Jump to content


So 3 Marks Are Kinda BS, eh?


  • Please log in to reply
27 replies to this topic

TheRealSerapth #1 Posted Sep 22 2022 - 22:44

    Major

  • Players
  • 14744 battles
  • 2,261
  • [FNWG] FNWG
  • Member since:
    05-18-2016

I don't mean in terms of the accomplishment... I mean in terms of their literal definition.

Here is the verbiage on the 3 mark from the game...

 

I'm floating pretty constantly around the 85-87% percentile marker in terms of 3 marking the tank...  Meanwhile in August I was 150th on the leaderboard.  This month I'm doing even better (although 7 battles from the 30 required).  The oddest part is, I 2 marked the tank months ago and I've steadily improved my performance since then but it still floats around the 2 mark level.

 

Considering the August leaderboard tapped out at 1,000 and that player 1000th is a 56% winrate, 1600DPG player, we can safely assume a few thousand more players played the Borat but didn't make the leaderboard.  So how can you be topping out the leaderboard of active players, but not in the top 5%, especially on such a heavily played tank.  To be top 5% at 150th rank should simply require a total of 3,000 players using the Borat in August.  Considering it's the 2nd most played tank in the entire game (literally), this doesn't exactly seem unlikely.

 

You can see the leaderboard here: https://worldoftanks...Bourrasque/#wot

 

Something is off with these calculations.  It can't be a purely population based calculation. Either that, or the leaderboards are somehow wrong.  It can't be a matter of being top 5% of the population, it's got to be top percentile of games played in the tank on some form of curve.  Meaning that the super unicorns averaging 5K games and playing tons of games, have to be counting multiple times.  You aren't competing against the general population, it seems you are playing against the top performance of the most played tankers in that tank.



Ortinoth #2 Posted Sep 22 2022 - 22:54

    Major

  • Players
  • 29571 battles
  • 2,007
  • [CBRO] CBRO
  • Member since:
    09-10-2011

your not doing enough damage and combined damage to hit above teh 87% percentile, you need to do more damage and assisted damage for 8% more

 



TheRealSerapth #3 Posted Sep 22 2022 - 23:00

    Major

  • Players
  • 14744 battles
  • 2,261
  • [FNWG] FNWG
  • Member since:
    05-18-2016

View PostOrtinoth, on Sep 22 2022 - 22:54, said:

your not doing enough damage and combined damage to hit above teh 87% percentile, you need to do more damage and assisted damage for 8% more

 

 

Oh I get how 3 marks are calculated.

 

What I'm saying is, those targets are absolute BS when contrasted against the leaderboards of people actually playing the tank.

 

Right now the 3 mark requirement of the Borat is 3742 combine damage and assisting damage.

 

Now take a look at the leaderboard from Wargaming:

https://worldoftanks...ot&w_ts=2022-08

 

The 20th place player on the leader for example ( M3SS_V2_LoveULucky ) isn't pulling in those numbers at 3,696 combined.

 

I don't need to be a math genius to calculate 20th out of 1000 > 95%.

 

So how can the 20th best player out of WELL over 1000 players ( and more accurately probably like 5000+ players ), not be hitting the requirements to be "in the top 95% of players".

 

Put simply, mark of excellence verbiage is BS.



Anzhelia #4 Posted Sep 22 2022 - 23:06

    Major

  • Players
  • 12758 battles
  • 2,459
  • [TANAS] TANAS
  • Member since:
    04-29-2012
It's because it works off a floating average. 

How MoEs are gained is by playing extremely consistently, so you can have an average below the requirement. All that matters is that for a string of battles, you play such that your floating average gets above that 95% mark.

Edited by Anzhelia, Sep 22 2022 - 23:07.


__BillyJack__ #5 Posted Sep 22 2022 - 23:09

    Staff sergeant

  • -Players-
  • 4900 battles
  • 455
  • Member since:
    08-24-2021
Does the cheaters stats figure into this BS floating average? 

TheRealSerapth #6 Posted Sep 22 2022 - 23:20

    Major

  • Players
  • 14744 battles
  • 2,261
  • [FNWG] FNWG
  • Member since:
    05-18-2016

View PostAnzhelia, on Sep 22 2022 - 23:06, said:

It's because it works off a floating average. 

How MoEs are gained is by playing extremely consistently, so you can have an average below the requirement. All that matters is that for a string of battles, you play such that your floating average gets above that 95% mark.

Oh I get that, I get how they are calculated, I get what I need to do to hit the target level. (Coincidentally I've played that tank at a 3200WN8 level over the last 120+ games).

 

What I'm saying is, the mark requirements you need to hit, those are clearly BS, as illustrated by the WoT leaderboards vs target numbers. 



Anzhelia #7 Posted Sep 22 2022 - 23:25

    Major

  • Players
  • 12758 battles
  • 2,459
  • [TANAS] TANAS
  • Member since:
    04-29-2012

View PostTheRealSerapth, on Sep 22 2022 - 14:20, said:

Oh I get that, I get how they are calculated, I get what I need to do to hit the target level. (Coincidentally I've played that tank at a 3200WN8 level over the last 120+ games).

 

What I'm saying is, the mark requirements you need to hit, those are clearly BS, as illustrated by the WoT leaderboards vs target numbers. 


No, they aren't. Those leaderboards only show the averages people have. You only need to play consistently high enough to make that mark, maybe 20 battles or so if you're close. Then you can just play at whatever level as you'll still keep it. In other words, there's a lot of people with high floating averages just due to random chance. 

That said, I will agree that MoEs description is 100% misleading.


Edited by Anzhelia, Sep 22 2022 - 23:27.


TheRealSerapth #8 Posted Sep 22 2022 - 23:33

    Major

  • Players
  • 14744 battles
  • 2,261
  • [FNWG] FNWG
  • Member since:
    05-18-2016

View PostAnzhelia, on Sep 22 2022 - 23:25, said:


No, they aren't. Those leaderboards only show the averages people have. You only need to play consistently high enough to make that mark, maybe 20 battles or so if you're close. Then you can just play at whatever level as you'll still keep it. In other words, there's a lot of people with high floating averages just due to random chance. 

That said, I will agree that MoEs description is 100% misleading.

Not really.  Now that leaderboard individuals might spike in and out based on personal performance, but at a macro level, the performance of say the 50th person, should be roughly consistent, regardless to if the actual person occupying that spot changes.

 

You cannot say *I'm in top 5% currently so I should mark*, as as you have said unless you've played a ton of games you can spike on and out of the leaderboard on a good or bad performance.

 

But over enough games (my August + Sept is over 100) its pretty accurate, and as a generalize metric of the position and not the performance of an individual, the leaderboard will show performance.

 

 

What the leaderboard shows is the 3mark requirements, at least on certain tanks, are BS



Kramah313 #9 Posted Sep 22 2022 - 23:39

    Major

  • Players
  • 14398 battles
  • 2,285
  • Member since:
    01-26-2018
Don’t the leaderboards require 30 battles in that tank in like a month for you to show up on them? There could be a lot of good players playing the tanks less than that

Anzhelia #10 Posted Sep 22 2022 - 23:50

    Major

  • Players
  • 12758 battles
  • 2,459
  • [TANAS] TANAS
  • Member since:
    04-29-2012

View PostKramah313, on Sep 22 2022 - 14:39, said:

Don’t the leaderboards require 30 battles in that tank in like a month for you to show up on them? There could be a lot of good players playing the tanks less than that

 

This is something that I didn't consider. 

I'm not even that good of a player and I played my ELC for 43 battles to 3 mark it. It's entirely possible for people to play less than 30 battles each month in that specific vehicle. 



spud_tuber #11 Posted Sep 22 2022 - 23:53

    Major

  • Players
  • 74154 battles
  • 12,478
  • Member since:
    08-26-2013
What is the minimum battle count needed ro get on the leader boards?

Dont most or all of the websites rely on MoE mods "phoning home"?

And do we actually have any idea how the target values are calculated by WG?



SteelRonin #12 Posted Sep 22 2022 - 23:58

    Major

  • Beta Testers
  • 27453 battles
  • 6,239
  • Member since:
    09-13-2010

always has been, one bad game because the RNG and the % goes down.

 

It is 50% being constant and 50% luck



Saiphaec_Grieve #13 Posted Sep 23 2022 - 00:15

    Staff sergeant

  • -Players-
  • 78 battles
  • 381
  • Member since:
    04-19-2022
I'm averaging 3342 wn8 over my last 401 games in the Borat.  I still dropped from 92% to 80% in the last couple of weeks.  

Killertruth186 #14 Posted Sep 23 2022 - 00:20

    Captain

  • Players
  • 22078 battles
  • 1,102
  • [NGS] NGS
  • Member since:
    09-22-2019

View PostTheRealSerapth, on Sep 22 2022 - 15:44, said:

I don't mean in terms of the accomplishment... I mean in terms of their literal definition.

Here is the verbiage on the 3 mark from the game...

 

I'm floating pretty constantly around the 85-87% percentile marker in terms of 3 marking the tank...  Meanwhile in August I was 150th on the leaderboard.  This month I'm doing even better (although 7 battles from the 30 required).  The oddest part is, I 2 marked the tank months ago and I've steadily improved my performance since then but it still floats around the 2 mark level.

 

Considering the August leaderboard tapped out at 1,000 and that player 1000th is a 56% winrate, 1600DPG player, we can safely assume a few thousand more players played the Borat but didn't make the leaderboard.  So how can you be topping out the leaderboard of active players, but not in the top 5%, especially on such a heavily played tank.  To be top 5% at 150th rank should simply require a total of 3,000 players using the Borat in August.  Considering it's the 2nd most played tank in the entire game (literally), this doesn't exactly seem unlikely.

 

You can see the leaderboard here: https://worldoftanks...Bourrasque/#wot

 

Something is off with these calculations.  It can't be a purely population based calculation. Either that, or the leaderboards are somehow wrong.  It can't be a matter of being top 5% of the population, it's got to be top percentile of games played in the tank on some form of curve.  Meaning that the super unicorns averaging 5K games and playing tons of games, have to be counting multiple times.  You aren't competing against the general population, it seems you are playing against the top performance of the most played tankers in that tank.

I mean, there's no bigger gap than between 1 mark and 2 mark.



Yuri_Doujinshi #15 Posted Sep 23 2022 - 00:21

    Major

  • Players
  • 27449 battles
  • 3,417
  • [CZ4R] CZ4R
  • Member since:
    05-07-2012

View PostTheRealSerapth, on Sep 22 2022 - 22:33, said:

Not really.  Now that leaderboard individuals might spike in and out based on personal performance, but at a macro level, the performance of say the 50th person, should be roughly consistent, regardless to if the actual person occupying that spot changes.

 

You cannot say *I'm in top 5% currently so I should mark*, as as you have said unless you've played a ton of games you can spike on and out of the leaderboard on a good or bad performance.

 

But over enough games (my August + Sept is over 100) its pretty accurate, and as a generalize metric of the position and not the performance of an individual, the leaderboard will show performance.

 

 

What the leaderboard shows is the 3mark requirements, at least on certain tanks, are BS

What you're missing is that the mark requirements take into account any battles played by any player in the tank, while the leaderboards require 30 games played in the tank over the past month. There are a lot of players who will play less than that in a month, but still perform at an extremely high level in a tank, and the mark requirements will take those battles into account.

 

The other thing to consider is how the mark requirements work. The game takes a lot of shortcuts when calculating this; from what I understand, the game takes the average combined damage+assist of all battles played in the tank, and uses that as the median for a standard distribution, which is then used to find the 65%, 85%, and 95% requirements. This means that all of the values are extrapolated rather than actually comparing against the 95th percentile players, so the mark requirements could be artificially higher or lower based on how players are playing.

 

All that being said, that doesn't mean the marks are "BS" or unachievable. It just so happens that the Bourrasque has one of the highest skill ceilings in the game so good players consistently overperform in it, which makes the mark extremely difficult.



spud_tuber #16 Posted Sep 23 2022 - 01:39

    Major

  • Players
  • 74154 battles
  • 12,478
  • Member since:
    08-26-2013

View PostYuri_Doujinshi, on Sep 22 2022 - 17:21, said:

What you're missing is that the mark requirements take into account any battles played by any player in the tank, while the leaderboards require 30 games played in the tank over the past month. There are a lot of players who will play less than that in a month, but still perform at an extremely high level in a tank, and the mark requirements will take those battles into account.

 

The other thing to consider is how the mark requirements work. The game takes a lot of shortcuts when calculating this; from what I understand, the game takes the average combined damage+assist of all battles played in the tank, and uses that as the median for a standard distribution, which is then used to find the 65%, 85%, and 95% requirements. This means that all of the values are extrapolated rather than actually comparing against the 95th percentile players, so the mark requirements could be artificially higher or lower based on how players are playing.

 

All that being said, that doesn't mean the marks are "BS" or unachievable. It just so happens that the Bourrasque has one of the highest skill ceilings in the game so good players consistently overperform in it, which makes the mark extremely difficult.

To the bolded:

 

Considering that as of this post, the AMX 30B needs 2165 to 1 mark and 3972 to 3 mark, while the cent AX needs 2172(7 higher) to 1 mark and 3922(50 lower) to 2 mark, they aren't just calculating a single average and extrapolating from that.  Either they're using an entirely different method, or they're also calculating some form of standard deviation as well.

 

I'm sure there is some form of shortcut being taken, though.  Not like your personal performance is actually the average of your last 100 battles, after all, and for good reason of storage and processing resources used by the server.

 

Edit:  I should be clear.  My post is not intended to discredit the main point of Yuri's post, and is instead a moderate nit-pick about the exact shortcuts that may be in use while agreeing that some form of shortcut is likely in use for the same reasons that a shortcut for the "last 100 battles" is in place. 



Tankeroo #17 Posted Sep 23 2022 - 03:28

    Staff sergeant

  • Players
  • 13655 battles
  • 353
  • Member since:
    03-05-2011

View PostOrtinoth, on Sep 22 2022 - 16:54, said:

your not doing enough damage and combined damage to hit above teh 87% percentile, you need to do more damage and assisted damage for 8% more

 

yes, you need to buy more damage



Asyranok #18 Posted Sep 23 2022 - 03:41

    Major

  • Beta Testers
  • 34707 battles
  • 2,425
  • [BOWSR] BOWSR
  • Member since:
    12-13-2010

View PostTheRealSerapth, on Sep 22 2022 - 23:00, said:

What I'm saying is, those targets are absolute BS when contrasted against the leaderboards of people actually playing the tank.


The leaderboards definitely confused me. In my journey to 3 mark the T54E1, I saw that I was top 3 in the NA server every single day, yet I was at 90% on the mark for a while. It didn't add up for me. Frankly it still doesn't. How could I be top 3 on the server, often #1, and not get the 3 mark for so long...? I sort of stopped looking at the leaderboards for tanks after that. Seems meaningless.


Edited by Asyranok, Sep 23 2022 - 03:41.


Lukas_FishGang4Life #19 Posted Sep 23 2022 - 03:41

    Sergeant

  • Players
  • 18915 battles
  • 204
  • [YOUJO] YOUJO
  • Member since:
    10-30-2014
https://www.13discip...w-moe-work.html

TheJeep1 #20 Posted Sep 23 2022 - 05:18

    Major

  • Players
  • 78290 battles
  • 4,908
  • [OVERC] OVERC
  • Member since:
    06-01-2013
I'm sick of getting a tank within 1 or 2 percent then having it fall back down to 80%.   So, I am scheming.  Mien scheme is to get 5 or 6 tanks to 90% then at least one will make it......I hope.




1 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users