Jump to content


Why efficiency is flawed.

efficiency XVM

This topic has been archived. This means that you cannot reply to this topic.
23 replies to this topic

Nick_the_Tank_Driver #1 Posted Jul 22 2012 - 19:05

    Staff sergeant

  • Players
  • 4030 battles
  • 407
  • Member since:
    03-16-2012
Attached File   shot_004.gif   25.13K

I think this picture speaks for itself.

In case you can't see it, that BT-2 has an efficiency of over 3000 with under 1k battles.

(Yes, I do play my PzII to beat up on noobs relieve stress sometimes.)

tankcrunch #2 Posted Jul 22 2012 - 19:09

    Major

  • Players
  • 22654 battles
  • 8,598
  • Member since:
    04-21-2011
so its supposed to mean?

ScionVyse #3 Posted Jul 22 2012 - 19:10

    Captain

  • Players
  • 7780 battles
  • 1,007
  • [PONY] PONY
  • Member since:
    08-13-2011
That's called a small sample size. Statistics don't work properly without a large sample. While I agree efficiency is flawed, that really doesn't help the argument.

KarateF22 #4 Posted Jul 22 2012 - 19:10

    Major

  • Beta Testers
  • 12011 battles
  • 2,632
  • [VILIN] VILIN
  • Member since:
    01-21-2011
I don't think anyone has ever really said efficiency is the be all end all determination of skill. Its just a useful tool which correlates pretty strongly with how skilled someone is. It doesn't guarantee anything, its just a bell curve of probability... and like all bell curves, there are still things at each end of the bell curve.

Ironlight #5 Posted Jul 22 2012 - 19:11

    Staff sergeant

  • Players
  • 6368 battles
  • 443
  • Member since:
    04-18-2012
That's why number of battles is included, so if it's under 1000 you can put it in perspective. After all, the guy has an 83% win rate. Either he's really awesome or he's played so few games that his win rate is skewed by a few good win streaks. Who knows.

And as mentioned, posting one picture from XVM and saying it's an indication of why efficiency is flawed is not a very strong argument.

Edited by Ironlight, Jul 22 2012 - 19:15.


tallstar5 #6 Posted Jul 22 2012 - 19:12

    Major

  • Players
  • 9022 battles
  • 3,658
  • [KV-2] KV-2
  • Member since:
    01-14-2012
It's only flawed because of the small sample size. You could say that of any stat without an adequate number of samples.

Nick_the_Tank_Driver #7 Posted Jul 22 2012 - 19:12

    Staff sergeant

  • Players
  • 4030 battles
  • 407
  • Member since:
    03-16-2012
My point is that , if this guy continued playing his BT-2 like this (which was suiscouting) he could bump up his efficency without even being skilled.

Hogan_ #8 Posted Jul 22 2012 - 19:15

    Sergeant

  • Players
  • 41843 battles
  • 241
  • [HARM] HARM
  • Member since:
    04-02-2011
who cares?

The_Millard #9 Posted Jul 22 2012 - 19:15

    Major

  • Players
  • 23035 battles
  • 5,606
  • [M--M] M--M
  • Member since:
    04-01-2011
Everyone informed believes eff is flawed.

/nuffsaid

SourKraut13 #10 Posted Jul 22 2012 - 19:35

    Major

  • Players
  • 19300 battles
  • 4,332
  • [27PZR] 27PZR
  • Member since:
    12-14-2011
Efficiency Rating is very flawed. I've completed several battles where I have 35% or better on cap or defending cap, and we win because we killed all the tanks. In all these battles, my rating is over 5000. So it is obvious that it ranks capping, resetting cap higher than damage done. Also in battles I have done 3 times the damage (2500+) in a game with no capping and only max out at 1900 efficiency. Someone who camps at base and waits for the enemy to start capping our base, then resetting cap, although will have a terrible win rate, will have an awesome efficiency.

XVM is flawed in that it only shows the overall stats, and not the tank that's being played stats. That's why when it tries to "predict" the winner, it's way off...

Question: What is XVM going to show in game when you play me? Look at my signature and you will see why I get called a "hack" allot!

Cybergod #11 Posted Jul 22 2012 - 19:38

    Major

  • Beta Testers
  • 18064 battles
  • 3,383
  • [DERP] DERP
  • Member since:
    09-16-2010
Everything in the game or has to do with the game is flawed.

Azyur #12 Posted Jul 22 2012 - 19:45

    Staff sergeant

  • Beta Testers
  • 18132 battles
  • 440
  • [MOB] MOB
  • Member since:
    01-25-2011
The efficiency calculators are both flawed and irrelevant. However, fail pubbers and dilettantes love using them as troll bait, or exclusionary devices. So, even with sample size errors (and and category biases), the E-Calcs serve their intended purpose as trolling instruments for gaming elitism. As with many silly things in the WoT universe : Working As Intended.

Gryphon_ #13 Posted Jul 22 2012 - 19:58

    Captain

  • Players
  • 19369 battles
  • 1,149
  • [COD] COD
  • Member since:
    11-24-2011

View PostSourKraut13, on Jul 22 2012 - 19:35, said:

...
XVM is flawed in that it only shows the overall stats, and not the tank that's being played stats. That's why when it tries to "predict" the winner, it's way off...
That isnt accurate. XVM can show the tank that's being played stats, but you need to edit the config file to get that info (which is very useful, by the way)

Edited by Gryphon_, Jul 22 2012 - 20:00.


Cleostrasz #14 Posted Jul 22 2012 - 19:59

    Sergeant

  • Players
  • 13210 battles
  • 229
  • [CH4I] CH4I
  • Member since:
    04-11-2012

View PostNick_the_Tank_Driver, on Jul 22 2012 - 19:12, said:

My point is that , if this guy continued playing his BT-2 like this (which was suiscouting) he could bump up his efficency without even being skilled.


The point is it stilll wouldn't matter once you look at his record and only those few people with XVM can see.  A week ago someone bragged about eff...he had 50% marder III games.  The general chat pooped on him.  Take argumentative writing in college and you'll learn a lot.  Also, you died again in the first minute of a battle. Why do you keep posting about lemmings, stats, and XVM still?

Nick_the_Tank_Driver #15 Posted Jul 22 2012 - 20:04

    Staff sergeant

  • Players
  • 4030 battles
  • 407
  • Member since:
    03-16-2012

View PostJaparican, on Jul 22 2012 - 19:59, said:

The point is it stilll wouldn't matter once you look at his record and only those few people with XVM can see.  A week ago someone bragged about eff...he had 50% marder III games.  The general chat pooped on him.  Take argumentative writing in college and you'll learn a lot.  Also, you died again in the first minute of a battle. Why do you keep posting about lemmings, stats, and XVM still?
What do you mean I died again in the first minute of the battle? I survived this one. You stalking me or something? And "the point" is that if it is even possible to have a 3000 efficiency at less than 1000 games, something, regardless of how much it is negated with a sufficient sample size, is borked in the calculations. And you think you have the right to tell me what I can post about? Go stalk someone else, why don't you?

ApplesauceBandit #16 Posted Jul 22 2012 - 20:14

    Major

  • Wiki Staff
  • 21870 battles
  • 5,178
  • [CMFRT] CMFRT
  • Member since:
    05-17-2011
There's a guy that I platoon with sometimes who has only a BT-2, a MS-1, a T-26, and a T-50 and a very high winrate.  He might be the one in the screenshot.  Does the name start with k and end with s?  I think it may be an alt account because he tends to do just as good or better than me (I have over 9000 games).

Ekscalybur #17 Posted Jul 22 2012 - 20:16

    Captain

  • Players
  • 21745 battles
  • 1,826
  • Member since:
    05-29-2011

View PostSourKraut13, on Jul 22 2012 - 19:35, said:

Efficiency Rating is very flawed. I've completed several battles where I have 35% or better on cap or defending cap, and we win because we killed all the tanks. In all these battles, my rating is over 5000. So it is obvious that it ranks capping, resetting cap higher than damage done. Also in battles I have done 3 times the damage (2500+) in a game with no capping and only max out at 1900 efficiency. Someone who camps at base and waits for the enemy to start capping our base, then resetting cap, although will have a terrible win rate, will have an awesome efficiency.

XVM is flawed in that it only shows the overall stats, and not the tank that's being played stats. That's why when it tries to "predict" the winner, it's way off...

Question: What is XVM going to show in game when you play me? Look at my signature and you will see why I get called a "hack" allot!

You do realize XVM can be modified to show per tank stats, right? You do know that XVM's victory percentage is probability, and that it already shows overall AND per tank win percentages, right?

For the part I've bolded. How can XMV be way off, when it has never once, in the entire history of the mod, ever said "You will Win!". It gives chances, probability, says things may happen a certain way. It is IMPOSSIBLE for it to be inaccurate, since it never stats something will happen. It just says something is likely to happen.

Nick_the_Tank_Driver #18 Posted Jul 22 2012 - 20:18

    Staff sergeant

  • Players
  • 4030 battles
  • 407
  • Member since:
    03-16-2012

View PostApplesauceBandit, on Jul 22 2012 - 20:14, said:

There's a guy that I platoon with sometimes who has only a BT-2, a MS-1, a T-26, and a T-50 and a very high winrate.  He might be the one in the screenshot.  Does the name start with k and end with s?  I think it may be an alt account because he tends to do just as good or better than me (I have over 9000 games).
Nope, that isn't him. This player was decent, but there was no way he deserved a 3k efficency. He mostly ran around the map lighting a few tanks up.

Lhos #19 Posted Jul 22 2012 - 20:27

    Major

  • Players
  • 9932 battles
  • 3,357
  • Member since:
    05-11-2011
HEY GUYS, LOOK HOW BROKEN THIS STUPID SYSTEM IS

IT CAN'T EVEN GENERATE ACCURATE STATS WITHOUT ENOUGH BATTLES TO USE

JEEZ WHAT A USELESS THING

/notsureiftrolling

KilgorSoS #20 Posted Jul 22 2012 - 20:40

    Major

  • Players
  • 26602 battles
  • 11,792
  • Member since:
    07-23-2011

View PostScionVyse, on Jul 22 2012 - 19:10, said:

That's called a small sample size. Statistics don't work properly without a large sample. While I agree efficiency is flawed, that really doesn't help the argument.

Yes, very correct. I saw one like this while driving my Tetrarch, and the person had 25 battles, and did well, hence the anamoly. I wrote the player's name down, and within a couple of weeks it was at 540 efficiency, and 200+ battles.

Anamolies do not negate, anything...regardless of which efficiency calc used. WR for any tank, or player is worthless for the first 100/1000 battles, and so is efficiency.