Gnalvl

To give a basic example of where the sound engine bothers me - the scale of minor sounds like brass hitting the floor is too loud compared to major sounds like gunshots. In my experience shooting, even if you're wearing hearing protection that filters in quiet noises and filters out loud ones, you're not really aware of the sound of the brass. Even when it's hitting hard surfaces like concrete, the sound it really too dull to stand out over the firing - occasionally a particularly loud brass impact might be barely audible, but it's not this constant "BOOM!.. put! BOOM!... put! BOOM!...put!" you hear in Arma 2 and 3. Honestly I find that sound rather annoying and distracting when I'm trying to concentrate on visually referencing where my bullet hit.

I'm glad to hear the current grenade system is a placeholder, as IMO the offhand grenade is horrible. It's way too quick - yes, modern tactical slings allow you to essentially just "drop" your rifle and toss a grenade , but you've still got to pull the pin and wind up a pitch. Given how little time is taken to wind up, the grenade also travels way too far. I definitely hope to see this improved in the future...tapping G as a quick bind to reach grenades is fine, but the time taken to pull the pin is crucial. At that point, you may as well use the time pulling the pin to give the player a chance to pick overhand or underhand throw, cooking off, etc.

What we really need is someone to do a test where various shooters fire, using rifles of various calibers, with laser sights affixed to each. With a camera recording the path of the laser during each shot, we could find the appropriate range of muzzle deviation during recoil in minutes of angle, how fast the muzzle should realign on target, and how accurately it should do so. Assuming such a test could be done, I think the best test subjects would be people who'd been through basic training and qualified with rifles, but not with years and years of experience. After all, shooting in a videogame will always be easier, and players will always compensate by pulling down and correcting with the mouse. Therefore, if you base the recoil on a SEAL with 20 years of experience, expert control of recoil, and then the player corrects with the mouse on top of that, you're back to excessively low recoil and easy long range shooting. Also, I meant to add, it's kind of clear what happened with the 6.5mm rifles... BIS attempted to "balance" or compensate for the high capacity magazines by artificially beefing up the recoil. Since it's a fictional weapon and a fictional caliber, one can't really say it's incorrect. Though, if the goal is to emulate 6.5mm Grendel, it's not really in-between the 5.56,, and 7.62 rifles...it's over both.

What we really need is someone to do a test where various shooters fire, using rifles of various calibers, with laser sights affixed to each. With a camera recording the path of the laser during each shot, we could find the appropriate range of muzzle deviation during recoil in minutes of angle, how fast the muzzle should realign on target, and how accurately it should do so. Assuming such a test could be done, I think the best test subjects would be people who'd been through basic training and qualified with rifles, but not with years and years of experience. After all, shooting in a videogame will always be easier, and players will always compensate by pulling down and correcting with the mouse. Therefore, if you base the recoil on a SEAL with 20 years of experience, expert control of recoil, and then the player corrects with the mouse on top of that, you're back to excessively low recoil and easy long range shooting. As for the 6.5mm recoil, I think I can guess what's going on here. Since BIS has implemented fictional weapons for the opfor, with innordinately high magazine capacities, they figured they could "balance" things by giving 'em innordinately high recoil. Given that it's blatantly higher than 7.62x51mm rifles we see in game, it couldn't be called a 6.5 Grendel analog. Personally, I'd have rather seen the enemies armed with contentional, real rifles.

Yeah, I mean temporary cavity could have a role in a few specific things, but you'd need to model permanent and temporary damage as 3D space with volume, rather than as points, and as well as model individual organs. In that case, if the temporary cavity area intersects with inelastic organs like the liver, spleen, or kidneys, that could be counted as additional permanent damage. There is also a speculated stun effect from temporary cavities impacting the spine, I would be ok with things like screen blur, camera flinching, etc resulting from such hits - in general I might expect a bullet like 5.45 fmj to cause more "stun" effects than a bullet which deals the same amount of permanent damage without the same temporary cavitation. Those sorts of details would be nice to have, but not as important as the points in my earlier posts. Yeah, armor is sort of a whole other can of worms. Low level armor doesn't stop rifle rounds - in some cases it may affect how the bullet does or does not tumble or fragment, but in others it may not. These days ammo companies design their bullets to be "barrier blind" and shooting gel through barriers like glass, wood, cinder block, and steel sheets frequently produces the same sorts of wounds as without any barrier. For higher level armor, we know its guaranteed to stop certain numbers of rounds from certain ammunition, but we have little info on how many rounds of other ammunition it will withstand, or what happens when the "guaranteed hit number" is exceeded. For example, if a trauma plate is guaranteed to stop three hits of 7.62 NATO AP, what happens on the 4th hit? It may be destroyed on the 3rd hit and immediately start letting rounds though as if it wasn't there, or it may let rounds through with reduced wounding ability, or it may stop a certain number of other rounds - this info generally isn't given out. Moreover, we typically aren't told specifically how many rounds it will stop of 5.56 fmj, 5.56 AP, 7.62x39, or other miscelaneous mil-spec loads. Higher level armor is also allegedly more vulnerable to pistol caliber rounds than to rifle rounds, since it is prioritized to stop the latter. But without buying a bunch of armor and shooting it first hand, I don't know that you could realistically answer all these questions, even though it is important in creating an accurate simulator. A refined bleeding system could definitely go a long way to making wound ballistics in games more realistic. In Red Orchestra 2, many shots that don't kill you outright bleed you instead, about half of which you can bandage if you react fast enough. In many cases, it works out pretty nicely - a lot of the time, you shoot someone and even if it doesn't kill them outright, they just crawl into a hole and try to bandage themselves, so it still effectively stops them from whatever they were trying to do when you shot them. They might eventually come back, but many times they still don't manage to bandage themselves in time, or since they're stuck in a ditch in no man's land, they'll just get shot again as soon as they poke their head out. Other times, a particularly determined player might choose to shoot you back rather than worry about themselves. There are also many times where a wounded player bleeds out in only a few seconds, and their screen is too dimmed to do anything, but it does tie up your attention for longer than if they just dropped instantly. The variation and uncertainty is pretty good. Of course there are things to be worked on - pistol rounds in particular in RO2 are too easily shrugged off. Basically anything that's not a headshot can be bandaged in 3 seconds and then you go on like nothing ever happened. There's room for improvement and things to watch out for in these sorts of models. Yeah, I mean you don't want things to be too random, and the more results that can be based on actual cause and effect, the better. But when X bullet pretty much always kills in 2 hits, while Y bullet always kills in 3 hits, things are getting too arcade.

To clarify, I'm not suggesting that it should be impossible to die from extremity shots. On the contrary, I would say that a shot to any part of the body should have a chance to kill you. However, there is a huge difference between real life deaths from extremity shots, and videogame insta-deaths. Even if we take one of the most brutal real life extremity shot deaths as an example: a shot to the femoral artery, which can bleed you out within 30-60 seconds, that still allows a gigantic window in which the victim might continue letting loose stray fire at his attackers. By comparison, if you attempt to carry this over under a conventional videogame damage system, you just wind up with soldiers that die instantaneously, the second they are hit anywhere. Arcade apologists will tell you that the solder isn't really dieing instantly from a leg wound, they are merely being "incapacitated" and bleeding to death gradually...but the game already counts them as dead because they "can't" fight. It's a poor excuse; this system is a simplified representation of reality which entirely removes the risk inherent in real life combat. Any reputable source will tell you, shots which don't damage the central nervous system don't guarantee instantaneous incapacitation. Without the possibility that the enemy could continue firing for 1,2, 3, 5, 10, or 30 seconds after a lethal extremity shot, then all a "one shot can kill" mentality does is promote twitch shooting mechanics; "I can go rambo because I know if I round the corner and shoot them first, they can't get me".

Recently, I measured the total area of permanent cavities in Fackler's wound profiles as square centimeters. Damage within the first 20cm was counted for 100%, while damage between 20-30cm depth was counted for only 25% since this it's outside the average depth of the human torso. The results were as follows: 5.56 no frag - 88cm (9) 7.62x39 fmj - 113cm (11) 5.45x39 fmj - 149cm (15) 7.62x51 fmj - 181cm (18) 5.56 with frag - 386cm (38) 6.8 SPC - 795cm (79) 7.62 otm - 1088cm (108) I chopped a decimal off in the parentheses to give an idea of what it'd look like on a 100-ish point scale like most games use. I also did some calculations counting temporary cavity as 10% extra damage within the 20cm mark and 2.5% between the 20 and 30cm mark, but since most bullets deal permanent and temporary damage in equal proportions, it made little difference and I discarded them. The only exception was 5.45x39, which suddenly looked just as damaging as 7.62x51mm using the added temporary numbers, but since temporary damage is not a confirmed factor in incapacitation, why go with calculations which yield entirely identical results except artificially boosting a single caliber? It's not like the permanent-only calculations don't show a strong advantage in the round's early yaw, so I stuck with those. Regardless of how the damage gets calculated, I think generally in a good damage system, the caliber should perform in propertion to these values. In other words, a 5.56mm which fragments properly should be twice as likely to drop the enemy as a 7.62x51 FMJ, and a 7.62x51fmj should be twice as likely to drop the enemy as a 7.62x39mm or 5.56 through-and-through. On that note, the ranges at which a 5.56 round either fragments or doesn't should be based on the weapon's barrel length, the distance of the shot, and ammo type. Beyond that, there's a few factors I think would be important in perfecting Arma 3's damage system: 1) Precise Center Mass Hit Detection - damage should be scaled based on how precisely center-of-mass it hits. By extension, this means that the lethal hitzone in the chest will vary based on the power of the round - maybe a weaker bullet would need to hit a center area the size of a golf ball, while a more powerful one could hit in a center mass area the size of a baseball. 2) Non-Instant Incapacitation - IRL, shots which don't hit the brain or base of the spine have no guarantee to stop an enemy instantaneously. Therefore, the time it take for an enemy to drop from a lethal non-head wounds should be randomized to a degree. Even if it only takes 1-2 extra seconds for the enemy to drop after a lethal shot, this makes a big difference. Basically, the typical mentality of FPS is "as long as I can round the corner and put 5-6 bullets in the other guy first, he'll die before he can kill me and I win". If the enemy no longer stops the milisecond the server recognizes the lethal shot, then he might use the extra time to shoot you back. Hence, unrealistic rambo-style twitch run-n-gun play becomes far more risky, far less glamorous and practical. Basically, there should be a chance to bleed to death, or bleed till you feint or go into shock, or simply go into shock from tissue and nervouse system damage, getting hit in a large number of places, but those methods of incapacitation should take place over the course of anywhere from miliseconds to minutes, depending on the caliber/ammo-type of the bullet, the placement, player health, and the main subject of my next point: 3) Randomization - The human body doesn't always react to being shot the same way, and bullets don't alway react to piercing human tissue the same way. IRL every bullet from the same factory batch of ammo doesn't perform absolutely identically, and due to random variances in their spin and yaw cycle, they won't always fragment, expand, or yaw properly in the enemy's body. Combined with the inconsistancy of human physiology and human emotions, there will always be an element of uncertainty to terminal ballistics. More damaging bullets should have a higher chance to incapacitate more quickly, but there should always be a range of possible results.

I think wound profiles of various ammo types need to be considered rather than simply leaving it up to only muzzle energy, penetration, or "energy transfer" calculations. If you're going to ignore the fact that 5.56 FMJ fragments [sometimes!], 5.45 FMJ yaws early, and both 7.62 caliber's FMJ merely go in and out in most instances, then when it comes to terminal ballistics Arma 3 would be no better than COD, CS, BF, or any other military-themed arcade shooter. What's more, when it comes to JHP, JSP, OTM, AP etc rounds, the ball game changes completely, and you can't really come up with one model to summarize the performance of all bullets available in a single caliber, particularly when rounds that perform the worst in FMJ suddenly perform the best when the right ammo is used. As others have stated, it's probably too much to expect real-time calculation of bullets penetrating, yawing, fragmenting, expanding, etc through organs, tissue, and bone. However, even if each round were to be assigned a simple numeric value based on an average of its penetration, temporary cavity, permanent cavity, etc. which was factored in with velocity and shot placement, then that would take the damage model light years ahead of arcadey alternatives.