Jump to content


My expectation scale based on a player's wn8/win%

XVM

  • Please log in to reply
101 replies to this topic

DiSaBor #21 Posted 14 May 2020 - 12:58 PM

    Corporal

  • Player
  • 22590 battles
  • 167
  • Member since:
    11-05-2013

View PostPrivate_Miros, on 14 May 2020 - 12:46 PM, said:

 

1vs1 is a skill. In a 1vs1.

 

Skill in randoms is adapting to what goes on around you and forcing as best as possible for the enemy to only engage you on your terms (basically avoid a 1vs1 on equal terms). That means that good players consistently do better than you. They still have losses, but do better there as well. And early damage and good decisions prevent losses that a worse player would have had. And those are the skills that matter, since they make your teams win more than average.

 

Short term battle results mean little, it's the consistency over battles that shows skill. And that consistency means they generally can be relied on in a battle, though mistakes and bad decisions are human and bad luck in an individual battle is perfectly possible. They just make less mistakes and bad decisions than others.

Exactly.

I take my self as exemple.

My wot-life profile shows dark green for my 1776 wn8, but light green for my 52.78% win.

This means percentile-wise, I am better at dealing damage than winning games.

And I am actively trying to improve my gameplay to win more games, rather than passively waiting for things to happen.

That's exactly why I put "avoid 1v1" when I see a player is green or +.



Venom7000 #22 Posted 14 May 2020 - 01:01 PM

    Warrant Officer

  • Player
  • 8649 battles
  • 841
  • Member since:
    03-10-2018

View PostPrivate_Miros, on 14 May 2020 - 12:54 PM, said:

 

Somewhat true.

 

But high(ish) PR and low(ish) WR means the player is bad at adapting early battle, and only good at seeking a defensive position when it hits the fan.

 

High(ish) WR for PR means more skilled in early damage and allowing early damage to team mates.

 

Hence, higher (solo) WR over a statistically significantly number of battles with consistent performance is always a better player.

Agreed.

However,  on the part about high % and low ish pl rating.

I mean, hypothetically. I can hide for 100 games behind  a bush. And pray that my teammates can win. Yes i got higher win% but I was useless in that game.

 

If we ran a mm simulation where in 100games MM puts you in a team of unicums. And you never leave the spawn but win 96 of those games. You are going to have high win rating and non existent pl rating.

 

Its possible to have high win rating and low pl score. Based on pure luck of the mm. Its possible.

 

Its nearly impossible to have 6000player rating and rock 47% win rate. :D

But I agree. Both indicators are good to have.

I just prefer to see what you did in battle via pl rating. And not judge you based on how MM blessed you (or d-cked you :teethhappy:)



Private_Miros #23 Posted 14 May 2020 - 01:09 PM

    Field Marshal

  • Player
  • 28660 battles
  • 11,224
  • [EMU87] EMU87
  • Member since:
    07-09-2011

View PostVenom7000, on 14 May 2020 - 12:01 PM, said:

Agreed.

However,  on the part about high % and low ish pl rating.

I mean, hypothetically. I can hide for 100 games behind  a bush. And pray that my teammates can win. Yes i got higher win% but I was useless in that game.

 

If we ran a mm simulation where in 100games MM puts you in a team of unicums. And you never leave the spawn but win 96 of those games. You are going to have high win rating and non existent pl rating.

 

Its possible to have high win rating and low pl score. Based on pure luck of the mm. Its possible.

 

Its nearly impossible to have 6000player rating and rock 47% win rate. :D

But I agree. Both indicators are good to have.

I just prefer to see what you did in battle via pl rating. And not judge you based on how MM blessed you (or d-cked you :teethhappy:)

 

MM blesses everyone alike once you look at a few hundred battles. Luck evens out pretty quickly at 100 battles already to reliably say the difference between someone good and someone bad, certainly in combination with their other stats.

So you're not going to get 100 battles more with unicums than other people (unless you are the unicum, or constantly platoon with unicums).

 

If you have high WR and low PL score, and it's solo WR, that player is making consistent great decisions that help his team.

 

You currently have 52% for 1282WN8. That's not bad. But if you improve your decision making, you will see that WR go up, with exactly the same luck as you do now. Yes, you'll obviously still have defeats, but you'll be consistently better there too (I currently have 1666 WN8 in my defeats alone).



Dramya #24 Posted 14 May 2020 - 01:10 PM

    Warrant Officer

  • Player
  • 17465 battles
  • 566
  • [_PZF_] _PZF_
  • Member since:
    03-30-2014

View PostVenom7000, on 14 May 2020 - 01:15 PM, said:

Win% is somewhat irrelevant. Why?

Because it only tells you how many times the won a match.

 

And unless we have 1v1 matches. Its not exactly showing his skill.

 

I know people who have 52% win rate but are dumber than bots and have sub 3000 pl rating.

 

If you have luck, MM places you on the roflstomp team. You will have higher win%, even though you sat in base with zero dmg.

 

Even a unicum can lose, because he is fighting 1v12 and his team are donkeys.

 

Pl rating is where its at

 

It's not irrelevant, it tells you that in a certain amount of the games he has played he was (part of) the deciding factor on winning said games. It doesn't mean he has done so singlehandedly, but it does mean he can pull his weight. The more consistently he can do this (eg. higher winrate) the more you can expect him to do well in the game you are playing.

 

If winrate was irrelevant we would all have 49% winrate as a statistical average, and it would mean we have absolutely no effect on the outcome. And this is decidedly not true.

 

But of course we can't win them all ...



DiSaBor #25 Posted 14 May 2020 - 01:13 PM

    Corporal

  • Player
  • 22590 battles
  • 167
  • Member since:
    11-05-2013

View PostBlack_Rojer, on 14 May 2020 - 12:48 PM, said:

This is really funny "unicum", I have seen a lot of "unicum" players, who camp all the time near the base, did nothing and at the ans battle lost. is it skill?

Funny fact : when I see this I predicted you had 47% win or so. Actually you have 48%, which is better but in the same range.

 

Usually a bad player cannot understand what a good player is doing.

Years ago I was surprised to see unicums crossing kill zones while staying in one piece like a boss. I tried to do the same next time in the same map, but got exploded very, very soon.

 

Lately I had a game in my HWK 12 in the map Airfield. I was playing middle early, did a few cross fire damage to enemy obj 257. Then I managed to surprise an enemy blue player with Skoda T50, putting 2 shots into him while receiving one in return, putting him in one-shot for everyone.

I then flipped my tank in a useless position, but our remaining two TDs managed to win the game.

For this particular game, I consider my self having done key contribution to victory. But after I flipped, someone was laughing at me and saying I was useless green player and was dancing the whole game doing nothing.

In post game stats I found myself having done most damage, despite missing the late part.

I messeged the player of this, who never messeged me back.


Edited by DiSaBor, 14 May 2020 - 01:15 PM.


UserZer00 #26 Posted 14 May 2020 - 01:21 PM

    Second Lieutenant

  • Player
  • 14298 battles
  • 1,080
  • Member since:
    03-05-2013
Since I play vanilla it doesn't matter, but....

Until they break the WR into WR playing solo, WR +1 and WR +2 (platooning with one or 2 other players) I don't care what anyone's WR is.

More factors would be
- WR while grinding
- WR with fully loaded, high skill crews
- WR with OPed tanks
- WR with crap tanks
Etc. Etc.

All in all, too many things can influence the statistics so you are better off playing like the enemy team is full of unicoms and your team full of new players.  

Play based on what you see on the map.

This is all from a just about average occasional player with no interest in grinding, just occasional stress relief.

saxsan4 #27 Posted 14 May 2020 - 01:22 PM

    Colonel

  • Player
  • 22878 battles
  • 3,893
  • [-WR--] -WR--
  • Member since:
    11-08-2011

View Postroboslash, on 14 May 2020 - 11:54 AM, said:

I simply do not use XVM and make judgements from the displayed stats, as people improve/become worse at the game. Overall stats do not always reflect a current player's ability to play.

 

I simply treat every enemy as a threat, and don't rely on any ally, though that judgement changes throughout the battle, according to what I observe.


This

 

 

yet o my thread many people take issue with the above statement,

 

In truth humans cannot be decried accurately by numbers 



Long_Range_Sniper #28 Posted 14 May 2020 - 01:33 PM

    Field Marshal

  • Player
  • 40767 battles
  • 11,598
  • [RGT] RGT
  • Member since:
    04-04-2011

Just stick with the classics.

 

Stats Question - General Discussion - World of Tanks official forum



Jauhesammutin #29 Posted 14 May 2020 - 01:34 PM

    Second Lieutenant

  • Player
  • 24255 battles
  • 1,195
  • [KANKI] KANKI
  • Member since:
    11-05-2013

View PostPrivate_Miros, on 14 May 2020 - 12:09 PM, said:

 (I currently have 1666 WN8 in my defeats alone).

Where can you see that?



Private_Miros #30 Posted 14 May 2020 - 01:36 PM

    Field Marshal

  • Player
  • 28660 battles
  • 11,224
  • [EMU87] EMU87
  • Member since:
    07-09-2011

View PostJauhesammutin, on 14 May 2020 - 12:34 PM, said:

Where can you see that?

 

You cannot. I take manual notes for my statistics thread.



DiSaBor #31 Posted 14 May 2020 - 01:45 PM

    Corporal

  • Player
  • 22590 battles
  • 167
  • Member since:
    11-05-2013

View PostPrivate_Miros, on 14 May 2020 - 01:36 PM, said:

 

You cannot. I take manual notes for my statistics thread.

You can check http://wn8lab.com/ if you are intersted in indepth analysis of your stats.



dan_dix #32 Posted 14 May 2020 - 01:48 PM

    Staff Sergeant

  • Player
  • 21351 battles
  • 351
  • Member since:
    01-09-2016

This is why stats in game are a bad thing.

 

You know the standard disclaimer "previous performance is not an indication of future performance".

 

You just don't know if the player behind is going to do - any unicum can have a bad game and any tomato can top the chart every once in a while. You can not decide how to play a particular game based on the random occurrence of people in your or the enemy team. You decide how to play based on how the game pans out



Jauhesammutin #33 Posted 14 May 2020 - 01:52 PM

    Second Lieutenant

  • Player
  • 24255 battles
  • 1,195
  • [KANKI] KANKI
  • Member since:
    11-05-2013

View PostDiSaBor, on 14 May 2020 - 12:45 PM, said:

You can check http://wn8lab.com/ if you are intersted in indepth analysis of your stats.

Almost purple



DiSaBor #34 Posted 14 May 2020 - 02:10 PM

    Corporal

  • Player
  • 22590 battles
  • 167
  • Member since:
    11-05-2013

View Postdan_dix, on 14 May 2020 - 01:48 PM, said:

You know the standard disclaimer "previous performance is not an indication of future performance".

 

You just don't know if the player behind is going to do - any unicum can have a bad game and any tomato can top the chart every once in a while. You can not decide how to play a particular game based on the random occurrence of people in your or the enemy team. You decide how to play based on how the game pans out

 

Same applies to almost everything : econocims, finance,... Even physics is about statistics, only are we living in a scale where individual atoms do not matter.

 

Do stats predict individuals' behavior ? No

Do I know every player individually in-game ? No

 

Here is why stats helps. It gives a somewhat reliable scale to start with. And of course you need to adjust your expectation by observing how a player is doing things ingame.

It happend to me to see an orange player doing everything extremely well in a game. I messaged him after the game, and platooned with him several times. He is actually a very skillful player; only is he sharing his account with his farther.



TheDrownedApe #35 Posted 14 May 2020 - 02:16 PM

    Lieutenant General

  • Player
  • 52942 battles
  • 6,785
  • [IDEAL] IDEAL
  • Member since:
    03-27-2013

View PostCmdRatScabies, on 14 May 2020 - 11:30 AM, said:

Exactly - so always keep an eye on him and when his tracks turn ..... run before he abandons you.  :)


When im in that position i ALWAYS click "follow me" to the fellow tank - they rarely do though, then die, then slag me off for abandoning them. it's funny how can don't read or chat until it's babyrage time.

 

But then i often tell myself "danger Will Robinson, run away" but then half of the time i don't and die stupidly :(



Ricox #36 Posted 14 May 2020 - 02:19 PM

    Lance-corporal

  • Beta Tester
  • 696 battles
  • 71
  • Member since:
    07-22-2010

View PostCmdRatScabies, on 14 May 2020 - 11:21 AM, said:

What I learnt early on:

60%+ : Unicum

      As teamate : Extreme attention.  Expect him to use his mobility to run off and leave me to die alone.  Always run before he does.

 

But it's often fair, no? It's not even a unicum thing. I try to help my teammates as much as possible, even by sacrificing my HP pool when I got plenty, but if I see 4 equal-tier heavies turning corner to attack two of us and they're not complete tomatoes who will keep trading one-by-one - you run. You warn your teammate and in nearly all cases he/she will just ignore you, then cry abandonment because their tunnel vision prevents them from noticing the minimap or the chat.

 

You're not making your team any more likely to win by just sacrificing yourself in some "heroic last stand" when you could save your HP and gun for an engagement with better chances, even if the engagement you ran away could have been won, it's still a game of best chances not delusionary expectations.

 

View Postdan_dix, on 14 May 2020 - 12:48 PM, said:

This is why stats in game are a bad thing.

 

You know the standard disclaimer "previous performance is not an indication of future performance".

 

You just don't know if the player behind is going to do - any unicum can have a bad game and any tomato can top the chart every once in a while. You can not decide how to play a particular game based on the random occurrence of people in your or the enemy team. You decide how to play based on how the game pans out

 

People misusing them is a bad thing. Having stats in-game is not a bad thing, it can give you a much more informed approach to the specific battle and make the right calls when it matters. Your team's going to fare much better if you have 2 identical tanks on killshot HP and you pick the unicum not the one who keeps firing at your angled upper plate.


Edited by Ricox, 14 May 2020 - 02:24 PM.


DeepDishPizza #37 Posted 14 May 2020 - 04:07 PM

    Private

  • Player
  • 359 battles
  • 13
  • Member since:
    09-18-2019

View Postsaxsan4, on 14 May 2020 - 01:22 PM, said:

 

In truth humans cannot be decried accurately by numbers 

 

Exactly. Good stats dont make you a good player.



lassekille #38 Posted 14 May 2020 - 04:50 PM

    Corporal

  • Player
  • 17923 battles
  • 115
  • [_V_A_] _V_A_
  • Member since:
    03-23-2014

A bit simplified, but not wrong.

Some have high win rates because the often platoon and seal club at low tiers. Doesn't mean they are good at tier 10.

Some might have high win rate in one specific tank (or class of tank) even if they have an overall low win rate.

Recent win rate is more relevant than overall win rate.



CmdRatScabies #39 Posted 14 May 2020 - 05:26 PM

    Lieutenant General

  • Player
  • 39034 battles
  • 6,277
  • [-MM] -MM
  • Member since:
    10-12-2015

View PostTheDrownedApe, on 14 May 2020 - 02:16 PM, said:


When im in that position i ALWAYS click "follow me" to the fellow tank - they rarely do though, then die, then slag me off for abandoning them. it's funny how can don't read or chat until it's babyrage time.

 

But then i often tell myself "danger Will Robinson, run away" but then half of the time i don't and die stupidly :(

I often feel a twinge of guilt as I run from a flank,    but it passes.  :D

16:30 Added after 3 minute

View PostRicox, on 14 May 2020 - 02:19 PM, said:

But it's often fair, no? It's not even a unicum thing. I try to help my teammates as much as possible, even by sacrificing my HP pool when I got plenty, but if I see 4 equal-tier heavies turning corner to attack two of us and they're not complete tomatoes who will keep trading one-by-one - you run. You warn your teammate and in nearly all cases he/she will just ignore you, then cry abandonment because their tunnel vision prevents them from noticing the minimap or the chat.

 

You're not making your team any more likely to win by just sacrificing yourself in some "heroic last stand" when you could save your HP and gun for an engagement with better chances, even if the engagement you ran away could have been won, it's still a game of best chances not delusionary expectations.

Part of learning the game is not feeling like you have to dig in or stay with your team mates.  Relocating is often the best, but bloody annoying if you're in a slow tank and everyone else buggers off.  :) 



fwhaatpiraat #40 Posted 14 May 2020 - 05:44 PM

    Major

  • Player
  • 63175 battles
  • 2,620
  • [RGT] RGT
  • Member since:
    05-04-2013
Just play vanilla and you will judge players on their moves and not on their history. You can already predict if a player is decent or not by the way he/she drives away from the spawn. Or how they move the turret. :)





Also tagged with XVM

4 user(s) are reading this topic

0 members, 4 guests, 0 anonymous users