Fragmentation/Explosion Damage rises with number of nearby men

[V0ll808]

Mods (complete and add to the following information):

• Arma 3: 2.16
• CBA: 3.17.1
• ACE3: 3.17.1

Description:
The number of injuries caused by an explosion rises with the number of nearby men.

Steps to reproduce:

1. Start up Editor in Virtual Reality
2. Place down a man (player)
3. (Optional) Remove all armor/weapons from the man to ease the tracking of injuries
4. Place down the "Systems -> Modules -> Effects -> Ordnance" ~1,5 large white squares away from the man
5. Place down multiple men directly behind the player
6. (Optional) Go into the ACE Medical Config and do everything to reduce the chance of going unconscious. (Fatal Damage Source: Sum of trauma; Player Critical Damage Threshold: 25; Bleeding Coefficient: 0; Pain Coefficient: 0)
7. Start the scenario as the man placed down in Step 2
8. Stand still and wait for the mortar shell to land,
9. Check injuries after Explosion
10. Repeat scenario with different numbers of men (Step 5) and multiple times with the same number of men. Keep track of number of injuries.

Expected behavior:
Injuries caused by should be independant of the number of men behind the player.

Where did the issue occur?

• Editor (Singleplayer)

Log Files:
Nothing to see in the logs
https://pastebin.com/7iUJVTV7

Additional context:
During the Tests, the number of Large Avulsions, Large Velocity Wounds and Minor Bruises was fairly constant, the number of other injuries rose with the number of men. The rise might be connected to the fact, that a certain number of fragments is created per man and these can hit other nearby men aswell.

Screenshots:
Setup for testing:

Table of Injuries during testing:

[LinkIsGrim]

Should be fixed by the Frag rewrite.

[LinkIsGrim]

Please try again with current master (should be available @https://steamcommunity.com/sharedfiles/filedetails/?id=1882627645 soon, or just get it from https://github.com/acemod/ACE3/actions/runs/10907795420/artifacts/1943909823) and see if the issue persists.

[V0ll808]

Cant get the current master to run in the editor. Not sure about the reason, but I will post an update if there is an version I can get running and test everything.

[lambdatiger]

Don't think it will be with the current master version, and maybe reduced by #9728.

There are some limitations to the way ACE approaches the frag simulation. Specifically that spawning all the fragments for a specific explosion would not work for the engine. Instead ACE frag generates up to 50 fragments it believes would hit. It fills the difference if there are fewer than 50 with random fragments. While the simulation takes into account the number of fragments being spit out at certain angle, there should be significant overhead to limit the quantity for sufficiently small solid angles to prevent the issue.

If I find time, I'll take a look into it being more precise with tracking azimuth and elevation to avoid conflicts and see if there's a method that has a reasonable performance impact and not O(n^2) computations.

[lambdatiger]

There are methods to breaking a sphere into a certain number of quasi-uniform solid angles, but still looking for a way to handle simple enumeration/formulaic creation of angles. I may settle on a predefined LUT or something similiar, but I'll save that for #9728 if I can't find a better way to do it either dynamically or by equation.