ace_frag - way too overpowered

[severgun]

Arma 3 Version: 1.80 (stable / rc / dev) CBA Version: 3.5 (stable / dev + commit hash) ACE3 Version: 3.12.1 (stable / dev + commit hash)

Mods:

- CBA_A3
- ace

Description: ace_frag - way too overpowered. It is impossible to throw M67 and be safe at open area. WIKI: The M67 can be thrown 30 to 35 metres by the average male soldier. Steel fragments (not to be confused with shrapnel) are provided by the grenade body and produce an injury radius of 15 metres , with a fatality radius of 5 metres. Yes, fragments can travel much more than 15m but not every time and they should not deal so much damage.

Steps to reproduce: Throw M67 grenade at open area and don't take cover. Get hit almost 100% chance at 40m throw

Where did the issue occur?

• Everywhere

Placed Modules:

• None

Frags did not hit or flyby near camera even on failed throw https://www.youtube.com/watch?v=VkyLGHpys68 https://www.youtube.com/watch?v=FXPeWdmazPI

[severgun]

Seems like dispersion of self-guided fragments should be increased

[commy2]

The reason they are guided towards players is to have as much of them noticable near the player as possible without wrecking performance. Increasing dispersion would mean they drain performance without being noticable. Should rather decrease their densitiy.

[severgun]

May be it was during development as idea. But in current release self-guided fragments is here to kill or injure units and not to scare them. Now they are shooting like 'snipers'. At 30m-40m you have really big chance to get hit. All user manuals for hand grenades say that radius of injure hit is 15-20m. So at 15m it should hit units with 90-100% chance but at 30-50m chance should be much less.(because density of flying fragmets reducing with distance.) Because there is single fragmets and not wall of them flying towards units I call this dispersion, not density. #define FRAG_VEC_VAR 0.004 I think this one should be tuned may be to 0.01 or even more(need testing with debug tracers)

[commy2]

But in current release self-guided fragments is here to kill or injure units and not to scare them.

Where did you got this from? And how would a shrapnel not headed towards a unit injure the unit?

I call this dispersion, not density.

Well then you are using the terms wrong. There is no point in simulating projectiles dispersed too far away from the player, as they will have no effects. What you want is to decrease the density, the number of frags alltogether.

[severgun]

Where did you got this from?

Enabled debug tracers, logic and slack chat.

And how would a shrapnel not headed towards a unit injure the unit?

I'm not talking about projectiles not heading to unit. This solution is good. But they(which heading to you) are too accurate at long range. Effective range of regular grenade 15-20m.

Let's do some math. Grenade have 210 gramm metal shell. 2+gramm frags are actually capable to injure. So, let's say, that all 210g split into equal 2 gramm fragments(of course it will not). We have 105 fragments flying away as sphere.

Surface area of a sphere with radius 20m is (S = 4πR^2) 5026.548m2 105/5026.548 = 0.020 frags per m2 Man profile avg. 1.8m*0.5m = 0.9m2 0.01872 frags per 0.9m2 So 1.8% chance to get frag in standing man profile. At 20m.

At 50m - (105/(4pi50^2))*0.9 = 0.3% Now I got ~70%...

Correct me if my thoughts are wrong.

What you want is to decrease the density, the number of frags alltogether.

Density of 1 fragment? Every unit(50 units is the coded limit?) have 1 frag flying into his body. If you turn on debug tracers and throw 10 grenades you will see "cone" of 10 red lines going towards you. Cone is pretty narrow. I think that if increase dispersion of this projectiles, cone will be wider. And because there is one frag at a time, there will be more chance that it will miss your body profile at long range. https://imgur.com/a/nuqvR 0 hits on second image. Still lethal at 10m

IMHO Grenades should hit ~100% with 5m. and ~1% with 30m. That is what I'm asking about. Name it how you like it.

[kymckay]

The issue sound to me like the bias toward players is too significant and thus fragments are too accurate instead of seemingly random.

Ideally it should take a linearConversion of the distance from the frag point to the unit, such that it becomes a dispersion range within which we randomise the direction of the frag toward the units. This way the chance of being hit decreases with distance, but there's still a chance at all distances (or something along these lines, I haven't looked at the actual implementation).

[severgun]

Frags needs total rewrite... 👎 https://imgur.com/WpPcB8I

[dedmen]

You mean like https://github.com/dedmen/ACE3/commits/InterceptFrag ? :D Yes we already know it does. But I actually like how it currently works. I think.

[jonpas]

You only see actually simulated fragments, there are others, we just don't simulate them for performance reasons, nothing is there, so no one cares if something goes there.

[severgun]

You only see actually simulated fragments, there are others

is there a way to see them for debug?

I just want to show that now if behind your back there is group of people in radius 480m!(insane for grenade) you will be ripped in parts by selfguided shrapnel. When, as i calc earlier, there should be 0.02 framents per m2 at 20m and even less at 30m.

[jonpas]

No, they are not simulated at all, basically imagine them. I am not saying dispersion/amount overall is correct, just that you have to take into account that only those in path towards objects are calculated.

[kymckay]

@severgun I haven't looked at the code myself, but if I understand right the fragments aren't "selfguided", the physics calculations are all being done normally and then only the fragments that are heading toward targets actually get created

@jonpas can confirm if I've got that right

[jonpas]

Yeah, well described @SilentSpike 👍

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[stale[bot]]

This issue has been automatically closed due to inactivity. If this is still an issue, please feel free to re-open this. If necessary, provide any additional details to help us solve this issue. If you wish to assist us resolving this issue, please re-open or create a new issue stating you wish to help us out. Thank you for your contributions.