Jump to content

Psr Tiers - Are They Flawed?

Balance General Metagame

135 replies to this topic

#81 Wintersdark

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • 13,375 posts
  • Google+: Link
  • Twitter: Link
  • LocationCalgary, AB

Posted 22 September 2015 - 03:03 PM

View PostXX Sulla XX, on 22 September 2015 - 08:43 AM, said:

Its fine the way its setup. If you notice the better players do more consistant damage do matter what mechs and weapons they are using. And they also do it in losses. Bad players will hit a wall pretty quickly as they run into better players no matter what weapons etc they are using.
True, but my concern isn't the ends of the scale but the middle.

This tournament was particularly enlightening to me, because I don't normally pay a heck of a lot of attention to match scores. What I found was:

Matches where I players very well and Was Credit To Team where consistently scored lower (though still reasonably high) than matches where I rolled along doing lots of damage but not really accomplishing anything.

I stuck with my Direwolf Siegebreaker through most of the tournament. With a roughly 100pt pinpoint alpha, with just a little accuracy it pushes through mechs quickly. But all my best scoring matches where odnes where I wrecked lots of mechs, but was sloppy.


Now, to be clear and avoid Mr. Blasterman issues, I do not think kills should be ranked higher, because the killing blow is effectively meaningless. Rating killing blows higher would just lead to different obnoxious play.

I wanted to give PGI a chance to see if their new scoring system - which claimed to be based off much more than just damage - actually panned out. It's not. It's pretty much just damage. I mean, +1, +2 for various things, when you get 300 points for damage done? Those other factors become irrelevant very fast.

Honestly, I don't think they can make a scoring system that'll actually work though. Before PSR, as I said, I always felt Elo was the best of a list of bad options for this very reason. It's too hard to actually quanify numerically good play.

#82 Roadkill

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,610 posts

Posted 22 September 2015 - 03:05 PM

View PostFate 6, on 22 September 2015 - 02:34 PM, said:

Yeah, the match score is like 70% of it.

More like 80%. I didn't check my bad games, but in all of my good games the non-damage component was ~20% of match score.

Under the old system damage was more like 65% of match score, so its reliance on damage has gotten worse.

#83 Koniks

    Member

  • PipPipPipPipPipPipPipPip
  • Moderate Giver
  • Moderate Giver
  • 1,301 posts

Posted 22 September 2015 - 03:19 PM

View PostUltimatum X, on 22 September 2015 - 07:07 AM, said:

What's your opinion of this?


My opinion is we should've stuck to Elo, kept the weight class distinction, and just put in the hard tier restrictions for matchmaker. An individual's wins and losses are the best prediction of whether someone's average performance makes them more or less likely to win, regardless of what role they use to do so.

With enough of a sample size, you've either figured out how to win or you haven't. The game modes don't even have enough variety to use them as a separate variable.

#84 Roadkill

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,610 posts

Posted 22 September 2015 - 03:19 PM

View PostWintersdark, on 22 September 2015 - 03:03 PM, said:

Before PSR, as I said, I always felt Elo was the best of a list of bad options for this very reason. It's too hard to actually quanify numerically good play.

I think they have PSR backwards. What they currently have is match score (read: damage) modified by W/L. It should be the other way around - W/L modified by match score.

The single most important piece of data you can take from an end result is "did you win?" After that you can look at how you did within your own team to determine your relative contribution to the win, but whether or not you won should be the biggest factor in PSR.

#85 YourSaviorLegion

    Member

  • PipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 418 posts
  • LocationSpace The Final Frontier

Posted 22 September 2015 - 03:37 PM

View PostFate 6, on 22 September 2015 - 01:44 PM, said:

No if you want to go up you just need to play better. If you are in Tier 3 with good builds it's because you are average. Another reason is that you're blaming the system - if you don't blame yourself for being average you will always stay average. The only way to improve your Tier is to play better not blame PGI.

Match Score doesn't really reflect wins/losses though which is significant in PSR. You could get **** match scores all day but if you always win your PSR won't go down

That's cute... I'll still beat you in a Zellbriggen Jade Turkey

#86 Wintersdark

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • 13,375 posts
  • Google+: Link
  • Twitter: Link
  • LocationCalgary, AB

Posted 22 September 2015 - 03:42 PM

I should add:

I'm not concerned that bad players are going to become T1, or great players are going to fall to T5. That's not really a danger.

My concern here is the middle ground. My concern is that sloppy players are going to be pushed up too fast, and good-average players held back - basically, that this effect will cause players to gravitate towards the center, and that'll lead to a strange, hard to understand "road up".

The acts that get you from T4 to T3 to T2 are the acts that will ensure you never see T1; and that strong, precise play in lighter mechs will ensure you're held in a lower tier.

Basically: I do my best when running dual LBX10 Jagermechs/LRM Stalkers (look at all these crazy high match scores I got!) and that got me to T2 fast, but now I lose all these matches because obviously my team sucks.

#87 Kaeb Odellas

    Member

  • PipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 2,934 posts
  • LocationKill the meat, save the metal

Posted 22 September 2015 - 04:09 PM

Ideas for improving Match Score:

Award Killing Blows a component destruction for each component still intact on the mech when it went down, and give an "efficiency" bonus the more components were intact when the mech went down. Whoever did the most damage should get maybe 50% of whatever the killing blow guy got, and of course Solo kills get both. Then reduce score for damage dealt by a bit.

Players should also get a match score award (Tank bonus?) for taking damage without losing components, though this would have a hard cap to prevent abuse. This would perhaps incentivize Assaults and Heavies to take some damage for the team.

UAV kills should get bonus points for however many friendly mechs that UAV was spotting.



Values could look something like this:
  • Killing Blow: 5 points, plus Component Destruction for every component still intact.
  • Component Destruction: 2 points
  • Efficiency Bonus: 2 points for every component left intact on target's death
  • Kill Most Damage Dealt: 3 points, plus half of the component destructions and efficiency bonus that whoever did the killing blow received.
  • Solo Kill: 8 points, plus half of the component destructions and efficiency bonus that you got from Killing Blow.
  • Assist: 4 points
  • Death: -10
  • Suicide: -15
  • Out of bounds Suicide: -50 (seriously, there's no excuse)
  • Damage: 30% of damage dealt
  • Team Damage: -50% of damage dealt.
  • Win by Kills: 30
  • Assault base capture or Resource win: 20 + 5 per enemy mech still standing
  • Loss: 0
  • Tie: 10
  • Capture control point: 5
  • Neutralize enemy control point: 4
  • Capture per tick: 1
  • Tank bonus: 0.05 score points per point of damage taken until death or end of match. Max 20 score points.
  • UAV kill: 2 + 0.5 point per friendly mech visible to that UAV.
  • Everything else: 1 point or whatever its current value is.


#88 BLOOD WOLF

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The Jaws
  • The Jaws
  • 6,368 posts
  • Locationnowhere

Posted 22 September 2015 - 04:11 PM

I thought the new PSR was weighting by a number of factors?

#89 J0anna

    Member

  • PipPipPipPipPipPipPip
  • The Merciless
  • The Merciless
  • 939 posts

Posted 22 September 2015 - 04:29 PM

How did he come up with those numbers? If true it is amazingly easy to sandbag your tier....

#90 riverslq

    Member

  • PipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 443 posts

Posted 22 September 2015 - 04:30 PM

If I play a ****** light mech, but one I enjoy, my tier tanks.
good job balancing pgi.
if I play an op meta mech, my tier shoots up.
good job balancing pgi.

guess I just tank while having fun and wipe noobs in tier 5 then. :)
tier 3 right now. i'm surprised though, i usually play ****** light mechs i enjoy.

Edited by riverslq, 22 September 2015 - 04:32 PM.


#91 Mcgral18

    Member

  • PipPipPipPipPipPipPipPipPipPipPipPip
  • CS 2019 Top 8 Qualifier
  • CS 2019 Top 8 Qualifier
  • 17,987 posts
  • LocationSnow

Posted 22 September 2015 - 04:42 PM

View PostMechWarrior5152251, on 22 September 2015 - 01:46 PM, said:

Deaths were glaring. Some people had 1000 damage but no deaths and had match scores over 600. So that table is way way off.


Over 1000 match score, actually.

Those follow the Match Score formula perfectly.
If you're talking CW, it divides the score for each mech you use.

#92 crashlogic

    Member

  • PipPipPipPipPipPip
  • Mercenary
  • 318 posts

Posted 22 September 2015 - 05:01 PM

View PostMister Blastman, on 22 September 2015 - 08:51 AM, said:

That's doing your job as a light 'mech. :) I wasn't referring to that.

What I was talking about is folks who stay away from their team, refuse to take damage and plink from far, far away, waiting for the hole to open up and finish the job. By staying out of the action they don't help spread the enemy damage making it harder for the rest of their team to win.


I do this in an lrm. If I am forcing guys to take cover at 1100 meters, then the rest of the team benefits. I can do my share of brawling in a maddog, and I do, but being able to hit at long range is essential to good teamwork. I like the saying "Damage belongs to the individual, but kills belong to the team."
Which is the basic flaw of PSR..out success in any match depends on teamwork, yet we are put in teams by individual ratings. They ought to count keeping in formation, protection and providng ecm cover as much as they count assists damage or kills.

#93 Splatshot

    Member

  • PipPipPipPipPip
  • Stone Cold
  • Stone Cold
  • 179 posts

Posted 22 September 2015 - 05:37 PM

Honest truth, this only really matters to the top 10%.

Let them all opt in and play each other, and let the rest play and have fun.


I also wonder if map choice has an effect as I play is Conquest unless its an event.

Edited by Splatshot, 22 September 2015 - 05:45 PM.


#94 Wintersdark

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • 13,375 posts
  • Google+: Link
  • Twitter: Link
  • LocationCalgary, AB

Posted 22 September 2015 - 06:13 PM

View PostRoadkill, on 22 September 2015 - 03:19 PM, said:

I think they have PSR backwards. What they currently have is match score (read: damage) modified by W/L. It should be the other way around - W/L modified by match score.

The single most important piece of data you can take from an end result is "did you win?" After that you can look at how you did within your own team to determine your relative contribution to the win, but whether or not you won should be the biggest factor in PSR.

That's exactly why I prefered Elo - not that I claim Elo was great; but it was the only system that actually included every intangible.

Their approach, however, was deliberately opposite: They wanted to unlink win/loss from your ranking, because often you'd lose matches despite playing well, or win matches despite playing poorly. That's true, but I personally believe that in aggregate, it'll balance out to reflect your contributions primarily as you are the common denominator in every match result you get.

I understand their goal, but I feel it's very dangerous. It promotes poor play from lower ranked players (rank up by playing worse!) and allows skilled players to deliberately sandbag their score - in order to face poor players they can continue to dominate while dropping in score.

#95 Quaamik

    Member

  • PipPipPipPipPipPip
  • Philanthropist
  • 413 posts

Posted 22 September 2015 - 06:29 PM

There are some simple scoring solutions that would improve it. Based off of the idea that half of damage adds directly to your score, these seem like they would balance it nicely:

(numbers are effects to score)
Solo Kill: +30
Most damage dealt to a kill: +30 (note this actually makes a solo kill worth +60)
Savior kill: +20
Assist: +10
Kill: +5 (assumes you would get the assist reward as well, does not add to solo kill)

Suicide: -30
Team Kill: -50
Team damage: - half damage dealt to teammates

That would give the proverbial "kill stealer" who did 200 damage and killed 4 mechs that were already cored a match score of: 160

If that same player was actually good enough to make their shots really count, doing headshots, or coring mechs from behind, or catching lighter mechs where 50 points of damage could kill the CT, they would get a score of 280 - 380, depending on how many solo kills they achieved.

That type of system rewards effective damage more than just general damage.

#96 Wintersdark

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • 13,375 posts
  • Google+: Link
  • Twitter: Link
  • LocationCalgary, AB

Posted 22 September 2015 - 07:29 PM

I would like to see players get points based on how low their health was when destroyed: the lower before the killing blow, the more damage you've tanked. This is a good thing - even if you're just being crushed, you're doing your team a favour by making it take as long as possible, and it encourages strong play practices.

#97 ThisMachineKillsFascists

    Member

  • PipPipPipPipPipPipPip
  • 871 posts

Posted 22 September 2015 - 07:49 PM

View PostFupDup, on 22 September 2015 - 07:45 AM, said:

Posted Image

Hey! i want the tier paul be represented in that pic . it would be somewhere here^^^^

Edited by ThisMachineKillsFascists, 22 September 2015 - 07:52 PM.


#98 Deathlike

    Member

  • PipPipPipPipPipPipPipPipPipPipPipPipPipPip
  • Littlest Helper
  • Littlest Helper
  • 29,240 posts
  • Location#NOToTaterBalance #BadBalanceOverlordIsBad

Posted 22 September 2015 - 08:38 PM

View PostThisMachineKillsFascists, on 22 September 2015 - 07:49 PM, said:

Hey! i want the tier paul be represented in that pic . it would be somewhere here^^^^


What if I told you that he was at the base of the mountain?

Edited by Deathlike, 22 September 2015 - 08:39 PM.


#99 ThisMachineKillsFascists

    Member

  • PipPipPipPipPipPipPip
  • 871 posts

Posted 22 September 2015 - 08:43 PM

View PostDeathlike, on 22 September 2015 - 08:38 PM, said:


What if I told you that he was at the base of the mountain?

o.O

Edited by ThisMachineKillsFascists, 22 September 2015 - 08:51 PM.


#100 Mechwarrior Buddah

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 13,459 posts
  • LocationUSA

Posted 22 September 2015 - 08:44 PM

Hell yes its flawed, once you hit tier 3, youre matched with every other tier. Whats the point of a tier system then?





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users