Jump to content

Psr Update And Hold On Patch.


716 replies to this topic

#701 Sjorpha

    Member

  • PipPipPipPipPipPipPipPipPip
  • Philanthropist
  • Philanthropist
  • 4,235 posts
  • LocationSweden

Posted 17 June 2020 - 05:15 AM

View PostXiphias, on 15 June 2020 - 12:53 PM, said:

Posted Image
You can see that the W/L system arrives at a good steady state with clearly defined tiers. Including the MS in the calculation causes the system to diverge and ends up putting things back to where they are right now. How long that takes will depend on the values used and the playerbase, but that's the eventual result.


This is great.

PGI please pay attention to this above simulation and do simulations for whatever models you're considering.

We want a distribution along the lines of the blue one, we don't want a distribution like the red one.

#702 Sergeant Destroy

    Member

  • PipPipPipPipPip
  • Bad Company
  • 156 posts

Posted 17 June 2020 - 08:51 AM

So is this **** gonna happen anytime soon or what?

#703 spannerturner

    Member

  • PipPip
  • The Spear
  • The Spear
  • 31 posts

Posted 17 June 2020 - 10:19 AM

New thread on the topic was started:

https://mwomercs.com...edback-round-1/

#704 Munkeyed

    Member

  • Pip
  • Philanthropist
  • 17 posts
  • LocationUS

Posted 17 June 2020 - 09:48 PM

Would make it more directly competitive... HUZZAH!
should still have a slight penalty/bonus for win/loss, just too keep it a "team sport"

#705 Exhall

    Rookie

  • Philanthropist
  • Philanthropist
  • 4 posts

Posted 18 June 2020 - 05:24 AM

PGI essentially said "We can't fix this, you do it". And when it all falls down over semantics they can say "Well, we tried". With player numbers so small everyone will be playing everyone else anyway so it won't make any difference.

#706 Barely Good

    Rookie

  • Bad Company
  • Bad Company
  • 6 posts

Posted 18 June 2020 - 01:42 PM

Not sure if this is the place to suggest this but here goes.

A while ago I had a email conversation with someone there about Pilot score after a match.

For example, you could run around kicking butt, kill 10 on other team, loose the game due to score or whatever, and your Pilot Rating will be = or go DOWN Solely because your team lost the match!

I mean really? If you kicked butt that match, team lost, you shouldn't be penalized just because your team lost the match!


Another suggestion I had was the loading screens when exiting a game. IMO, a great deal of time in this game is devoted to skills on a new mech.
So when the game is over, why can't the game put you back on the last screen you were in before the match launched?

Instead, the match ends, you are back on Home, then have to click to Mechlab, then have to click to skills.
Why not just put us back in the screen we were in when we clicked Quick Play?

#707 Xiphias

    Member

  • PipPipPipPipPipPipPip
  • Littlest Helper
  • Littlest Helper
  • 845 posts

Posted 18 June 2020 - 04:46 PM

View PostBarely Good, on 18 June 2020 - 01:42 PM, said:

Not sure if this is the place to suggest this but here goes.

A while ago I had a email conversation with someone there about Pilot score after a match.

For example, you could run around kicking butt, kill 10 on other team, loose the game due to score or whatever, and your Pilot Rating will be = or go DOWN Solely because your team lost the match!

I mean really? If you kicked butt that match, team lost, you shouldn't be penalized just because your team lost the match!

That's perfectly fine though. I know it doesn't seem fair to go down on a great match, but as long as the movement on average is in the correct direction the individual matches don't matter. A player that is consistently putting up good results (10 kills) will win more often than not so they will go up. A player that isn't contributing might go up (undeserved) some matches, but will end up losing more often than they win.

Individual match results don't matter, only the average of the match results. If you balance the MM around individual matches by letting players on the losing team move up (and winners move down) you will cause the system to diverge over time and split most players into either T1 or T5. That's not a good system in the long run.

#708 Akillius II

    Rookie

  • 6 posts

Posted 18 June 2020 - 06:01 PM

If PSR change works then Devs gonna know it from all the haters yelling about being up against equals. And number of steam rolls will go down while number of close matches should go up. Therefore their precious win/loss ratios will be (correctly) a thing of the past as the number of matches won/lost vs equals should over time become 50/50 equal.
But from many of the comments I can see a lot just don't get it, and that's caused by way too many Years of a bad faux "PSR".

I vote Core 1A as it does not favor a reverse flow to tier 5, and it does not favor the old ways.
+ Match scores already account for win/loss.
+ One "bad" or good match doesn't equal a whole hill of beans after a few dozen matches if PSR is formulated correctly.
If there was say 500 points per tier then it'll take several matches to change tiers, so a few bad/good matches won't change it.
+ Core 1B and 2B are not zero sum and both act like the current PSR just in reverse down to tier 5.
+ A real PSR would not require monthly/quarterly resets/adjustments/etc which (as others pointed out) 1B/2B/2C would.
+ 1A favors the floating middle and players float or sink depending on skill, as it should be.

Caveats:
- Mechs that move 0 distance for entire match get No match score.
- Mechs that overheat and explode within first 2 minutes of match starting get Negative match score.
- Mechs that go out of bounds and die within first 2 minutes of match starting get Negative match score.
- Players that severely damage or kill "friendlies" within first 45 seconds of match starting get Negative match score.
To avoid exploiting PSR to drop tiers, these players PSR for that match should be flagged to stay Unchanged.

#709 Horseman

    Member

  • PipPipPipPipPipPipPipPipPip
  • CS 2019 Top 8 Qualifier
  • CS 2019 Top 8 Qualifier
  • 3,086 posts
  • LocationPoland

Posted 18 June 2020 - 10:04 PM

View PostAkillius II, on 18 June 2020 - 06:01 PM, said:

+ Core 1B and 2B are not zero sum and both act like the current PSR just in reverse down to tier 5.
They very much are zero sum if the PSR adjustment is calculated correctly. Remember, a zero sum system just has to ensure that the total amount of PSR increase is equal to the amount of PSR decrease (ie, that the sum of all PSR changes after the match equals zero). How that amount is distributed overall is not relevant.

#710 Akillius II

    Rookie

  • 6 posts

Posted 19 June 2020 - 05:07 PM

Cognitive biases aside, lets avoid systems that require never ending "adjustments" therefore the amount distributed does matter.
Otherwise PGI will have a "MMR Hell" ranking system like league of legends & some other games use.

Also PSR must not be linked to a history of a players previously played matches like some kind of persistent leaderboard-like ranking system to avoid smurfing among other issues.

However to encourage team play PGI just needs to simply increase the matchscore rewards for:
Lance Formation, Protected Light, Protected Medium, Protect Proximity

Ideally there would be a Solo-PSR, and a Team-PSR, and a CW-PSR, but that won't happen.
Because PGI is not looking to reinvent match maker, PSR and match scores all at the same time.
So PGI is unlikely to mess with matchscores beyond the above as that is an established known working system.
Moreover if PGI changes too much in matchscores or match maker then interpretation of changes to PSR will be flawed.

Fear not because in the end it doesn't matter what I write here, as I'm not on twitter, and I'm not one of Russ's pets, plus I'm not a whale and only purchased half the mechs via preorder over the years on my main account so I'm definitely persona non grata.

Edited by Akillius II, 19 June 2020 - 05:07 PM.


#711 Ishaldre Conner

    Rookie

  • The Bowman
  • The Bowman
  • 2 posts

Posted 19 June 2020 - 05:54 PM

I haven't played in years due to the wildly unbalanced weapons and mechs, a complete lack of a meta-game with any true BattleTech canon or feel to it, the lack of constructive/meaningful/important missions and gameplay, and the static maps.

My two-cents on what needs to happen?

1. Procedurally generated maps, so no two battlefields are the same (unless the battle is at a particular city/landmark/etc.)
2. Fix PUGs, S7 and Faction Play with a Canon timeframe selector, which will allow people to play period-specific battles across the IS from the Star League to the Word of Blake (or beyond). This gives everyone a chance to play to their likes and maybe try another BT era they are not familiar with. This also allows the development team to produce more Mechs from all eras and prevents any Mech produced from being totally nerfed as almost all Mechs had their glory day at some point in the BT canon.
3. Make the Clans operate like the Clans. Batchall must be in force during Faction Play (and possibly PUGs, but I think not having it in PUGs would be better). One of the big draws to the Clans is the Batchall and to have more accurate Faction play, 12 IS Mechs vs a smaller Clan force not only helps balance the superior weaponry of the Clan, it gives the IS Mechs a more sporting chance. And to make the Batchall more attractive, the lower a Clan team's bid, the more swag they will walk away with if they win (C-Bills, etc).
4. Make Faction play missions have more significance. Scouting missions, hit and run supply lines, frontal assaults, sneak attacks, "hold the line" and other mission types that all feed in to each other based on success or failure. Points, rankings, and badges are great for S7, but ultimately useless when it comes to Faction play if you want it to be a true canon-based game mode. Also, every fight being an even number of Mechs of the same tonnage is lame. War is never lined-up so neatly. Players want to see a lance of scout Mechs do a hit and run on two lances of lights and mediums defending an ammo dump or to see a star of Assault Mechs fighting to the last against waves of attacking Mechs.

I know I am asking a lot, but I played for many years here and put a lot of money in to this game in the hopes that the development would come around to a model of gameplay that was more engaging, jived with BT canon, and made more sense. I know that this post will likely fall by the wayside and MWO will fade away, no longer profitable or enjoyable.

And that makes me very sad...

#712 Surn

    Member

  • PipPipPipPipPipPipPip
  • The God of Death
  • The God of Death
  • 987 posts
  • Twitter: Link
  • LocationSan Diego

Posted 23 June 2020 - 08:23 AM

Sorry, as the a 99% player that has actual experience in this type of system I must apologize for not keeping up with the restrictions as I am working on a project at the moment.
Although, I was not aware of the restriction in finding average player PSR. My old league RPI could easily handle that, so I assumed 24 years later that would not be a problem.
I had also relied upon a rpi adjustment system to simplify match tracking and player RPI and team RPI. Think of it as college basketball RPI rankings, with player rankings included. You know, March Madness...how the brackets are calculated... maybe that is unfamiliar to some of my fellow mechwarriors.
my proposed, forumla was ( from post):
Win/Loss variable:
W = +1 for winning pilots and -1 for losing pilots
Match variables:
P = Individual Pilot Matchscore
TA = Team Average Matchscore
WA = Winning Team Average Matchscore
LA = Losing Team Average Matchscore
A = Average Matchscore of match
Historical variables:
PPSR = Individual Pilot Current PSR
APSR = Average PSR of Pilots in the game.
PSR = (P * (APSR/PPSR)/TA) + W * ( 1 + (WA/LA))
This makes it harder to gain matchscore as you have a higher PSR relative to the match participants, and the opposite is true.
However, it is not Zero sum, but that can be alleviated with a lookup table.

My new formula :
PSR = P/A + W * ( 1 + WA/LA)
How the pilot did in the game + win/loss factor that is consistent across winners and losers
Rank the players in order of psr, with top psr worth 24 points and bottom worth -24.




Also, how did the "Community PSR Fixes / Proposals" seem to happen the same day as the announcement? I literally have not had time to put into this yet.

Edited by Surn, 23 June 2020 - 08:26 AM.


#713 Dakkalistic

    Member

  • PipPipPipPipPipPipPip
  • 661 posts

Posted 23 June 2020 - 09:01 AM

I see Mr. 99% is above figuring out the right thread to post in, great start dude...

#714 Munkeyed

    Member

  • Pip
  • Philanthropist
  • 17 posts
  • LocationUS

Posted 23 June 2020 - 06:08 PM

Why not rank each team 1st through 12th with a modifier for W/L?

#715 OneTeamPlayer

    Member

  • PipPipPipPipPip
  • Bad Company
  • 113 posts

Posted 23 June 2020 - 06:11 PM

This thread is multiple threads behind the current point of this topic.

It has been defunct for weeks at this point.

Stop Replying, look for the most current thread.

#716 Tongo Rad

    Rookie

  • Bad Company
  • 6 posts

Posted 25 June 2020 - 11:47 AM

View PostOneTeamPlayer, on 23 June 2020 - 06:11 PM, said:


This thread is multiple threads behind the current point of this topic.

It has been defunct for weeks at this point.

Stop Replying, look for the most current thread.



Where IS the most recent thread?

#717 Horseman

    Member

  • PipPipPipPipPipPipPipPipPip
  • CS 2019 Top 8 Qualifier
  • CS 2019 Top 8 Qualifier
  • 3,086 posts
  • LocationPoland

Posted 25 June 2020 - 01:53 PM

View PostTongo Rad, on 25 June 2020 - 11:47 AM, said:



Where IS the most recent thread?

https://mwomercs.com...ity-version-10/





2 user(s) are reading this topic

0 members, 2 guests, 0 anonymous users