Jump to content


How often do you high/low roll?(Experiment)


  • Please log in to reply
32 replies to this topic

macedoniawarrior #1 Posted 26 October 2017 - 01:07 PM

    Lance-corporal

  • Player
  • 13732 battles
  • 85
  • Member since:
    01-07-2014

I decided to conduct an experiment because i think that the +/- 25% difference from the average roll is a bit too much. I recorded my shots from a dozen of battles, however some shots are excluded like finishing shots: For example if my average damage was 300 and i killed an enemy at 250hp, i would not record the shot. If i rolled for let's say 240, then i would record the shot. Splash damage from HE is also excluded.

 

I intentionally kept some battles where i only managed to land very few shots that met these requirements mentioned above, in order to show how deceiving the figures can be if you do not have enough values. So in order to calculate some numbers I only used the damage rolls from battles where I had at least 5 shots recorded.

 

However take these figures with a pinch of salt, as the experiment is random, i could potentially min-roll 5 times in one game so these numbers i have may vary if you try it yourself however it seems that in the end you end up in the area around your average damage.

 

Its still annoying missing a kill because you low rolled nonetheless :facepalm:

And yes, i still think that the +/- 25% deviation is too much :justwait:

 

What do you think of it?

Attached Files

  • Attached File   Low-High Rolls.xlsx   10.25K

Edited by macedoniawarrior, 26 October 2017 - 01:08 PM.


_EXODUZ_ #2 Posted 26 October 2017 - 01:09 PM

    Lieutenant

  • Player
  • 34165 battles
  • 1,935
  • [RGT] RGT
  • Member since:
    11-05-2014
What do I think? I don't care about RNG and other issues. For me this game is about dynamics, positioning, etc.

macedoniawarrior #3 Posted 26 October 2017 - 01:12 PM

    Lance-corporal

  • Player
  • 13732 battles
  • 85
  • Member since:
    01-07-2014

View Post_EXODUZ_, on 26 October 2017 - 12:09 PM, said:

What do I think? I don't care about RNG and other issues. For me this game is about dynamics, positioning, etc.

 

but reducing the RNG would, at least in my opinion make the game more about dynamics positioning etc., as luck would be less of a factor..

_EXODUZ_ #4 Posted 26 October 2017 - 01:14 PM

    Lieutenant

  • Player
  • 34165 battles
  • 1,935
  • [RGT] RGT
  • Member since:
    11-05-2014

View Postmacedoniawarrior, on 26 October 2017 - 01:12 PM, said:

 

but reducing the RNG would, at least in my opinion make the game more about dynamics positioning etc., as luck would be less of a factor..

 

I've learned to embrace the RNG factor in life. It's everywhere, it adds some spice. :trollface:

CrySpy #5 Posted 26 October 2017 - 01:17 PM

    Second Lieutenant

  • Player
  • 21754 battles
  • 1,001
  • [LEWD] LEWD
  • Member since:
    04-07-2011
You shouldn't keep any of the finishing shots. You can only know if you lowroll(below a certain threshold), but if you roll high enough you have no idea whether oyu got average, high rolled or even lowrolled, but above the threshold. This will introduce a bias in your data.

macedoniawarrior #6 Posted 26 October 2017 - 01:20 PM

    Lance-corporal

  • Player
  • 13732 battles
  • 85
  • Member since:
    01-07-2014

View PostCrySpy, on 26 October 2017 - 12:17 PM, said:

You shouldn't keep any of the finishing shots. You can only know if you lowroll(below a certain threshold), but if you roll high enough you have no idea whether oyu got average, high rolled or even lowrolled, but above the threshold. This will introduce a bias in your data.

 

I did not keep any finishing shots, "however some shots are excluded like finishing shots: For example if my average damage was 300 and i killed an enemy at 250hp, i would not record the shot." 

CrySpy #7 Posted 26 October 2017 - 01:21 PM

    Second Lieutenant

  • Player
  • 21754 battles
  • 1,001
  • [LEWD] LEWD
  • Member since:
    04-07-2011

View Postmacedoniawarrior, on 26 October 2017 - 02:07 PM, said:

. If i rolled for let's say 240, then i would record the

 

Then you should remove this part

macedoniawarrior #8 Posted 26 October 2017 - 01:23 PM

    Lance-corporal

  • Player
  • 13732 battles
  • 85
  • Member since:
    01-07-2014

View PostCrySpy, on 26 October 2017 - 12:21 PM, said:

 

Then you should remove this part

 

Why should i not keep a failed finishing shot? It's a low roll isn't it?

CmdRatScabies #9 Posted 26 October 2017 - 01:26 PM

    Colonel

  • Player
  • 35172 battles
  • 3,538
  • Member since:
    10-12-2015

View Postmacedoniawarrior, on 26 October 2017 - 01:23 PM, said:

 

Why should i not keep a failed finishing shot? It's a low roll isn't it?

 

Keep all the shots that do not kill - i.e. those that aren't limited by the target's HP..

macedoniawarrior #10 Posted 26 October 2017 - 01:28 PM

    Lance-corporal

  • Player
  • 13732 battles
  • 85
  • Member since:
    01-07-2014

View PostCmdRatScabies, on 26 October 2017 - 12:26 PM, said:

 

Keep all the shots that do not kill - i.e. those that aren't limited by the target's HP..

 

this is exactly what i did 

malachi6 #11 Posted 26 October 2017 - 01:31 PM

    Lieutenant Сolonel

  • Player
  • 48582 battles
  • 3,083
  • Member since:
    04-14-2011

View Postmacedoniawarrior, on 26 October 2017 - 01:23 PM, said:

 

Why should i not keep a failed finishing shot? It's a low roll isn't it?

 

But you may be rolling high on other kill shots and excluding them. Skewing your data.  You cannot include any kill shots. 

 

You would be much better using the training room and using the same tanks for a series of tests.  To have any degree of veracity you need to remove as many variables as possible.



macedoniawarrior #12 Posted 26 October 2017 - 01:33 PM

    Lance-corporal

  • Player
  • 13732 battles
  • 85
  • Member since:
    01-07-2014

View Postmalachi6, on 26 October 2017 - 12:31 PM, said:

 

But you may be rolling high on other kill shots and excluding them. Skewing your data.  You cannot include any kill shots. 

 

You would be much better using the training room and using the same tanks for a series of tests.  To have any degree of veracity you need to remove as many variables as possible.

 

For the 2nd time, i did not include ANY kill shot

jabster #13 Posted 26 October 2017 - 01:34 PM

    Field Marshal

  • Beta Tester
  • 12516 battles
  • 21,645
  • [WSAT] WSAT
  • Member since:
    12-30-2010

View Postmacedoniawarrior, on 26 October 2017 - 12:28 PM, said:

 

this is exactly what i did 

 

What’s being said is that you should exclude all shots that could potential kill a tank not just those that do. So for example, if you do on average 100 damage all shots on tanks up to 125hp should be excluded. If you don’t do that you bias your results to lower rolls. Saying that good job on actually collecting data.

 

To make it more obvious, let’s say you shot at a tank with 100hp with a 100hp gun 100 times. You’d only record those shots under 100.


Edited by jabster, 26 October 2017 - 01:43 PM.


macedoniawarrior #14 Posted 26 October 2017 - 01:37 PM

    Lance-corporal

  • Player
  • 13732 battles
  • 85
  • Member since:
    01-07-2014

View Postjabster, on 26 October 2017 - 12:34 PM, said:

 

What’s being said is that you should exclude all shots that could potential kill a tank not just those that do. So for example, if you do on average 100 damage all shots on tanks up to 125hp should be excluded. If you don’t do that you bias your results to lower rolls. Says no all that good job on actually collecting data.  

 

so you mean that if i shot a guy on 200 hp with my gun that has 240 alpha but failed to kill him, that the shot shouldnt be included at all?(from what i remember i did not have any situations like that on the games that i collected the data from)

CmdRatScabies #15 Posted 26 October 2017 - 01:43 PM

    Colonel

  • Player
  • 35172 battles
  • 3,538
  • Member since:
    10-12-2015

View Postmacedoniawarrior, on 26 October 2017 - 01:37 PM, said:

 

so you mean that if i shot a guy on 200 hp with my gun that has 240 alpha but failed to kill him, that the shot shouldnt be included at all?(from what i remember i did not have any situations like that on the games that i collected the data from)

 

I think it should be included - RNG is based on your gun alpha not the target's HP unless you kill him.

__H3H3__ #16 Posted 26 October 2017 - 01:45 PM

    Warrant Officer

  • Player
  • 29972 battles
  • 577
  • [OXIDE] OXIDE
  • Member since:
    12-09-2013

View Postmacedoniawarrior, on 26 October 2017 - 12:37 PM, said:

 

so you mean that if i shot a guy on 200 hp with my gun that has 240 alpha but failed to kill him, that the shot shouldnt be included at all?(from what i remember i did not have any situations like that on the games that i collected the data from)

 

NO he doesnt mean that. Oke look. Youre alpha is 240 and the enemy is on 260 HP. But you kill him. So its was a high roll. But you wont count it because its a killing shot. And that will screw up youre data.

jabster #17 Posted 26 October 2017 - 01:45 PM

    Field Marshal

  • Beta Tester
  • 12516 battles
  • 21,645
  • [WSAT] WSAT
  • Member since:
    12-30-2010

View Postmacedoniawarrior, on 26 October 2017 - 12:37 PM, said:

 

so you mean that if i shot a guy on 200 hp with my gun that has 240 alpha but failed to kill him, that the shot shouldnt be included at all?(from what i remember i did not have any situations like that on the games that i collected the data from)

 

Basically yes as otherwise you’re biasing your data to low rolls. I edited my post above to include an extreme example.

jabster #18 Posted 26 October 2017 - 01:46 PM

    Field Marshal

  • Beta Tester
  • 12516 battles
  • 21,645
  • [WSAT] WSAT
  • Member since:
    12-30-2010

View PostCmdRatScabies, on 26 October 2017 - 12:43 PM, said:

 

I think it should be included - RNG is based on your gun alpha not the target's HP unless you kill him.

 

No it shouldn’t as you’re only going to include it if you got a low roll.

CrySpy #19 Posted 26 October 2017 - 01:46 PM

    Second Lieutenant

  • Player
  • 21754 battles
  • 1,001
  • [LEWD] LEWD
  • Member since:
    04-07-2011

View Postmacedoniawarrior, on 26 October 2017 - 02:23 PM, said:

 

Why should i not keep a failed finishing shot? It's a low roll isn't it?

 

Yes iti s a lowroll, but you have no way of evaluating that a highroll may occur. Therefore you should ignore any shots where the target has less HP than the average roll of your gun.

 

An example.

 

250 HP enemy and 320 DMG gun.

You don't kill the enemy -> lowroll
You kill the enemy -> ???             It could have rolled for 260 DMG, it could also have rolled for 350.

 

There are so many unknowns about that one shot, that you have to discard them. If you can't evaluate highrolls and lowrolls on a shot you will introduce a bias in your data if oyu only include one of them. in this case lowrolls. What oyu are infact doing is saying that if it highrolls you don't include it, but if it lowrolls you include it. And in this case it's lowrolling below a certain threshold. This is not a sound way of representing your data at all



macedoniawarrior #20 Posted 26 October 2017 - 01:51 PM

    Lance-corporal

  • Player
  • 13732 battles
  • 85
  • Member since:
    01-07-2014

View Postjabster, on 26 October 2017 - 12:45 PM, said:

 

Basically yes as otherwise you’re biasing your data to low rolls. I edited my post above to include an extreme example.

 

ok thx i understood now.




1 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users