Jump to content


This statement made about HE is rather astounding.


  • Please log in to reply
10 replies to this topic

Jaspo #1 Posted Oct 18 2019 - 01:38

    Staff sergeant

  • -Players-
  • 31377 battles
  • 485
  • Member since:
    03-12-2015

"The current algorithm often "doesn't find" obvious weak spots, and HE shells may not cause any damage at all."

 

So what WG just admitted, is that since the game's beginning, HE has actually been broken and not just excessively RNGy, and they've not seen the need to do anything about it until now, and even now they're only doing something about it since they're changing how the other shells work and it just happened to be a parameter that will be changing due to other changes.

 

Hmm.



nuclearguy931 #2 Posted Oct 18 2019 - 01:42

    Major

  • Players
  • 82534 battles
  • 10,986
  • [UMM] UMM
  • Member since:
    08-16-2011
It's probably a poor Russian to English translation, and it wouldn't be the first time we've seen that with this game.

24_inch_pythons #3 Posted Oct 18 2019 - 01:53

    Major

  • Players
  • 30359 battles
  • 5,733
  • [HSOLO] HSOLO
  • Member since:
    03-28-2012

View Postnuclearguy931, on Oct 17 2019 - 18:42, said:

It's probably a poor Russian to English translation, and it wouldn't be the first time we've seen that with this game.

 

 Possibly, yes, but I don't think so. I think it makes perfect sense and is probably legit. Think of all the different armor thicknesses a 1m bubble would encounter on a tank. Everything from tracks which won't cause the tank any damage to possibly hard turret armor and possibly very thin deck armor, just as an generic example. The HE should calculate vs. the deck, which has the weakest armor - provided the deck is within the 1m bubble in the example, but is instead "wasting" itself calculating off of the tracks or the turret.



Siergen #4 Posted Oct 18 2019 - 01:56

    Sergeant

  • -Players-
  • 10239 battles
  • 248
  • [H_FUN] H_FUN
  • Member since:
    07-12-2016

View Post24_inch_pythons, on Oct 18 2019 - 01:53, said:

 

 Possibly, yes, but I don't think so. I think it makes perfect sense and is probably legit. Think of all the different armor thicknesses a 1m bubble would encounter on a tank. Everything from tracks which won't cause the tank any damage to possibly hard turret armor and possibly very thin deck armor, just as an generic example. The HE should calculate vs. the deck (provided it is within the 1m bubble in the example) which has the weakest armor, but is instead "wasting" itself calculating off of the tracks or the turret.

 

Especially since they probably can only allocate a fixed number of clock cycles to search for a weak spot for each HE hit.



24_inch_pythons #5 Posted Oct 18 2019 - 01:58

    Major

  • Players
  • 30359 battles
  • 5,733
  • [HSOLO] HSOLO
  • Member since:
    03-28-2012

View PostSiergen, on Oct 17 2019 - 18:56, said:

 

Especially since they probably can only allocate a fixed number of clock cycles to search for a weak spot for each HE hit.


 Yup



spud_tuber #6 Posted Oct 18 2019 - 06:24

    Major

  • Players
  • 59346 battles
  • 8,864
  • Member since:
    08-26-2013
Yeah.  Saw that.  Personally, I'm not sure that it is so much broken, as a compromise between performance and functionality that WG neglected to tell us about. It was also probably made worse by the HD armor models, which increased the number of modeled segments.

Presumably, the new process has had heavy optimization, and of course computers have gained at least some performance since the original code was written.   Hopefully, this will mean the new method gains functionality with no significant loss of relative performance compared to the old code.  But, even so, you can bet there will be some compromises for performance sake.

golruul #7 Posted Oct 18 2019 - 21:33

    Captain

  • Players
  • 22162 battles
  • 1,742
  • Member since:
    11-05-2011

In the beginning, from what I recall, they used a spherical model from where the round impacted.  So if an HE round exploded on the lower part of the upper glacis plate, it would use the lower armor value of the lower glacis to calculate damage against (since it was the lowest armor thickness in the generated sphere).

 

That was kinda dumb.  They then have since moved to the current conical system, which apparently still has issues.



ShadowDancer27 #8 Posted Oct 18 2019 - 22:29

    Major

  • Players
  • 30872 battles
  • 3,918
  • Member since:
    07-26-2013
The old algorithm might have simply checked perhaps 6 points +/- X,Y,Z from the point of impact for an intersection with tank surface and picked the thinnest one.  The new model may have a more general intersection model.

Flarvin #9 Posted Oct 18 2019 - 22:42

    Major

  • Players
  • 54888 battles
  • 17,053
  • Member since:
    03-29-2013

View Postspud_tuber, on Oct 18 2019 - 00:24, said:

Yeah. Saw that. Personally, I'm not sure that it is so much broken, as a compromise between performance and functionality that WG neglected to tell us about. 


I see it as with how spotting check rates were increased. While older spotting check rate were not broken, but the increase helped functionality.

 

A simple performance verse functionality decision.  
 

Why would WG tell players about every performance verse functionality decision they made? 
 

The software companies I worked for never told customers about those types of decisions. 

 



Flarvin #10 Posted Oct 18 2019 - 22:47

    Major

  • Players
  • 54888 battles
  • 17,053
  • Member since:
    03-29-2013

View PostShadowDancer27, on Oct 18 2019 - 16:29, said:

The old algorithm might have simply checked perhaps 6 points +/- X,Y,Z from the point of impact for an intersection with tank surface and picked the thinnest one.  The new model may have a more general intersection model.


I agree.

 

WG probably has been able add more thorough weak point checking algorithm, without affect performance through various means. 
 

With or without this HE redesign going live, I hope WG at least adds this new non-pen HE damage path algorithm. 



SquishySupreme #11 Posted Oct 18 2019 - 23:22

    First lieutenant

  • -Players-
  • 3707 battles
  • 760
  • Member since:
    10-29-2018

View PostJaspo, on Oct 18 2019 - 01:38, said:

"The current algorithm often "doesn't find" obvious weak spots, and HE shells may not cause any damage at all."

 

So what WG just admitted, is that since the game's beginning, HE has actually been broken and not just excessively RNGy, and they've not seen the need to do anything about it until now, and even now they're only doing something about it since they're changing how the other shells work and it just happened to be a parameter that will be changing due to other changes.

 

Hmm.



This is because in May, they changed the damage falloff from linear to exponential.

Now, for 95% of players, this will go past their heads and they simply won't GET what a drastic change this was.  It was hidden in a ton of boiler plate text and blathering, but was the single most significant change to HE yet.
http://www.learnabou..._log_action.gif


The damage modelling assumes a linear spread.  So if your plate you hit is large enough, the damage spreads to the edge but has fallen off so much by then that it "finds" nothing weak enough to penetrate.  You get 0 damage "hits" .

That graph shows the problem.  If you miss by 5 ft or it is a solid enough pate, it's lost half of its damage almost instantly.  So they are trying to make it do more consistent damage, but the problem is that the effective output of HE if implemented would be such that it takes 8-10 shots to kill even light tanks.  They quickly learn to laugh at you, rush in, take the HE shot, and then kite you with impunity.  


Edited by SquishySupreme, Oct 18 2019 - 23:24.





1 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users