Jump to content

Balance Elo With Kda


22 replies to this topic

#1 MadCat02

    Member

  • PipPipPipPipPipPipPip
  • The 1 Percent
  • 668 posts

Posted 30 September 2013 - 03:01 PM

As 1/12 of your team you can only effect your win rate so much .

For those of you who don't know MWO ELO is different for every weight class . Thats why i wait 20 minutes to play as assault and 30 seconds as a light mech .

Lets say i have 2.0 win rate with 4.0 KDA on assaults . Its about 66% win rate so if you match me agaist 60% win rate assaults with 1.8 kda its not very fair .

Unfornutly WIN/LOSS is mostly determined by a match making luck with some rare close games .

This is not Quacke where 1 good player can kill everyone . If i am matched agaist a group of skilled players it really dosen't matter how much i try . Nobody likes that .

......................................

I undestand that a light mech can't have as many kills as heavys or assault .

However we already have elo based on a weight class so kda can be specific to a weight class as well .

These avareg scores are based on my stats so i apologize if i may have overated this community XD.

My gauess is 1.8 KDA is above avareg on a light mech pilot .

2.0 For mediums

2.5-4.0 For heavys and assaults .

I think balance between elo and KDA for every class would result in better gameplay.

Edited by MadCat02, 01 October 2013 - 06:04 AM.


#2 PEEFsmash

    Member

  • PipPipPipPipPipPipPipPip
  • Philanthropist
  • Philanthropist
  • 1,280 posts
  • LocationLos Angeles

Posted 30 September 2013 - 03:31 PM

View PostMadCat02, on 30 September 2013 - 03:01 PM, said:

My gauess is 1.8 KDA is above avareg on a light mech pilot .

2.0 For mediums

2.5-4.0 For heavys and assaults .


Lol you must be really bad at math. All of your statistics are way above average. 1.8 K/D is massively above average for a light. I would bet that 1.0 K/D is above average for a light.

And no, KDR should have nothing to do with matchmaker, as stealing kills at low-Elo is a completely meaningless statistic, but would suddenly matter with your frankencalculation.

Edited by PEEFsmash, 30 September 2013 - 03:32 PM.


#3 Rascula

    Member

  • PipPipPipPipPipPip
  • Talon
  • Talon
  • 387 posts
  • LocationWord of Blake Protectorate, Epsilon Eridani.

Posted 30 September 2013 - 03:36 PM

Im not aware that ELO looks at weight classes when matching people, although i'm hardly an expert on it.

But as to "balancing elo and KDA for every class resulting in better gameplay" you may well be right if you like hanging around at the back and stealing kills...

Edited by Rascula, 30 September 2013 - 03:37 PM.


#4 Farix

    Member

  • PipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 890 posts

Posted 30 September 2013 - 03:42 PM

View PostRascula, on 30 September 2013 - 03:36 PM, said:

But as to "balancing elo and KDA for every class resulting in better gameplay" you may well be right if you like hanging around at the back and stealing kills...


Given his other topic where he complains that someone with a kill assist with lots of damage gets more c-bills than someone who gets a kill with very little damage, you are probably not far off the mark.

#5 Homeless Bill

    Member

  • PipPipPipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 1,968 posts
  • LocationA Box Near You

Posted 30 September 2013 - 04:12 PM

I know a lot of high KDR players that only have that KDR because they camp, feed their PUGs to the meat grinder, and then swoop in and clean up the mess. It encourages playing like an opportunistic *******. Bill does not approve.

Also, average KDR should be almost exactly 1.0. There is a death for every kill. There is not always a kill for every death, but it's pretty close. My guess would be 0.9 for lights, 0.8 for mediums, 1.1 for heavies, and 1.2 for assaults.

#6 Hot Rod

    Member

  • PipPipPipPipPipPip
  • 279 posts
  • LocationGermany

Posted 30 September 2013 - 04:25 PM

View PostHomeless Bill, on 30 September 2013 - 04:12 PM, said:

I know a lot of high KDR players that only have that KDR because they camp, feed their PUGs to the meat grinder, and then swoop in and clean up the mess. It encourages playing like an opportunistic *******. Bill does not approve.

Also, average KDR should be almost exactly 1.0. There is a death for every kill. There is not always a kill for every death, but it's pretty close. My guess would be 0.9 for lights, 0.8 for mediums, 1.1 for heavies, and 1.2 for assaults.


i second that,besides the k/d ratios...

i always pug, i think i am an average player but even my k/d is

3.x for lights
1.8 for meds
1.6 for heavys
and 2.5 for assaults.

Contrary, my win/loss is Wins / Losses 646 / 673
Its balanced, 10 bad games, 9 good ones. And i am sure i don't want to get grouped with killstealers. Because they cant kill something by themselves, period.

#7 OneEyed Jack

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,500 posts

Posted 30 September 2013 - 05:48 PM

This idea is about the level of half-baked nonsense I've come to associate with the OP, for reasons already sufficiently enumerated.

However,I would agree that more personal control of individual progression would be a huge boon. Though I'm not sure what criteria could be used determine contribution while not being abuseable. It would have to be a broad spectrum of contributions, many of which offer no good way to measure them, to encompass various roles while remaining balanced between those roles.

#8 Hellcat420

    Member

  • PipPipPipPipPipPipPipPip
  • Bad Company
  • 1,520 posts

Posted 30 September 2013 - 06:16 PM

View PostMadCat02, on 30 September 2013 - 03:01 PM, said:

As 1/12 of your team you can only effect your win rate so much .

For those of you who don't know MWO ELO is different for every weight class . Thats why i wait 20 minutes to play as assault and 30 seconds as a light mech .

Lets say i have 2.0 win rate with 4.0 KDA on assaults . Its about 66% win rate so if you match me agaist 60% win rate assaults with 1.8 kda its not very fair .

Unfornutly WIN/LOSS is mostly determined by a match making luck with some rare close games .

This is not Quacke where 1 good player can kill everyone . If i am matched agaist a group of skilled players it really dosen't matter how much i try . Nobody likes that .

......................................

I undestand that a light mech can't have as many kills as heavys or assault .

However we already have elo based on a weight class so kda can be specific to a weight class as well .

My gauess is 1.8 KDA is above avareg on a light mech pilot .

2.0 For mediums

2.5-4.0 For heavys and assaults .

I think balance between elo and KDA for every class would result in better gameplay.


lol please let me know when you get back to reality. i dont know what game you are playing but it certainly isnt mechwarrior online if you really think that would result in better gameplay.

#9 ICEFANG13

    Member

  • PipPipPipPipPipPipPipPip
  • Elite Founder
  • Elite Founder
  • 1,718 posts

Posted 30 September 2013 - 06:29 PM

The average KDR of all players, although some are not playing anymore is 1, anything above that is above average and anything below is below average.

#10 PEEFsmash

    Member

  • PipPipPipPipPipPipPipPip
  • Philanthropist
  • Philanthropist
  • 1,280 posts
  • LocationLos Angeles

Posted 30 September 2013 - 06:42 PM

View PostHot Rod, on 30 September 2013 - 04:25 PM, said:


i second that,besides the k/d ratios...

i always pug, i think i am an average player but even my k/d is

3.x for lights
1.8 for meds
1.6 for heavys
and 2.5 for assaults.

Contrary, my win/loss is Wins / Losses 646 / 673
Its balanced, 10 bad games, 9 good ones. And i am sure i don't want to get grouped with killstealers. Because they cant kill something by themselves, period.


It looks like you are the kill-stealer, and you aren't doing enough to actually win games. I know plenty of players with a far higher winrate than you, but a lower K/D.

Winning is what matters, not last-hitting.

Edited by PEEFsmash, 30 September 2013 - 06:43 PM.


#11 Jin Ma

    Member

  • PipPipPipPipPipPipPipPip
  • 1,323 posts

Posted 30 September 2013 - 06:45 PM

what does your KDA matter if you can't win

#12 Roland

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • 8,260 posts

Posted 30 September 2013 - 07:28 PM

If you are pugging though, your kdr is pretty much your best representation of your contribution. You are unlikely to be able to contribute via some complex coordination with pugs, and if you are killing multiple mechs before you go down, then you generally did your part.

Of course, at least in my experience, a high kdr tends to go hand in hand with a higher than average w/l ratio. However, in 12v12 games it is harder to carry a team to victory. You can score a thousand damage and get 6 kills and still somehow lose. I haven't played enough since release to see if the new matchmaking changes really helped this.

Sure, sometimes you might get lucky and just happen to get the last shot in, but I doubt someone could pull that off often enough to have kill stealing be statistically significant.

Getting the kill means you were shooting at a crit target, which in itself is generally the right thing to do. Certainly, I'd rather have someone on my team stealing kills as opposed to folks who walk around without even targeting mechs .

#13 xMEPHISTOx

    Member

  • PipPipPipPipPipPipPipPip
  • Elite Founder
  • Elite Founder
  • 1,396 posts

Posted 30 September 2013 - 07:30 PM

KDR is cool, but high WIN % is OP.
Would rather win w/no kills and a death than lose with kills and being alive at end.
But an ELO system that took both into account would not be a bad thing.

#14 Marj

    Member

  • PipPipPipPipPipPip
  • 215 posts

Posted 30 September 2013 - 08:22 PM

KDR probably isn't the best stat to use, but I agree something needs to be done to fix the matchmaker. Maybe something along the lines of the top 8 match scores (anyone know how is this calculated?) get a boost to ELO separate from the change due to win/loss, the middle 8 stay the same and the bottom 8 go down.

I've been grinding centurions and I'm sick of regularly getting a high match score/damage (usually top 3 unless I do something stupid) relative to other people on my team but still getting stuck with some really bad teams where 4-6 people have sub 50 damage because my w/l ratio is 1:1.

Maybe the brackets are just too large (not enough players) and it's sucking in too wide a range of skill levels to create a balanced match. I don't play much anymore, has anyone noticed a significant difference in the range of skill levels within a match between peak and off peak times?

#15 Roland

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • 8,260 posts

Posted 30 September 2013 - 08:29 PM

Just to be clear, certainly real matches can't just boil down to KDR...

But if you are on a team, and you kill 6 mechs and do a mountain of damage, only to lose through your team's incompetence.. but some guy on the other team did literally NOTHING, but just happened to be carried to victory... does it really make sense for his Elo to go up and yours to go down?

I guess the theory is that it'll balance out, and my experience seems to be that it does in fact do so... I tend to see the same folks, who I regard as competent players, more often than I see the random mouthbreathers... so Elo is doing something. But at the same time, I can certainly see a flaw with the application of Elo to a team based game.

#16 Hot Rod

    Member

  • PipPipPipPipPipPip
  • 279 posts
  • LocationGermany

Posted 01 October 2013 - 01:08 AM

View PostPEEFsmash, on 30 September 2013 - 06:42 PM, said:


It looks like you are the kill-stealer, and you aren't doing enough to actually win games. I know plenty of players with a far higher winrate than you, but a lower K/D.

Winning is what matters, not last-hitting.


Winning is what matters...Oh you got the point XD

but try to tell that to the other 11 "random" players.


When they sometimes manage to get killed till minute 2 into the game, all 11? while you killed your first enemy? that walked up behind your assault? that was left alone on the way to the front line?

When your the light that fakecaps only to come back minutes later to see 7+ dead friendly players?

What would you do?

I fight till i'm dead and usually i take someone with me. Sometimes more sometimes less.


Roland got that right.

View PostRoland, on 30 September 2013 - 07:28 PM, said:

If you are pugging though, your kdr is pretty much your best representation of your contribution. You are unlikely to be able to contribute via some complex coordination with pugs, and if you are killing multiple mechs before you go down, then you generally did your part.

Of course, at least in my experience, a high kdr tends to go hand in hand with a higher than average w/l ratio. However, in 12v12 games it is harder to carry a team to victory. You can score a thousand damage and get 6 kills and still somehow lose. I haven't played enough since release to see if the new matchmaking changes really helped this.

Sure, sometimes you might get lucky and just happen to get the last shot in, but I doubt someone could pull that off often enough to have kill stealing be statistically significant.

Getting the kill means you were shooting at a crit target, which in itself is generally the right thing to do. Certainly, I'd rather have someone on my team stealing kills as opposed to folks who walk around without even targeting mechs .


The worst thing is "support" isn't a stat.

In pug, esp now with 12 vs 12, you can't control the outcome, In 8 vs 8 it was possible but not now.
Now you just kill things and hope your teammates know how to survive the 5 minute breakpoint and know how to pull the trigger effectively.

Ultimately your elo depends on you AND your team. You can be as good as you want to be but when your team sucks, it takes your elo down with them.

#17 Krivvan

    Member

  • PipPipPipPipPipPipPipPipPip
  • Littlest Helper
  • Littlest Helper
  • 4,318 posts
  • LocationUSA/Canada

Posted 01 October 2013 - 01:14 AM

Your Elo changes slower, yes, but you have a bias towards more wins or more losses depending on your play. Otherwise we wouldn't have people with 0.7 w/l ratios and people with 1.8 w/l ratios.

Edited by Krivvan, 01 October 2013 - 01:14 AM.


#18 Devil Fox

    Member

  • PipPipPipPipPipPipPipPip
  • The Infernal
  • The Infernal
  • 1,393 posts
  • LocationThe Fox Den

Posted 01 October 2013 - 02:16 AM

I like how this topic draws out all the "top" light pilots...

I don't agree with KDA being associated with ELO, it's bad enough that high tier players have to carry the offensive because the lower players are used to sitting back trying to preserve armour in machines meant for a fight (seriously you have an Atlas DDC, they split up... and you still don't push to core out their damaged machine?!).

What they need to a tighter balancing of ELO and the vaunted "tonnage based" matchmaking to come into effect. That will then determine how the game scene will change. Whether for the good or bad, we need the system to see.

#19 Mycrus

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Overlord
  • Overlord
  • 5,160 posts
  • Google+: Link
  • Twitter: Link
  • LocationFilipino @ Singapore

Posted 01 October 2013 - 02:24 AM

We need to balance elo with e-peen...

You haz a slider that allows you to most likely drop against higher elo than you...

I'd also like a hardcore 4-man mode... allowing your 4-man to group with other random 4-mans vs 12s...

#20 Devil Fox

    Member

  • PipPipPipPipPipPipPipPip
  • The Infernal
  • The Infernal
  • 1,393 posts
  • LocationThe Fox Den

Posted 01 October 2013 - 03:58 AM

View PostMycrus, on 01 October 2013 - 02:24 AM, said:

We need to balance elo with e-peen...

You haz a slider that allows you to most likely drop against higher elo than you...

I'd also like a hardcore 4-man mode... allowing your 4-man to group with other random 4-mans vs 12s...


That would open up the 12's queues more... nothing like spending an hour and a half getting a group and trying to drop into a dead 12's queue before giving up. I think the slider idea has to come with an increased reward VS risk system to earn more cash... but then this could just result in feeding low ELO player to the vultures that are the high ELO bracket.





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users