'[R-DEV wrote:M42 Zwilling']
Back then, Mats and I (mostly Mats) set up a spreadsheet to collect data on the RHA-equivalent armor thickness for the front/side/rear/top of our vehicles. We got frontal values at least for most of the tanks, but there's a lot missing elsewhere that would need to be estimated. I might be able to ask and get you the link if you want to play around with it. If we want to rework this, I think it's high priority that we do it properly and use RL stats, even if that does mean duplicating and editing a ton of materials. And if we're doing that for armor, it would only make sense to do it for the projectiles that impact the armor as well... this would get pretty complicated.
This is the part where I wonder whether it's really worth it. There would definitely be tons of annoying skrubs everywhere like "omg why did you change a system that worked fine for years and break balance? do you even play the game? @#$%^&* devs" But it's all or nothing IMO. Either we leave it alone, or stupid copy-and-paste balance must die as violently as possible. It'll just be more work later if we end up overhauling it incrementally.
Anyway, if you want to help with trying to make this happen, here's my thoughts on what the model for KE damage should look like. From your posts on the forums, it seems like you want damage to correspond exactly with the most likely outcome of any scenario, i.e. if a 3BM42 penetrator from a T-72M1 hits an M1A1HA on the frontal arc, it should do absolutely nothing. What I would propose is something slightly different, though the end result is pretty similar.
Here's what its armor looks like from the front according to Steel Beasts.
Now the 3BM42 is supposed to have 450mm penetration at 2000m. Looking at the image, this tank could stop one of those for most of the frontal arc, but not all of it. Possibly the center lower glacis at close range, the area just below the gun, and the turret ring are weak points one could make it thorough.
I don't think reworking the colmeshes to represent these weak areas is feasible. I am told it would be a ton of work, and I don't think we have anyone willing to take that on. We have to simplify and use one material each for front, side, and rear as it is currently ingame (more or less).
Rather than basing damage on the most likely outcome, I think the damage output ingame should prioritize scaling with the
probability of making it through, when the projectile is dealing with a large amount of armor. Basing it on the average damage caused by a number of possible RL outcomes is the best we can do IMO. So just for example let's say in real life a T-72 would fire 10 shells at the front of an M1, and only 1 hits a weak area and penetrates, causing, let's say, 1000 damage to destroy the tank. Ingame, we just make each shell do 100 damage. The end result is the same, but the method of getting there is different due to the game engine and limitations we have in time/motivation.
Damage vs. armor it can go through like paper, like a HMMWV's, is another matter. The after-armor effects of solid penetrators like all our tanks use for AP rounds depend on them going through a lot of armor and spraying the resulting fragments everywhere when they make it through. If there's not much armor, there's a decent probability that it overpenetrates.
War Thunder represents this if you've ever played it. I took like 5 hits from T-34s a couple days ago in my M15 AA halftrack and only took a damage to parts that happened to be in the direct path of the projectiles, like 1 unfortunate crew member. I just drove away with minor damage. That's probably what would happen in real life too, and what should be represented ingame by reduced damage as armor decreases. Basically, I think our damage for AP rounds should be based on a formula, which, when graphed would look something like this (13/37 Paint skills amirite?):
This is where HEAT projectiles might be made actually useful, since they have explosives for damage and don't rely so much on bits of metal from the impact, much better than AP against light targets. Their damage graph would look more like a simple curve falling from left to right, again representing the probability of penetrating on the lower end.
One more thing that might influence how we implement this: randomizing damage a little may be possible. K4on discovered that having duplicate materials can result in weird cool stuff. E.g., he said if I were to duplicate the 5.56 material, and have the damage modifier vs. torso as 5 on one duplicate and 0 on the other, I might sometimes one-shot people and sometimes do nothing. When I tested it a little a couple months ago, I only got about 5% variation, but that doesn't mean more isn't possible. I just haven't figured out yet how to control the amount of randomness.
So yeah, what d'ya think about all this?
Do you need me to rephrase anything? Agree/disagree with my line of thinking on the damage model?
