Jump to content

- - - - -

Why Not To Judge Performance By Score


58 replies to this topic

#1 Victor Morson

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Elite Founder
  • 6,370 posts
  • LocationAnder's Moon

Posted 18 November 2013 - 07:04 PM

So, I see this a lot in these threads: Someone posts a screenshot as "proof" of how awesome their build is, with a high damage number. "See, my MGs are awesome because look how much damage I did!" is a common sentiment, but we also see the same thing for franken builds, designs that will explode if the armor gets cracked, etc.

However this doesn't mean anything at all really. And here's why:
  • The bottom line is in pug games, even at higher ELOs - it always mixes in newbies. Newbies that will sit there and let you shoot at them. Forever. Many games I can score over 1k damage easily just off of farming people who aren't paying attention.
  • Damage is not everything in MechWarrior. Case in point: Two pilots battle it out. One guy gave 150 damage. One guy gave 450. It's entirely likely the guy with 150 won the fight. I routinely end my Centurion runs 25-35% intact or lower, giving the enemy team huge damage numbers... but I'm still standing and they are dead. Accurate damage beats "just damage" anytime.
  • Because of the first two factors, it rewards pilots who blow apart newbies piece by piece with huge damage numbers. More efficient pilots would have ended them in half the damage or better.
  • All these screenshots are incredibly cherry picked to skew the results, even if the poster claims it's an "average."
So next time you are talking about mech performance and want to argue a design, don't go whipping out the screenshot-of-proof. It doesn't mean anything. I take that back, it does mean something: It means you don't understand why it's not effective evidence, and that probably hurts your case more than helps it.

#2 Alaskan Nobody

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • The Determined
  • The Determined
  • 10,358 posts
  • LocationAlaska!

Posted 18 November 2013 - 07:12 PM

I agree with the above, but always remember the 1st rule of rules:
There are exceptions to every rule including (and excluding :D) this one.

Having said that - the best builds are the fun ones - and fun is subjective.
I have been mocked since day 3 for my builds - but I have fun, and I (usually) help my team. ;)
(Woulda been day 1 but it took a while to get your own mech before they added the new player bonus)

#3 Fenris Krinkovich

    Member

  • PipPipPipPipPip
  • 177 posts
  • LocationWestfall, OK

Posted 18 November 2013 - 07:22 PM

This have anything to do with a certain Cataphract build?

#4 BillyM

    Member

  • PipPipPipPipPipPipPip
  • 530 posts

Posted 18 November 2013 - 07:46 PM

Average kills/game
Average deaths/game

...those are the stats to watch.

That said, damage is a good indicator of your comfort between loadouts.
Running all lasers and doing 150dmg, switch to AC5's and doing 400?

Well, that counts!

--billyM

Edited by BillyM, 18 November 2013 - 07:48 PM.


#5 Mycrus

    Member

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

Posted 18 November 2013 - 07:47 PM

View PostVictor Morson, on 18 November 2013 - 07:04 PM, said:

Screeeenshots are baaaad...



What about 12v12 videos??

#6 Buckminster

    Member

  • PipPipPipPipPipPipPipPip
  • Elite Founder
  • 1,577 posts
  • LocationBaltimore, MD

Posted 18 November 2013 - 07:53 PM

But ehrmagerd, look at all the damage I did in my Locust! I rule!!!! ;) (I know people have done more, but it's the example I have)

Posted Image

But then you look at it more closely... All that damage was against one mech - a lone 'phract that I pecked to death with my streaks and small laser. He was obviously a newer pilot, because I could just run around him in circles all match and he couldn't really do much to me. He was panicking to the point of shutting down, on Alpine. And the simple fact is, we lost. So I spent all match playing tag with a heavy, when I probably should have been more focused on getting a cap and winning the game.

Are high damage numbers nice? Sure. But it's not everything.

Edited by Buckminster, 18 November 2013 - 07:53 PM.


#7 WVAnonymous

    Member

  • PipPipPipPipPipPipPipPip
  • Wrath
  • Wrath
  • 1,691 posts
  • LocationEvery world has a South Bay. That's where I am.

Posted 18 November 2013 - 08:19 PM

I really enjoy matches where I have 3 kills and under 200 damage.

Of course I also enjoy 3 kills and 750 damage!

Why, my KDR has climbed all the way from 0.28 to 0.65 now. :(

#8 Alaskan Nobody

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • The Determined
  • The Determined
  • 10,358 posts
  • LocationAlaska!

Posted 18 November 2013 - 08:25 PM

View PostWVAnonymous, on 18 November 2013 - 08:19 PM, said:

I really enjoy matches where I have 3 kills and under 200 damage.

Of course I also enjoy 3 kills and 750 damage!

Why, my KDR has climbed all the way from 0.28 to 0.65 now. :(


Agreeing with you, but playing devils advocate
(maybe I aughta make another account with that as a name.... :ph34r: ....naw, to much effort)

A 3 Kill 200 damage match may have been you 'kill stealing!!1!'
and the 3 kill 700+ match may have been you 'kill stealing!!1!11111!!' and being a lucky-yet-crappy shot at the same time. :ph34r:

#9 Tesunie

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The Seeker
  • The Seeker
  • 8,586 posts
  • LocationSeraphim HQ: Asuncion

Posted 18 November 2013 - 08:32 PM

View PostBillyM, on 18 November 2013 - 07:46 PM, said:

Average kills/game
Average deaths/game

...those are the stats to watch.

That said, damage is a good indicator of your comfort between loadouts.
Running all lasers and doing 150dmg, switch to AC5's and doing 400?

Well, that counts!

--billyM


Even the old kill/death numbers are useless. I often times end a match with no kills, but a ton of assists. It is not uncommon for me to shoot up a foe, just for someone else to sneak in the last couple of points of damage to kill them.

Damage is a nice indicator, but also can be considered a poor indicator.

Often times its how you play the mech as much as anything else. Damage can be spread out ineffectively. Kills can be snatched at the last moment (or stolen even depending upon your point of view).

The biggest factor to a mech is how much fun it is to play, which is a stat that can't be determined.

#10 WVAnonymous

    Member

  • PipPipPipPipPipPipPipPip
  • Wrath
  • Wrath
  • 1,691 posts
  • LocationEvery world has a South Bay. That's where I am.

Posted 18 November 2013 - 10:40 PM

View PostShar Wolf, on 18 November 2013 - 08:25 PM, said:


Agreeing with you, but playing devils advocate
(maybe I aughta make another account with that as a name.... :ph34r: ....naw, to much effort)

A 3 Kill 200 damage match may have been you 'kill stealing!!1!'
and the 3 kill 700+ match may have been you 'kill stealing!!1!11111!!' and being a lucky-yet-crappy shot at the same time. :(


Better lucky than good...

#11 Koniving

    Welcoming Committee

  • PipPipPipPipPipPipPipPipPipPipPipPipPip
  • The Guide
  • The Guide
  • 23,384 posts

Posted 18 November 2013 - 10:45 PM

View PostWVAnonymous, on 18 November 2013 - 08:19 PM, said:

I really enjoy matches where I have 3 kills and under 200 damage.

Of course I also enjoy 3 kills and 750 damage!

Why, my KDR has climbed all the way from 0.28 to 0.65 now. :(


View PostShar Wolf, on 18 November 2013 - 08:25 PM, said:

Agreeing with you, but playing devils advocate
(maybe I aughta make another account with that as a name.... :ph34r: ....naw, to much effort)

A 3 Kill 200 damage match may have been you 'kill stealing!!1!'
and the 3 kill 700+ match may have been you 'kill stealing!!1!11111!!' and being a lucky-yet-crappy shot at the same time. :ph34r:


You mean like this?
Posted Image

o.O; Source match.

#12 Koniving

    Welcoming Committee

  • PipPipPipPipPipPipPipPipPipPipPipPipPip
  • The Guide
  • The Guide
  • 23,384 posts

Posted 18 November 2013 - 10:50 PM

View PostTesunie, on 18 November 2013 - 08:32 PM, said:

Often times its how you play the mech as much as anything else. Damage can be spread out ineffectively. Kills can be snatched at the last moment (or stolen even depending upon your point of view).

The biggest factor to a mech is how much fun it is to play, which is a stat that can't be determined.


The two most important things of all.

Mechs that are awful at one play style can perform superbly in a completely different play style. Just because a mech doesn't fit how one person plays doesn't mean it's as awful as one might think. It's why I try to cater any designs made specifically for someone to play to that person's strengths so far as I'm told of them. (Thus it's important to tell me everything about how you play and what you enjoy doing). This also means that just because I built something for (Random name) doesn't mean it'll be ideal for you; becase you and (Random name) probably don't play the same way. Meanwhile many others build mechs for their playstyle and somehow expect a new player to perform just as well. Sorry to say but it doesn't really work that way. You can't give the wizard's wand to the fighter and you can't give heavy chainmail to the nimble archer and expect them to be able to handle it. What you can do is design builds to enable their strengths and preferences.

I'm a jack of all trades and enjoy lots of games and playstyles, so I can get away with some pretty crazy stuff.

On to the other important thing and honestly this is more important than the one above. "How fun it is to play" is a factor that can be measured by the size of the grin on your face and how loud you are when you do something awesome. This factor can only be multiplied if you can transfer that smile onto your mech beside the barrel that's going to ruin your opponent's day.

Posted Image
Smile!

Edited by Koniving, 18 November 2013 - 11:01 PM.


#13 100mile

    Member

  • PipPipPipPipPipPipPipPip
  • Elite Founder
  • Elite Founder
  • 1,235 posts
  • LocationAlegro: Ramora Province fighting Pirates. and the occasional Drac

Posted 18 November 2013 - 10:56 PM

View PostBillyM, on 18 November 2013 - 07:46 PM, said:

Average kills/game
Average deaths/game

...those are the stats to watch.



--billyM

No the only stat that matters is Win Ratio....Yes I realize that depends on your teams...but isn't that the point? If you can maintain a high win ratio even when you pug...doesn't that mean you are being extremely effective at supporting your team...which by the way is the basis for this game...

#14 Koniving

    Welcoming Committee

  • PipPipPipPipPipPipPipPipPipPipPipPipPip
  • The Guide
  • The Guide
  • 23,384 posts

Posted 18 November 2013 - 11:15 PM

View Post100mile, on 18 November 2013 - 10:56 PM, said:

No the only stat that matters is Win Ratio....Yes I realize that depends on your teams...but isn't that the point? If you can maintain a high win ratio even when you pug...doesn't that mean you are being extremely effective at supporting your team...which by the way is the basis for this game...

Agreed for performance of the player as a whole. I agree completely.

But by "performance" the original poster is referring to people like myself who say: "zomg my build is so good, look at mah score!" Which is a way to judge whether or not the mech's build is or weapons are good. Not whether the player is good. Win/loss can't help determine if the build itself is good.

For example I could have a win/loss ratio of 5,000,000/1. But I could also have 5,000,000 deaths and a single kill. Doesn't say a thing about my mech's build.

I could pull out a score like this, and say my blackjack is incredible everyone should try it. But what's to say the two enemies weren't afk players? (They weren't, but you get my point). 8 small lasers, btw.
Posted Image

But if I go into matches and rack in a thousand damage and multiple kills consistently, it says something about my mech.

Like this from my infamous Heavy Metal rig.
Posted Image
Edit: The match below can be found on this video. The build's basics can be found at the start (it is time skipped in as it's part of the Battlegrid tutorial vid in which I'm trying, during a match recording the instructions live, to teach people how to use the battlegrid.)
Posted Image

Or these, from my AC/2, AC/5 + 2 streak Trebuchet Joke build.
Posted Image
Posted Image

A single image doesn't mean much. But when the results are consistent, it can say something towards the performance. However, even with consistency, what if these were two matches out of 9,387? Truth be told both Trebuchet shots here are when my entire team had essentially fallen and left me to clean up and in both cases I am completely out of ammo (the build has 2 tons of ammo per weapon and it burns fast). But how would you know? You wouldn't.

This is why I prefer video myself.

----------

If you want to really be able to gauge the performance of a build, run it. Have several similarly minded people pick up a build and just roll with it at the same time and get their impressions.

For example, the flamer Awesome. (Vid 2)

Or Lordred's unnamed fast attack Cataphract. (Vid 2)

Edited by Koniving, 18 November 2013 - 11:38 PM.


#15 WVAnonymous

    Member

  • PipPipPipPipPipPipPipPip
  • Wrath
  • Wrath
  • 1,691 posts
  • LocationEvery world has a South Bay. That's where I am.

Posted 19 November 2013 - 12:08 AM

View PostKoniving, on 18 November 2013 - 11:15 PM, said:


But if I go into matches and rack in a thousand damage and multiple kills consistently, it says something about my mech.

Like this from my infamous Heavy Metal rig.
Posted Image
Ed


Unfair comparison, you launched one short. More opportunity to rack up damage. :(

#16 Koniving

    Welcoming Committee

  • PipPipPipPipPipPipPipPipPipPipPipPipPip
  • The Guide
  • The Guide
  • 23,384 posts

Posted 19 November 2013 - 12:32 AM

View PostWVAnonymous, on 19 November 2013 - 12:08 AM, said:

Unfair comparison, you launched one short. More opportunity to rack up damage. :(


Worse. Within the first minute, we had 3 disconnects and they had 2. All simulataineously. While some of my teammates were busy killing disconnected mechs I was swarmed with enemies as they dropped like flies. So we were 4 short, versus 2 short. 8 to 10. They even got free kills because of some of my team's players completely oblivious to being attacked while they concentrated on the two disconnects.

Thankfully Stalkers don't have much of a chance against a Heavy Metal leaping over their heads (they can't turn for jack) and some of those mechs weren't really a threat.

#17 Greyboots

    Member

  • PipPipPipPipPipPip
  • Liquid Metal
  • Liquid Metal
  • 396 posts

Posted 19 November 2013 - 04:18 AM

View Post100mile, on 18 November 2013 - 10:56 PM, said:

No the only stat that matters is Win Ratio....Yes I realize that depends on your teams...but isn't that the point? If you can maintain a high win ratio even when you pug...doesn't that mean you are being extremely effective at supporting your team...which by the way is the basis for this game...


Win/Loss is the only stat that matters to the game and is a great insight into how the Devs have structured MWO. While it depends on your team for sure, it's a measure of how well you are capable of performing as a member of a functional team.

When you are trying to show something else like how effective a particular build is? It becomes necessary to use different parameters. This tends to be when the offensive stats you get handed matter because you are trying to show something that win/loss ratio isn't much of a guide for.

It's always good to remember that a black and white viewpoint can sometimes blind you to to how the tools on offer can offer you insights into what's going on in the game :(. Remaining objective is always my advice when looking at stats people post!

#18 42and19

    Member

  • PipPipPipPipPip
  • Philanthropist
  • Philanthropist
  • 197 posts

Posted 19 November 2013 - 04:50 AM

I think Koniving has it right. When I test new builds I stay with the build for at least 10 drops. Constant performance is what you are looking for. I've put together builds were my first 5 drops were completely trash but they have turned out to be rather good. The opposite is true as well. There are so many factors that can effect your game performance that going off just one or two matches is completely silly.

Something to consider with the damage stat is what weapons your mech is using. I expect a much lower damage output with my hunchback than I do with my splatmaster. One is precision damage and the other is spread damage. However, both are viable builds with many successful games.

#19 Modo44

    Member

  • PipPipPipPipPipPipPipPipPip
  • Bad Company
  • 3,559 posts

Posted 19 November 2013 - 05:16 AM

All values should be considered, with different weights attached to them -- both for match score calculations, as well as for Elo purposes. The current "Cash based on damage/Elo based on wins/losses" system is too basic in a team game.

#20 ManaValkyrie

    Member

  • PipPipPipPipPipPipPip
  • Mercenary
  • Mercenary
  • 507 posts
  • LocationUK

Posted 19 November 2013 - 06:05 AM

Unfortunately the 'Match Score' is not hugely a great indicator of performance as plenty of people will and can attest too. I have seen players that don't help the team at all, sit back and do little damage till later stages of matches and once the enemy team is whittled down, come in and 'mop up' with high damage alpha strikes.

Match Scores presently are made up from a whole list of things, damage done to the enemy, kills, assists, component destructions, spotting bonuses, narc/tag bonues, UAV detection and assists, capping assists. Unfortunately the score is heavily biased towards damage, more direct damage and component destruction and your score is boosted significantly.

However if you have someone who has an average off 3-400 damage, maybe a kill or two, with assists with a very high match score then they have actually been doing a team effort by doing spotting/tagging.etc and helping others on the team. However, more still needs to be done for 'other' role players.

For example in a conquest match, a couple of your lights go and cap the points, keep the caps rolling in your favour. They see little or no combat then they get hammered for match rewards and also match score, is that right? no, did they not do enough for the team? no, they did there job and got penalized for it.

So ultimately matchscore is a 'varied' indicator of performance and an end of match scoreboard absolutely no indicator of what the match was actually like, I know plenty of good players who get a low score because they are just the unfortunate guy that gets the unlucky first around that corner and into enemy mechs or get the enemy lucky shot on them.





7 user(s) are reading this topic

0 members, 7 guests, 0 anonymous users