Jump to content


Difference between a WR of 49% and 52% is only 3 games out of 100


  • This topic is locked This topic is locked
272 replies to this topic

Gangerr #1 Posted 13 August 2017 - 11:29 AM

    Warrant Officer

  • Player
  • 19871 battles
  • 572
  • Member since:
    03-19-2011
I don't believe you can really take WR ratio as a base for player skill like some people believe it is.  It's information to some people but its only a small part of a players stats.  Unless my maths are really wrong here, a win/lose ratio of 48% like me, means I win 48 games out of every 100 (if im not mistaken, give or take 1 for a draw) but someone who has a W/R of 52 only wins 4 games more than myself.   As a heavy you expect a player to do more damage, as a light scout you expect them to spot more tanks and receive spotting damage...  Correct me if im wrong.

SABAOTH #2 Posted 13 August 2017 - 11:31 AM

    Major

  • Player
  • 36318 battles
  • 2,874
  • [-133-] -133-
  • Member since:
    08-28-2011

View PostGangerr, on 13 August 2017 - 11:29 AM, said:

I don't believe you can really take WR ratio as a base for player skill like some people believe it is.  It's information to some people but its only a small part of a players stats.  Unless my maths are really wrong here, a win/lose ratio of 48% like me, means I win 48 games out of every 100 (if im not mistaken, give or take 1 for a draw) but someone who has a W/R of 52 only wins 4 games more than myself.   As a heavy you expect a player to do more damage, as a light scout you expect them to spot more tanks and receive spotting damage...  Correct me if im wrong.

 

What exactly is the point you are making?

 

It means that for half of your games you do not make any positive difference. Trust me, once you get better you will see that you can influence the outcome of a match with some extra damage to the right target, more than one could say at first glance. :girl:

 

Let me try to be more specific, hoping this helps:

 

Block Quote

 I don't believe you can really take WR ratio as a base for player skill like some people believe it is. It's information to some people but its only a small part of a players stats.

 

It is part of the package, showing your influence over a match

 

Block Quote

Unless my maths are really wrong here, a win/lose ratio of 48% like me, means I win 48 games out of every 100 (if im not mistaken, give or take 1 for a draw) but someone who has a W/R of 52 only wins 4 games more than myself.

 

No there is draw into consideration. And it does not mean automatically that "you win" it means your team wins 48% of the time. It should be 50%, that 2% less means that with you on board the team is 2% more likely to lose, so you have a negative impact.

 

Block Quote

 As a heavy you expect a player to do more damage, as a light scout you expect them to spot more tanks and receive spotting damage...  Correct me if im wrong.

 

 

I am correcting you because you are wrong:

 

Heavy should potentially break a stalemate line, light can potentially spot. But the real deal is to take advantage of the situation. I can do 2000 spotting damage with my heavy while holding a line, as I can do 2000 damage with my light while advancing and lighting up targets.

 

Your job should be to disrupt enemy strategy regardless what class you play (ofc arty is out of the picture)

 

 

Hope this clarifies why people put attention to it :honoring:

 


Edited by SABAOTH, 13 August 2017 - 11:40 AM.


Lil_Dimitry #3 Posted 13 August 2017 - 11:32 AM

    Sergeant

  • Player
  • 2318 battles
  • 202
  • Member since:
    12-07-2014
 The point he's trying to make is that he's not bad at the game, just like every other under 50%er out there that thinks it's not his fault for being bad

Sfinski #4 Posted 13 August 2017 - 11:35 AM

    Lieutenant

  • Player
  • 29721 battles
  • 1,991
  • [-PJ-] -PJ-
  • Member since:
    09-26-2013
Now multiply that 4 game difference with the amount of games you have(18k / 100 = 180) and that's 720 games where someone with the same amount of games as you but with 52% winrate player has made a difference, but you haven't. Thats quite a lot I'd say. Difference grows the more games you play. 

Edited by Sfinski, 13 August 2017 - 11:36 AM.


HundeWurst #5 Posted 13 August 2017 - 11:54 AM

    Brigadier

  • Player
  • 67549 battles
  • 4,172
  • [FAME] FAME
  • Member since:
    02-06-2012

The math is obviously right but the conclusions you drew are fundamentally wrong:

First: 4 more won games does not sound a lot but it is in reality. That pretty much means that the player with 52% winrate carried 4 games in which you "sucked out" so to say. Thats already something.

 

Now to the important part:

You assume that skill and (in this case winrate) correlate in a linear way. However that is not true.

I guess its impossible to know the exact correlation between skill and winrate as skill is a stat which supposingly cannot be messured perfectly with a number. However in order to get these 4% more winrate the player with 52% does not only need to be in absolute number 4% better or in relative numbers 8.3333 percent better but more like 100% better, or even 200% better than the guy with with 48% winrate.

With other words: A player with 52% winrate needs a multiple amount of understanding and knowledge of the game (which then with experince is combined to skill) to get that 52% winrate.

 

So winrate is is a good messurement - as long as we talk about SOLO winrate - for skill. However in this example the guy with 52% winrate would be twice or even 3 times as skilled as the player with 48% winrate.

 

Thats all kinda hard to explain. However I hope I got the thing I wanted to point out across.



W0Z #6 Posted 13 August 2017 - 11:59 AM

    Warrant Officer

  • Player
  • 50255 battles
  • 802
  • Member since:
    01-30-2011

If it's that simple why are you not winning those four extra games?

 

The average W/R is 49% win, 49% lose and 2% draws.



DracheimFlug #7 Posted 13 August 2017 - 12:02 PM

    Colonel

  • Player
  • 8925 battles
  • 3,720
  • Member since:
    11-13-2014

View PostLil_Dimitry, on 13 August 2017 - 11:32 AM, said:

 The point he's trying to make is that he's not bad at the game, just like every other under 50%er out there that thinks it's not his fault for being bad

 

The point is that there is a difference between "worse" and "horrible" and a similar difference between "better" and "amazing."

AvengerOrion #8 Posted 13 August 2017 - 12:02 PM

    Warrant Officer

  • Player
  • 25317 battles
  • 991
  • Member since:
    12-21-2013

View PostJunglist_, on 13 August 2017 - 11:39 AM, said:

Reading these topics almost makes you wish for a nuclear winter

 

Alpha, beta and gamma rays don't care about IQ when they kill though.

Nishi_Kinuyo #9 Posted 13 August 2017 - 12:04 PM

    Lieutenant Сolonel

  • Player
  • 6977 battles
  • 3,469
  • [GUP] GUP
  • Member since:
    05-28-2011

:facepalm:

OP's comment only makes sense at small sample sizes.

Even at my modest battle count we're talking about almost 200 matches difference between 49% and 52%.

More if you include the 1% draws that you'll typically get.

At OP"s battlecount, it increases to almost 550 battle difference.

 

Which is why recents are both more important, and why it can be hard for some accounts to change their global WRs; for OP to increase his WR by even 1% will require him to win almost 200 consecutive matches. (Regardless of current WR.)

 

And that's only going by global winrates.

If you'd check most accounts with above 50% winrates, you'll probably see that their recent winrates are actually quite a few percent above their global; a 2,43% difference in my case, with my most-played tank being higher than  my global winrate as well by more thn 2%. (Trust me, that wasn't always the case!)

Even my second-most played tank has seen a WR increase of 1.5-2% recently, from 44% up to 46%.



AvengerOrion #10 Posted 13 August 2017 - 12:05 PM

    Warrant Officer

  • Player
  • 25317 battles
  • 991
  • Member since:
    12-21-2013

View PostDracheimFlug, on 13 August 2017 - 12:02 PM, said:

 

The point is that there is a difference between "worse" and "horrible" and a similar difference between "better" and "amazing."

 

In his PoV though there is only an accidental 10 games difference between amazing and horrible.

Edited by AvengerOrion, 13 August 2017 - 12:06 PM.


DracheimFlug #11 Posted 13 August 2017 - 12:18 PM

    Colonel

  • Player
  • 8925 battles
  • 3,720
  • Member since:
    11-13-2014

View PostNishi_Kinuyo, on 13 August 2017 - 12:04 PM, said:

:facepalm:

OP's comment only makes sense at small sample sizes.

Even at my modest battle count we're talking about almost 200 matches difference between 49% and 52%.

More if you include the 1% draws that you'll typically get.

At OP"s battlecount, it increases to almost 550 battle difference.

 

Which is why recents are both more important, and why it can be hard for some accounts to change their global WRs; for OP to increase his WR by even 1% will require him to win almost 200 consecutive matches. (Regardless of current WR.)

 

And that's only going by global winrates.

If you'd check most accounts with above 50% winrates, you'll probably see that their recent winrates are actually quite a few percent above their global; a 2,43% difference in my case, with my most-played tank being higher than  my global winrate as well by more thn 2%. (Trust me, that wasn't always the case!)

Even my second-most played tank has seen a WR increase of 1.5-2% recently, from 44% up to 46%.

 

3% is still only 3% regardless of how large a sample size.

 

View PostAvengerOrion, on 13 August 2017 - 12:05 PM, said:

 

In his PoV though there is only an accidental 10 games difference between amazing and horrible.

 

He did not say anything about accidental, merely that 3% is only 3%, which objectively is a small difference and not an "OMG we are doomed because they are the worst player EVAH!" difference.



AvengerOrion #12 Posted 13 August 2017 - 12:26 PM

    Warrant Officer

  • Player
  • 25317 battles
  • 991
  • Member since:
    12-21-2013

View PostDracheimFlug, on 13 August 2017 - 12:18 PM, said:

He did not say anything about accidental, merely that 3% is only 3%, which objectively is a small difference and not an "OMG we are doomed because they are the worst player EVAH!" difference.

 

The way he worded it makes me suspect he does mean it that way.

Edited by AvengerOrion, 13 August 2017 - 12:26 PM.


roachex #13 Posted 13 August 2017 - 12:28 PM

    Warrant Officer

  • Player
  • 20261 battles
  • 607
  • [RIGA] RIGA
  • Member since:
    06-01-2012
It is easy to become 48%, but significant harder to crawl out and get 52%.

And to get 15k games out of 48%, you actually need 55-60% recent with few thousand games.

If that is that easy and not important, then stay stable at 52%+ zone. It does not mean these players cant have 0 dmg games, but it is more likely that they will contribute way more then 48%.

Spurtung #14 Posted 13 August 2017 - 12:32 PM

    Major General

  • Player
  • 60081 battles
  • 5,474
  • [GW-UP] GW-UP
  • Member since:
    07-05-2013

View Postroachex, on 13 August 2017 - 01:28 PM, said:

And to get 15k games out of 48%, you actually need 55-60% recent with few thousand games.

And here lies the crux of it, to climb out of 'horrible' 48% one must be consistently 'amazing' 58%.

The difference then is not only 3 games out of 100.



DracheimFlug #15 Posted 13 August 2017 - 12:39 PM

    Colonel

  • Player
  • 8925 battles
  • 3,720
  • Member since:
    11-13-2014

View Postroachex, on 13 August 2017 - 12:28 PM, said:

It is easy to become 48%, but significant harder to crawl out and get 52%.

And to get 15k games out of 48%, you actually need 55-60% recent with few thousand games.

If that is that easy and not important, then stay stable at 52%+ zone. It does not mean these players cant have 0 dmg games, but it is more likely that they will contribute way more then 48%.

 

OP talks about 49% to 52% (3% difference) and now it is 48% to 52% (4% difference)

 

View PostSpurtung, on 13 August 2017 - 12:32 PM, said:

And here lies the crux of it, to climb out of 'horrible' 48% one must be consistently 'amazing' 58%.

The difference then is not only 3 games out of 100.

 

And suddenly it is 48% to 58%, a 10 % difference (which is indeed a significant difference, but not the range the OP was talking about)



ApocalypseSquad #16 Posted 13 August 2017 - 12:40 PM

    Lieutenant

  • Player
  • 26140 battles
  • 1,976
  • Member since:
    07-31-2011

Win rate is not everything.  

 

It's the only thing...



jabster #17 Posted 13 August 2017 - 12:41 PM

    Field Marshal

  • Beta Tester
  • 12516 battles
  • 21,682
  • [WSAT] WSAT
  • Member since:
    12-30-2010

View PostWunderWurst, on 13 August 2017 - 10:54 AM, said:

 

So winrate is is a good messurement - as long as we talk about SOLO winrate - for skill. However in this example the guy with 52% winrate would be twice or even 3 times as skilled as the player with 48% winrate.

 

I think someone over at WoT Labs tried to quantify the difference between players skill using damage and I beleive the rough answer was a 60% player contributes roughly twice as much as a 50% player. That ties in with a simplified model I played around with years ago where having twice the 'skill' as an average player meant you won 60% of your games.

Spurtung #18 Posted 13 August 2017 - 12:47 PM

    Major General

  • Player
  • 60081 battles
  • 5,474
  • [GW-UP] GW-UP
  • Member since:
    07-05-2013

View PostDracheimFlug, on 13 August 2017 - 01:39 PM, said:

View Postroachex, on 13 August 2017 - 12:28 PM, said:

It is easy to become 48%, but significant harder to crawl out and get 52%.

And to get 15k games out of 48%, you actually need 55-60% recent with few thousand games.

If that is that easy and not important, then stay stable at 52%+ zone. It does not mean these players cant have 0 dmg games, but it is more likely that they will contribute way more then 48%.

 

OP talks about 49% to 52% (3% difference) and now it is 48% to 52% (4% difference)

 

View PostSpurtung, on 13 August 2017 - 12:32 PM, said:

And here lies the crux of it, to climb out of 'horrible' 48% one must be consistently 'amazing' 58%.

The difference then is not only 3 games out of 100.

 

And suddenly it is 48% to 58%, a 10 % difference (which is indeed a significant difference, but not the range the OP was talking about)

I thought you knew numbers.

But feel free to explain how a 48% or 49% becomes a 52% without ever going over 52%. I'm all ears.


Edited by Spurtung, 13 August 2017 - 12:48 PM.


Long_Range_Sniper #19 Posted 13 August 2017 - 12:48 PM

    Lieutenant General

  • Player
  • 29440 battles
  • 7,512
  • [-MM] -MM
  • Member since:
    04-04-2011

View PostGangerr, on 13 August 2017 - 10:29 AM, said:

I don't believe you can really take WR ratio as a base for player skill like some people believe it is.  It's information to some people but its only a small part of a players stats.  Unless my maths are really wrong here, a win/lose ratio of 48% like me, means I win 48 games out of every 100 (if im not mistaken, give or take 1 for a draw) but someone who has a W/R of 52 only wins 4 games more than myself.   As a heavy you expect a player to do more damage, as a light scout you expect them to spot more tanks and receive spotting damage...  Correct me if im wrong.

 

Firstly you have to consider the upper and lower limits of the WR stats population that you are in, which is also skewed towards the lower end. This graph shows that distribution for the EU server.

 

Untitled.jpg

 

So when you think "it's just four games difference" to go from 48 to 52 games out of a hundred you're not considering the fact that the very best players only win 6/10, and a computer bot can win 4/10. The actual space you're working in is the 20% between 40 and 60 percent. 

 

Therefore to go from 48 to 52 games out of a hundred, you've got to win 4 games more out of the 20 that are available for you to work with, because the game will put a ceiling on how many you can win (there are statistical exceptions [looking at you StatPadder] but in general most players will have a ceiling of around 60% as you can see from the stats). These floors and ceilings are due to team composition, numbers, maps and RNG.

 

Four games from twenty then becomes a 20% increase in your normal playing level. To achieve that consistently would show that you have genuinely improved as that is a significant improvement.

 

Which is why winrate is a reasonable indication of skill because no matter what you do such as yolo, camp bush, active spot, passive spot, push in heavies, or whatever tactic you choose then in the long run any team with you on it comes out on top. 



Jigabachi #20 Posted 13 August 2017 - 01:03 PM

    Field Marshal

  • Player
  • 17746 battles
  • 17,623
  • [T-D-U] T-D-U
  • Member since:
    08-12-2011

Your math is correct, but you played more than 100 matches, right? Those four matches difference will turn into hundreds or even thousands of matches that you failed to contribute to.

 

Besides that, try to put that WR in perspective:

The difference of 4% between 48%WR and 52%WR looks very small, but you have to consider that we're not talking about a range of 0%WR-100%WR here, but about 46%WR being the minimum and 58%WR being the maximum: BOTTING/BRAINDEAD <--- 46%WR - 58%WR ---> STATSPADDING

The limits I chose are debatable of course, I just want to point out that the range of normal winrates we're talking about here is rather small. In this example it's 12. And 4 out of 12 is already one third, meaning that a 52% player is a lot better than a 48% player.

 

edit: Had this page idle for too long. Other people already did a better job explaining it...


Edited by Jigabachi, 13 August 2017 - 01:07 PM.





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users