Suggestion: RNG in WG (for WG4?)

Zark
Private First-Class
Posts: 6
Joined: Fri 1 Dec 2017 02:27
Contact:

Re: Suggestion: RNG in WG (for WG4?)

Postby Zark » Sun 15 Apr 2018 17:33

Firstly, have you played Steel Division: Normandy 44?

The reason I'm asking is because that game uses a two dice system much like what you are suggesting.
https://www.reddit.com/r/Steel_Division/comments/6ak9j9/veterancy_and_accuracy_calculations_explained/
I do not care for that system from a game-play stance because it reduces granularity between hit chances.

In addition to that and what other people have mentioned, I believe that such a system would increase the performance load on the game significantly. The current Wargame system calculates a hit simply based on (random(number) * 100 <= hit-chance + modifiers), where as Steel Division's is more like (random(num)*6 + random(num)*6 > 12 - unit accuracy value - modifiers). This may not seem like a big difference, but when you have >100 weapons on the field making hit calculations every second, as often happens in wargame, it starts to become an issue. This is probably part of the reason Steel division is a smaller scaled game. When it was released, the game would suffer performance-related crashes frequently in larger games. It was an achievement to make it to the end of a 10v10 match. This has since been fixed, but Steel Division is still more performance intensive despite being smaller in scale.

Now, I do like how Steel Division has units become more accurate for every shot they take, to simulate adjusting fire, but that could be done with a simple +10% for every shot. As for simulating ERA or active protection, that could be done with the hit calculation similar to how ECM works in game.

Booster
Corporal
Posts: 37
Joined: Fri 13 Apr 2018 13:32
Contact:

Re: Suggestion: RNG in WG (for WG4?)

Postby Booster » Sun 15 Apr 2018 17:50

Zark wrote:Firstly, have you played Steel Division: Normandy 44?

The reason I'm asking is because that game uses a two dice system much like what you are suggesting.
https://www.reddit.com/r/Steel_Division/comments/6ak9j9/veterancy_and_accuracy_calculations_explained/
I do not care for that system from a game-play stance because it reduces granularity between hit chances.

In addition to that and what other people have mentioned, I believe that such a system would increase the performance load on the game significantly. The current Wargame system calculates a hit simply based on (random(number) * 100 <= hit-chance + modifiers), where as Steel Division's is more like (random(num)*6 + random(num)*6 > 12 - unit accuracy value - modifiers). This may not seem like a big difference, but when you have >100 weapons on the field making hit calculations every second, as often happens in wargame, it starts to become an issue. This is probably part of the reason Steel division is a smaller scaled game. When it was released, the game would suffer performance-related crashes frequently in larger games. It was an achievement to make it to the end of a 10v10 match. This has since been fixed, but Steel Division is still more performance intensive despite being smaller in scale.

Now, I do like how Steel Division has units become more accurate for every shot they take, to simulate adjusting fire, but that could be done with a simple +10% for every shot. As for simulating ERA or active protection, that could be done with the hit calculation similar to how ECM works in game.

No, I do not play this game. Do not like snail WW1/2 setting :D

I think we can find a compromise between actual WG and SD:N44 RNG system. We can use WG system for an infantry and for the vehicles with the high rate of fire. The dice system we can use for tanks, missiles, etc...

EDITED:
I don't know how it will work, but it must reduce the number of calculations.
It's rough: we can use an 2dX (Xd2?) RNG 10 times in 1 second for all units at the battlefield. RNG gives us a numbers which will be used for all this units which is shoot in the current one tenth of a second.

User avatar
James-Bond
Master Sergeant
Posts: 198
Joined: Mon 9 Jul 2012 18:08
Contact:

Re: Suggestion: RNG in WG (for WG4?)

Postby James-Bond » Sun 15 Apr 2018 18:20

Booster wrote:
James-Bond wrote:End of the day multiple dice will just return a probability value anyway their is no reason to make it more complex?

2 head coin tosses in a row, 25%
Greater than 6 on an 8 sided dice, 25%
Greater than 75 out of 100, 25%

I'd prefer to deal directly with % probability than complex compound odds.

http://anydice.com/program/535 Here you can see a comparison of 1d12 and 2d6. The probability for 1d12 are equal for each side is 8.33%. The probability for 2d6 for 2 or 12 is 2.78% but for 7 is 16.67%.
The difference between 1d12 and 2d6 is: 6.50 average / 3.45 standard deviation for 1d12 and 7.00 average / 2.42 standard deviation for 2d6.
For the 1d12 we only have 12 variants of roll but for 2d6 we have 36 variants! And we can use even more dices to lower the deviation.

Want bring to mind - we are playing a strategy game which must relay on player's skill but not a random numbers.
The simple example in probabilities is - if we give a monkey an unlimited time the monkey will write War and Peace of Tolstoy someday but this is in theory, but in reality the monkey will never write it. But now in WG we have a monkey which can write War and Peace 100 times in 40 minutes.


Im am aware of probability.
But just seems you're making it unnecessarily complex. End of the day you are still returning odds / %.

There is player skill in this game strategy on how you command the forces.

This dice system still isnt player skill. Though...

Since in game we dont take over tank cannon etc like in a fps shooter.
The combat element of wargame isn't player skill.
Instead it is character skill of the unit. Which has probability values assigned to it.

The monkey has a very slight chance of writing war and peace, not impossible.

Like everyone who enters a lottery has a slim chance of winning. A lot of people who enter lose but there is a winner.

Booster
Corporal
Posts: 37
Joined: Fri 13 Apr 2018 13:32
Contact:

Re: Suggestion: RNG in WG (for WG4?)

Postby Booster » Mon 16 Apr 2018 06:06

James-Bond wrote:Im am aware of probability.
But just seems you're making it unnecessarily complex. End of the day you are still returning odds / %.


Yes, for the first i suggested to use a dice system to determine a percentage of hit chance, but after this I suggest to use it to determine a hit damage but we can still use some low values to determine a miss-shot in both: hit roll an safe roll. There is a difference.
For hit damage - 4d2:
4 dmg - 6.25% - gun jammed
5 dmg - 25.00% - hit
6 dmg - 37.50% - hit
7 dmg - 25.00% - hit
8 dmg - 6.25% - missed shot
For safe roll - 10d2:
10 - 0.10% - something
11 - 0.98% - engine destroyed
12 - 4.39% - damage to modules (random)
13 - 11.72% - reactive armor/active armor protection system (against HEAT) else it taking damage
14 - 20.51% - taking damage
15 - 24.61% - taking damage
16 - 20.51% - taking damage
17 - 11.72% - miss
18 - 4.39% - critical hit
19 - 0.98% - ammo explosion
20 - 0.10% - something
Actually we can use a variants of a safe roll for each type of damage (percentage).
Example:
for hit - 3d2:
3 - 12.50% - something
4 - 37.50% - taking damage
5 - 37.50% - taking damage
6 - 12.50% - critical hit
for critical hit - 3d2:
3 - 12.50% - ammo explosion
4 - 37.50% - something
5 - 37.50% - random: fuel leak/optical failure/track lost...
6 - 12.50% - crew member killed
This is just an example.
Profit - we have a statistically stable result. We have a hit chance of 87.5% and variable damage. Even if a shooter have a low damage (for example - average 4) it still can penetrate 6 armor if it's lucky.
The overall chance to hit will be 90-95% for all units with the differences in hit damage and HP and armor of the units must be reworked because of that.
So far, no thoughts for this yet...

User avatar
James-Bond
Master Sergeant
Posts: 198
Joined: Mon 9 Jul 2012 18:08
Contact:

Re: Suggestion: RNG in WG (for WG4?)

Postby James-Bond » Mon 16 Apr 2018 13:07

Ah right, so there are different layers of probability for different results.
I thought it was just multiple dice for the return of 1 value like accuracy.

Is this mechanism used in games like
World of Tanks or Armored Warfare?

Booster
Corporal
Posts: 37
Joined: Fri 13 Apr 2018 13:32
Contact:

Re: Suggestion: RNG in WG (for WG4?)

Postby Booster » Mon 16 Apr 2018 13:55

James-Bond wrote:Is this mechanism used in games like
World of Tanks or Armored Warfare?

Don't know about World of Tanks or Armored Warfare but It used in tabletop RPG games mostly. WH40k have the similar system.

Booster
Corporal
Posts: 37
Joined: Fri 13 Apr 2018 13:32
Contact:

Re: Suggestion: RNG in WG (for WG4?)

Postby Booster » Tue 17 Apr 2018 00:39

We can determine a critical hits for the inner and the outer modules of a vehicle.
We can use a 2 types of critical hit:
The low-damage guns will be viable against heavy armor because they can crit an outer modules like: optic, trucks, gun barrel, outer machineguns, etc...
If the gun can penetrate the armor it can crit the inner modules of a vehicle: crew, engine, ammunition, fuel tank, etc...

The way of implementation of reactive armor like in Aurora-4X:
XXXXXXXXX
XXHXXXXXX
This is a layers of a side reactive armor and when the gun hits it "H" (randomly) it'll explode.
After explosion "o":
XXoXXXXXX
XoooXXXXX
Then if it'll be hit in the "o" again with HEAT, there's no reactive armor now and the main armor will be penetrated if the dmg is higher then the armor.

Booster
Corporal
Posts: 37
Joined: Fri 13 Apr 2018 13:32
Contact:

Re: Suggestion: RNG in WG (for WG4?)

Postby Booster » Tue 17 Apr 2018 16:14

With this Lua script you can get to know how this dice system would work in actual game for the different number of dices and sides:
Spoiler : :

Code: Select all

--[[ DICE ROLL CHANCES ]]------------------------------------------------

math.randomseed( os.time( ))

-- INFO: Declaration of variables:
local rolls = 100
local dices = 4
local sides = 2
local lower = dices
local upper = dices * sides
local range = upper - lower
local mean  = range / 2 + dices
local pool  = 0
local data  = { }

-- INFO: Definition of functions:
function newData( )
  for i = 0, range do
    data[ tostring( i + lower )] = 0
  end
end

function castRolls( )
  local sum
  local num
  for i = 1, rolls do
    sum = 0
    for j = 1, dices do
      num = math.random( 1, sides )
      sum = sum + num
    end
    data[ tostring( sum )] = data[ tostring( sum )] + 1
  end
end

function getChance( )
  local rate = 0
  local k    = ""
  for i = lower, upper do
    k = data[ tostring( i )]
    rate = 100 * k / rolls
    pool = pool + rate
    io.write( string.format( "\n%7.3d%6.4d%8.3f%s", i, k, rate, " %" ))
    if i == mean then
      io.write( " mean" )
    end
  end
end

-- INFO: Main function:
newData( lower, range, data )
castRolls( rolls, dices, sides, data )
getChance( rolls, lower, upper, pool, data )

io.write( string.format( "\n\n%.3d-%.3d%6.4d%8.3f%s\n%21.f%s%.f", lower, upper, rolls, pool, " %", dices, "d", sides ))

The output:
Untitled 3.jpg
Untitled 3.jpg (7.01 KiB) Viewed 874 times
Last edited by Booster on Tue 17 Apr 2018 22:36, edited 1 time in total.

Booster
Corporal
Posts: 37
Joined: Fri 13 Apr 2018 13:32
Contact:

Re: Suggestion: RNG in WG (for WG4?)

Postby Booster » Tue 17 Apr 2018 20:53

Zark wrote:I believe that such a system would increase the performance load on the game significantly.

Just tested it for 100 000 rolls:

Code: Select all

    006   1551   1.551 %
    007   9287   9.287 %
    008  23568  23.568 %
    009  31298  31.298 %
    010  23394  23.394 %
    011   9352   9.352 %
    012   1550   1.550 %

006-012 100000 100.000 %
                     6d2
[Finished in 0.272s]

And this is even not a C++ :!:
Most likely there is a different problem with SD:N44.

EDITED: We actually can use an array of predefined rolls (when the map are loading) and just move through when it's needed (when a unit makes a shot).

Booster
Corporal
Posts: 37
Joined: Fri 13 Apr 2018 13:32
Contact:

Re: Suggestion: RNG in WG (for WG4?)

Postby Booster » Mon 23 Apr 2018 22:31

Pondered how to be with planes with "no misses system". So, we can give a countermeasures ( in amount of ??? ) which can divert missiles with probability of 100% (95%?), or which will fire with a certain frequency. It's still can be a little chance to avoid missiles with no countermeasures.

Return to “Wargame : Red Dragon”

Who is online

Users browsing this forum: Google [Bot] and 30 guests