 PM, said:


As of this post, I will be making the maximum Tier separation in the match maker, 1 (one). This means a Tier 1 player should not see any other players outside of Tier 1 or Tier 2.



I'm not sure the T1 population is large enough yet for that restriction.

View PostNightbird, on 15 July 2020 - 05:27 PM, said:

https://mwomercs.com...cy-with-graphs/

Proven here with math, stomp and match quality go hand in hand


Yeah, I know you're the guy that is salty because PGI didn't embrace your brainchild, you made your point.

#28 Nightbird

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The God of Death
  • The God of Death
  • 5,301 posts

Posted 15 July 2020 - 05:43 PM

View PostDakkalistic, on 15 July 2020 - 05:34 PM, said:

Yeah, I know you're the guy that is salty because PGI didn't embrace your brainchild, you made your point.


Can I take that as you being happy at the current state of the game?

#29 OneTeamPlayer

    Member

  • PipPipPipPipPipPip
  • Bad Company
  • 377 posts

Posted 15 July 2020 - 05:49 PM

View PostDakkalistic, on 15 July 2020 - 05:14 PM, said:

Why do ppl always make it about stomps? Stomps are lore:

https://www.sarna.ne..._Loss_Groupings

Final match score says very little how good the match actually was imho.


Stock mech builds are lore too with customizations being ridiculously expensive and rare.

Repair costs are also lore.

Re-arming costs are also lore.

Permanent destruction of mechs is also lore

Pilot knockouts with long recovery times are also lore

If every bit of "lore" were present in MWO literally no one would play this game.

I also can't believe that someone is arguing for unbalanced results in what is advertised as a competitive game.

2020 is some kind of year "yeah we're playing a game bro, but getting your teeth stomped in because of a lack of an ability to create fair matches like literally every other competitive game does is lore, bro."

#30 Clay Endfield

    Member

  • Pip
  • Bridesmaid
  • 16 posts

Posted 15 July 2020 - 06:13 PM

Does PGI have any plans to adjust matchscore calculation? Cause right now, I'm feeling kinda penalized for playing a pinpoint heavy versus the guy boating LRMs on a 100 ton Assault mech; they're gonna get higher matchscore for sitting in the back and leaching target data, when their assault mech is supposed to be applying pressure on the enemy frontline and sharing armor team closer to the front. I'm the one risking my neck, using a config designed to kill rather than soften targets, and I'm being awarded 1/3rd of the mooch's matchscore who is exerting little to no skill and detracting from the team by using 100 tons of mech to do a 65 ton mech's job. Because it only takes me 2-5 alphas to kill (AC 20; 2x Snub PPC combo), I'm doomed to come out behind the mooch anyways, because it takes less damage for me to kill a target than it does the LRM mooch.

Can you guys track LRM damage per player in a match, and award 1/2 the score for damage dealt for LRMs to the matchscore total? It might cut back on all the LURM farmers spoiling every event, if LRMs are worse at amassing matchscore than just your run of the mill brawler/laser vomit build.

#31 MrMadguy

    Member

  • PipPipPipPipPipPipPipPip
  • 1,896 posts

Posted 15 July 2020 - 06:43 PM

I'm still in Tier 3, but slowly going to Tier 4. Quality of game is much better now. Number of deathball stomps are much lower.

View PostClay Endfield, on 15 July 2020 - 06:13 PM, said:

Does PGI have any plans to adjust matchscore calculation? Cause right now, I'm feeling kinda penalized for playing a pinpoint heavy versus the guy boating LRMs on a 100 ton Assault mech; they're gonna get higher matchscore for sitting in the back and leaching target data, when their assault mech is supposed to be applying pressure on the enemy frontline and sharing armor team closer to the front. I'm the one risking my neck, using a config designed to kill rather than soften targets, and I'm being awarded 1/3rd of the mooch's matchscore who is exerting little to no skill and detracting from the team by using 100 tons of mech to do a 65 ton mech's job. Because it only takes me 2-5 alphas to kill (AC 20; 2x Snub PPC combo), I'm doomed to come out behind the mooch anyways, because it takes less damage for me to kill a target than it does the LRM mooch.

Can you guys track LRM damage per player in a match, and award 1/2 the score for damage dealt for LRMs to the matchscore total? It might cut back on all the LURM farmers spoiling every event, if LRMs are worse at amassing matchscore than just your run of the mill brawler/laser vomit build.

Current Meta - is speed. 100 ton Assault - is just one big stow target dummy. Any Medium has enough firepower to two-shot your ST now. LRM builds - are only viable now. I personally prefer mixed LRM + TAG + some support weapons as the most deadly variant. You do instant LRM rain on every enemy, that tries to get out of cover, but you also have something like Gauss to kick them, if they actually get out and try to get closer to you. In this case Assault is the best in this role, as it has large amount of tonnage to fit both kinds of weapons at the same time.

Edited by MrMadguy, 15 July 2020 - 06:48 PM.


#32 Snukums

    Member

  • PipPip
  • Ace Of Spades
  • Ace Of Spades
  • 22 posts

Posted 15 July 2020 - 06:51 PM

Just tried some drops and wow the wait is long if you're T1. Don't think there are enough players in these tiers as yet to restrict to just 1 Tier gap.

Hope that the Tiers rationalize soon.

#33 Dakkalistic

    Member

  • PipPipPipPipPipPipPip
  • 670 posts

Posted 15 July 2020 - 06:54 PM

View PostOneTeamPlayer, on 15 July 2020 - 05:49 PM, said:

Stock mech builds are lore too with customizations being ridiculously expensive and rare.

Repair costs are also lore.

Re-arming costs are also lore.

Permanent destruction of mechs is also lore

Pilot knockouts with long recovery times are also lore

If every bit of "lore" were present in MWO literally no one would play this game.

I also can't believe that someone is arguing for unbalanced results in what is advertised as a competitive game.

2020 is some kind of year "yeah we're playing a game bro, but getting your teeth stomped in because of a lack of an ability to create fair matches like literally every other competitive game does is lore, bro."


This is not about adhering to some old TT rules, this is just an observation on how mech combat works. In a perfect scenario, where every team had perfect armor sharing and cycling, ie equally shared the damage in the battle of attrition, every game would be a stomp. So it's really a question of how the stomp happened to determine whether it was gg or not.

#34 Nightbird

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The God of Death
  • The God of Death
  • 5,301 posts

Posted 15 July 2020 - 07:07 PM

The worse the PSR scale, the worse the MM, the more stomps we see. This is common sense. The best MM will still have some stomps, but noticeably fewer.

#35 Voice of Kerensky

    Member

  • PipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 58 posts

Posted 15 July 2020 - 07:37 PM

Interesting information. Thanks for the work you've done.
I do not know when exactly the rule for different levels + 1 / -1 was introduced, but yesterday I (level 3) met a cadet two or three times in battle. Once on my team, once on the opposing team.
I also wonder if the matchmaker will be trained to select opposing teams taking into account the tonnage of the mechs?
I can post a lot of screenshots when one of the teams had a significant advantage in assault mechs.
The brightest battle was, on the Grim Plexus, in skirmish or domination mode, when my team had 1 assault mech (Warhammer IIC), and the enemy team had 9 or 10. The battle ended very quickly, its result, I think, is obvious to everyone.

#36 Xaat Xuun

    Member

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

Posted 15 July 2020 - 07:45 PM

View PostClay Endfield, on 15 July 2020 - 06:13 PM, said:

Does PGI have any plans to adjust matchscore calculation? Cause right now, I'm feeling kinda penalized for playing a pinpoint heavy versus the guy boating LRMs on a 100 ton Assault mech; they're gonna get higher matchscore for sitting in the back and leaching target data, when their assault mech is supposed to be applying pressure on the enemy frontline and sharing armor team closer to the front. I'm the one risking my neck, using a config designed to kill rather than soften targets, and I'm being awarded 1/3rd of the mooch's matchscore who is exerting little to no skill and detracting from the team by using 100 tons of mech to do a 65 ton mech's job. Because it only takes me 2-5 alphas to kill (AC 20; 2x Snub PPC combo), I'm doomed to come out behind the mooch anyways, because it takes less damage for me to kill a target than it does the LRM mooch.

Can you guys track LRM damage per player in a match, and award 1/2 the score for damage dealt for LRMs to the matchscore total? It might cut back on all the LURM farmers spoiling every event, if LRMs are worse at amassing matchscore than just your run of the mill brawler/laser vomit build.

Ya know you could just move back by the LRM pilots, and have the front line move there, you don't have to always run to the center right away

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--=-=-=-=-=-

is there a way to see your QP PSR numbers ?
maybe I'm blind and just can't find it, if there is

Edited by Xaat Xuun, 15 July 2020 - 07:45 PM.


#37 Magic Pain Glove

    Member

  • PipPipPip
  • The Bushido
  • The Bushido
  • 86 posts

Posted 15 July 2020 - 07:47 PM

Wait times in T1 are infinite now almost . There is just not enough population in T1/T2 to do this move right now ...

Also , the chart is incorrect (slightly delayed) I think ? I got to T1 10 Hours after the PSR update patch and its not showing on the chart ? So there might be like 1 day of delay on tiers . It should have displayed 1 Tier 1 on the 2nd day of the chart . ( Overall irrelevant , but yeah there might be a day of delay on the chart )

#38 OneTeamPlayer

    Member

  • PipPipPipPipPipPip
  • Bad Company
  • 377 posts

Posted 15 July 2020 - 07:53 PM

View PostMagic Pain Glove, on 15 July 2020 - 07:47 PM, said:

Wait times in T1 are infinite now almost . There is just not enough population in T1/T2 to do this move right now ...

Also , the chart is incorrect (slightly delayed) I think ? I got to T1 10 Hours after the PSR update patch and its not showing on the chart ? So there might be like 1 day of delay on tiers . It should have displayed 1 Tier 1 on the 2nd day of the chart . ( Overall irrelevant , but yeah there might be a day of delay on the chart )


They knew this, they made the change simply to triage the uber groups away from T3 players for a while.

T1 can't make matches during primetime, imagine daytime play.

After two weeks this is what this change that took over a month to wade through and implement has to offer, putting us at months of the worst gameplay in MWO history.

I thought the exodus was going to be a sight to behold before, but now i'm truly interested in seeing next months numbers.

The game is in maintenance mode, the only job was to keep it running smoothly and it's like we're in one of those choose your own adventure books, the kind where you always choose the worst option first just to see all the endings.

Added solo to group queue -> pushed a mm change for months then cancel -> make players code your psr change -> make players vote on a mathematical system despite lack of mathematical aptitude -> take system, without telling anyone put all players into T3 -> note that there is barely any change in tiers after two weeks, cut off the ability to play for players whom obviously have put in the most time -> whatever terrible decision comes next -> Your have multiplay game has died of negligence and near malicious incompetence! Return to page 1 and try again!

Edited by OneTeamPlayer, 15 July 2020 - 07:54 PM.


#39 crazytimes

    Member

  • PipPipPipPipPipPip
  • Ace Of Spades
  • 433 posts

Posted 15 July 2020 - 08:00 PM

I had no idea there was so few in T1/2.

I like the idea of a 2 tier spread as 3 tier is clearly flawed from a skill perspective... but as I said when everyone was getting excited about how PSR is calculated- without dropping the tier spread it is useless. Drop the tier spread and people can't get matches. There aren't enough people to drop the tier spread.

There is no way that I am waiting +4 minutes for a match consistently. I enjoy the game, not looking at a waiting screen. Plenty of games in my steam library if the wait times stay high.

Edited by crazytimes, 15 July 2020 - 08:01 PM.


#40 Aidan Crenshaw

    Member

  • PipPipPipPipPipPipPipPip
  • The Mercenary
  • The Mercenary
  • 1,757 posts

Posted 15 July 2020 - 08:09 PM

View PostCluster Fox, on 15 July 2020 - 04:40 PM, said:

What information did you use? Own end of game screens?


Imagination, obviously. I don't take this seriously.

View PostVoice of Kerensky, on 15 July 2020 - 07:37 PM, said:

Interesting information. Thanks for the work you've done.
I do not know when exactly the rule for different levels + 1 / -1 was introduced, but yesterday I (level 3) met a cadet two or three times in battle. Once on my team, once on the opposing team.

That's no surprise. Cadets start in Tier 3 right now.

Edited by Aidan Crenshaw, 15 July 2020 - 08:17 PM.






5 user(s) are reading this topic

1 members, 4 guests, 0 anonymous users