Jump to content
Sign in to follow this  
pvt_ryan

The Elusive Stopping Power "Formula" and ArmA 3

Which of the following would be factored into your ideal damage formula for ArmA 3?  

362 members have voted

  1. 1. Which of the following would be factored into your ideal damage formula for ArmA 3?

    • Bullet Mass (heavier rounds have more energy)
      56
    • Bullet Caliber (bigger bullets create wider wounds)
      49
    • Bullet Shape/Form Factor (affects external ballistics, performance over range)
      39
    • Penetration (overpenetration causes an exit wound, but does not completely transfer energy)
      51
    • Performance vs. armor
      57
    • Other aspects of terminal ballistics (expansion, fragmentation, yawing, etc.)
      42
    • Whether vital organs were hit (as opposed to having a single upfront damage value with bleedout)
      68
    • Something else not mentioned here
      10


Recommended Posts

Let me preface this post by saying that I am not aware of how ArmA games traditionally assess how much "damage" a weapon does, so please feel free to fill me in. However, the idea is that this thread focuses more on the future of damage modeling (in ArmA 3 and beyond), and takes a more comparative aspect in terms of evaluating ArmA's current models with respect to other games currently on the market. I haven't seen any topics dealing with this, so excuse me if this is redundant, but I think a discussion of what "stopping power" entails can not only change how we approach FPS's, but also teach some of us about firearms in the process. Anyway, to the matter at hand:

So I'm sure as many of you are not only gamers but also firearm enthusiasts, you're aware of the controversy that goes along with the notion of "stopping power", at least in that sense that said quantity could be measured in a way that would facilitate any definitive form of comparison between firearms. Well, as both an FPS gamer and an amateur weapons aficionado, one of the most interesting things for me to observe, both over individual series' lifespans and between FPS series, is how this issue is tackled by developers. If there could be one single aspect of a game's mechanics that was most representative of its developers design philosophy, it would have to be damage models (although recoil models are almost equally interesting), at least in my opinion. The most obvious models for stopping power, in order of increasing complexity, are as follows:

  1. A single constant- Each gun deals a certain amount of damage that does not change across ranges (most older games use this model for the sake of its simplicity)
  2. Two constants with a damage gradient- Each gun has a certain maximum damage level that it deals up to a certain range, beyond which damage drops off until it reaches a second constant corresponding to the gun's minimum damage, regardless of range (as seen in CoD)
  3. A linear decrease- Damage drops off linearly from point blank until at very far ranges it practically equals zero
  4. Some sort of curve- Introducing a third variable into the damage equation in addition to range (such as a bullet's specific ballistic characteristics or the potential for over-penetration or fragmentation) will generally produce some sort of damage curve rather than a straight line.

That's all fine and well, but the crux of the matter is how said constants are determined and said curves are calculated. One of the most striking things is that the first two categories above are usually merely based off of popular conceptions of how much damage a weapon should do- often it even comes down to "bigger guns do more damage", or a precedent set by other earlier, popular FPS games.

  • Only slightly more ambitious would be actually factoring in the caliber of the bullet fired by the weapon in question. While such a model does have the advantage of being markedly more realistic than relying on how firearms are portrayed in the popular media, given that most modern-day cartridges are standardized this does have the potential drawback of ending up with many weapons having similar or identical stats. Some series such as CoD avoid this by somewhat arbitrarily readjusting damage in the interest of balance and variety and so cannot be considered to fall under this category.
  • The next level of complexity would account for either barrel length or muzzle velocity (of which barrel length is one of the primary determinants), which would be determined experimentally. The BF series claims to do this, at least in BF3, and it is a definite improvement over the aforementioned models, but it still ignores how various ballistic properties can affect a bullet's behavior mid-flight and so it cannot produce a true damage curve.
  • The following category of stopping power models would correspond to those designs that calculate how different bullets' dimensions and masses cause them to behave differently at varying ranges. Often said characteristics are summarized in a value known as the "ballistics coefficient"; I'm unsure of how many FPS series actually delve into this level of detail.
  • An even more advanced stage in modeling stopping power would be to make allowances for over-penetration, which refers to when a bullet completely passes through the target and therefore hypothetically does not transfer all of its energy. This would have the interesting consequence of concluding that each weapon has an optimally effective range, below and above which its performance decrease, meaning that from point blank range damage would increase for some distance before it finally began to decrease in the usual manner.
  • Finally, the most difficult (at least from my perspective) but comprehensive damage models would somehow predict certain aspects of terminal ballistics that are often seemingly random, including expansion, fragmentation, and yawing. The first refers to how a bullet deforms and increases surface area as it encounters the target, the second is rather self-explanatory, and the latter describes how a bullet can rotate perpendicular to its length (pardon the poor description). Obviously these things are extremely complicated to estimate and they have as much to do with what the target consists of as much as the actual gun and bullet themselves, which is why these models are far beyond the scope of mainstream FPS's.

For an ArmA noob like myself, where exactly do previous ArmA games fall on this scale? What improvements (if any) to damage models would you like to see with ArmA 3? What do you think the ideal "stopping power formula" is, in real life or in the context of video games? Also, feel free to correct me if you feel that I have anything wrong- part of the reason I have written this post is to serve as a little learning exercise for myself.

Also worthy of note are the decisions that must be made with regard to how these values should be measured; often one can compute a theoretical value for things like muzzle energy easily enough using basic physics formulas, or one could actually test fire these weapons and use the experimental data. Neither method is necessarily better than the other, since ultimately video games must rely on an equation to output a damage value. Which do you think is better in the context of a military simulation like ArmA 3?

inb4TL;DR

Edited by Pvt_Ryan

Share this post


Link to post
Share on other sites

ofp/arma and arma2 follows p3 more or less.

Now, since damage/stopping power whatever you want to call it is defined in the bullet class, the length of the barrel is not take into consideration. That said, one could simulate an increase of muzzle speed by setting up multiple bullets and therefore magazines for those different weapons. Dunno if that is actually needed.

Bullet calibre is taken into consideration within A2 configs (in the way those are set).

TBH i think that from this POV, things are quite well portrayed in Armaverse. I don't see any reason to go further down the road and simulate overpenetration.

From a ballistic POV, things are also pretty well setup = bullet drop for the weapons i have tested behave more or less the same with their RL counterparts (according to existing information).

I think that everyone around here is happy with the bullet ballistics. Or at least i never heard any complains. Rockets and their behaviour is a different matter.

I would much rather see BIS taking their time and focus on damage models for vehicles, or if you like, removing altogether the hitpoints based system.

Even if you don't own A2 or OA, you can test most of the stuff with A2 and make up your own mind there, even though weapon zeroing and similar are only available in OA

Edited by PuFu

Share this post


Link to post
Share on other sites

Now, since damage/stopping power whatever you want to call it is defined in the bullet class, the length of the barrel is not take into consideration. That said, one could simulate an increase of muzzle speed by setting up multiple bullets and therefore magazines for those different weapons. Dunno if that is actually needed.

Bullet calibre is taken into consideration within A2 configs (in the way those are set).

Hmmm... wow, I actually expected the damage models in ArmA to be a bit more complicated than that. That's not necessarily a hugely bad thing in the game, but I for one would like to see a more in-depth model, even if that just means taking barrel length into account (which is a very important part of a firearm's performance). I mean, I'm no programmer, but if I can easily calculate (relative and theoretical) stopping power while taking into account barrel length (or experimental muzzle velocity) on an Excel spreadsheet, it seems like it should be something they could put into the game, even if they had to move a couple of things around in the code.

TBH i think that from this POV, things are quite well portrayed in Armaverse. I don't see any reason to go further down the road and simulate overpenetration.

I'd generally agree, but ultimately such a complicated model can be useful for balance purposes because it generates a curve that allows each weapon class (ex. submachine gun, assault rifle, sniper rifle) to excel at its intended range. for example, one big problem that other FPS series have had with regard to weapon balance is that (in line with traditional thought) for sniper rifles to be powerful at far ranges they must also be powerful at close range, which gives rise to playstyles that blatantly contradict how sniper rifles are used in real life (don't worry, I'm not accusing a game like ArmA that has more hardcore fans of having a "quickscoping" problem, but hopefully you see the point). A damage model that takes over-penetration into account would actually predict that a sniper rifle bullet would (due to its high velocity) over-penetrate at point blank range, doing potentially less damage than it would from a more reasonable distance. Whether this is truly at all realistic (as this assumes that energy transfer is the primary determinant of injury), I am not sure, but it is a very interesting and arguably useful consequence to using such a model, and as such more complex models shouldn't be immediately dismissed as worthless.

I would much rather see BIS taking their time and focus on damage models for vehicles, or if you like, removing altogether the hitpoints based system.

Indeed, there are certainly other aspects of the game mechanics with room for improvement, and ultimately I'm sure everyone would love to see the end of seemingly oversimplified "hitpoints" systems. However, seeing as video games by nature will always be computer programs, damage/injury would always have to be reduced to numbers at some point, so we can never truly do away with quantified damage models. We can only improve our ability to hide the numerical nature of hitpoints by adopting more complicated damage models, which is why I think that a more in-depth model is called for even if there are changes in how injury is assessed in the game. Like I said though, I know very little about coding, so I could be completely off the mark.

Share this post


Link to post
Share on other sites

Over penetrate... I like where your going with this Pvt_Ryan. More complexity for wounding and types of wounding would be swell.

Share this post


Link to post
Share on other sites

I think that muzzle speed is taken into account in Arma. If I remember correctly the AK 74U will have a lower range than a AK 74

Share this post


Link to post
Share on other sites
Hmmm... wow, I actually expected the damage models in ArmA to be a bit more complicated than that. That's not necessarily a hugely bad thing in the game, but I for one would like to see a more in-depth model, even if that just means taking barrel length into account (which is a very important part of a firearm's performance). I mean, I'm no programmer, but if I can easily calculate (relative and theoretical) stopping power while taking into account barrel length (or experimental muzzle velocity) on an Excel spreadsheet, it seems like it should be something they could put into the game, even if they had to move a couple of things around in the code.

From all the games i have played so far (i doubt i missed any proper FPS in the last 10-15 years), arma is by far the most accurate. As previously said, the barrel lenght is configed in a simple way (muzzle velocity = range). The game does not calculate this by itself taking into acount the geometry or other special LOD from the model.

I'd generally agree, but ultimately such a complicated model can be useful for balance purposes because it generates a curve that allows each weapon class (ex. submachine gun, assault rifle, sniper rifle) to excel at its intended range. for example, one big problem that other FPS series have had with regard to weapon balance is that (in line with traditional thought) for sniper rifles to be powerful at far ranges they must also be powerful at close range, which gives rise to playstyles that blatantly contradict how sniper rifles are used in real life (don't worry, I'm not accusing a game like ArmA that has more hardcore fans of having a "quickscoping" problem, but hopefully you see the point). A damage model that takes over-penetration into account would actually predict that a sniper rifle bullet would (due to its high velocity) over-penetrate at point blank range, doing potentially less damage than it would from a more reasonable distance. Whether this is truly at all realistic (as this assumes that energy transfer is the primary determinant of injury), I am not sure, but it is a very interesting and arguably useful consequence to using such a model, and as such more complex models shouldn't be immediately dismissed as worthless.

Yes, that is something i agree with you on. But the issues is this sort of thing would need to be simplified somewhat for the game, since simulating it in real time would be overkill (things like ricochet, bullet fragmenting, rotation, angle of penetration in relation with point of entry, body armor type, bones and organs hit and so on) would be a bit too much for a game that is already CPU intensive.

That said, at this very moment, a bullet does NOT go through bodies (as in killing 2 guys with the same bullet is impossible even at close ranges with a 50cal). It does go through different types of materials of the environment.

Indeed, there are certainly other aspects of the game mechanics with room for improvement, and ultimately I'm sure everyone would love to see the end of seemingly oversimplified "hitpoints" systems. However, seeing as video games by nature will always be computer programs, damage/injury would always have to be reduced to numbers at some point, so we can never truly do away with quantified damage models. We can only improve our ability to hide the numerical nature of hitpoints by adopting more complicated damage models, which is why I think that a more in-depth model is called for even if there are changes in how injury is assessed in the game. Like I said though, I know very little about coding, so I could be completely off the mark.

BIS doesn't need to change the ballistic system in order to improve the injury system. I feel that the way it is now is real enough. Making things more complicated will NOT change the end user experience in any notable way.

I think that muzzle speed is taken into account in Arma. If I remember correctly the AK 74U will have a lower range than a AK 74

Yes, it is, as i have already said in my post

Edited by PuFu

Share this post


Link to post
Share on other sites

Various barrel lengths and their effects can be easily scripted, and is implemented like that in ACE; http://ace.dev-heaven.net/wagn/Muzzle_Velocity

Penetration etc is modeled to some extent in the Engine, in ACE we try to raise the bar somewhat; http://ace.dev-heaven.net/wagn/Wounding_System

More hitpoints (being able to find out where a bullet hit on the body other than just "Overall", "Head', "Body", "Legs", "Hands") could offer more accurate / interesting wound modelling.

Edited by Sickboy

Share this post


Link to post
Share on other sites

Your suggestion for optimal ranges is a pretty neat idea and would make choosing and using the right weapons for the right jobs more important. But personally I would rather see more detail in the way a unit takes damage rather than the way in which damage is dealt. Basically hits should produce more random effects because everyone will react to being shot differently in reality.

More hitpoints (being able to find out where a bullet hit on the body other than just "Overall", "Head', "Body", "Legs", "Arms") could offer more accurate / interesting wound modelling.

shot placement is more important than calibre when determining the stopping power of a bullet. Therefore I agree that it would be more benificial to add more hitzones to create realistic wounds rather than changing the way in which bullets act.

Share this post


Link to post
Share on other sites

One thing to understand - in ballistics, there is no one true formula. There are well-accepted formulas that model various aspects of projectiles as well as popular simplifications. The most controversial section is terminal ballistics - where a projectile interacts with another object.

Muzzle velocity is the topic of interior ballistics. Stock Arma only allows for muzzle-velocity to vary by weapon - meaning that it's very limited in the situations where muzzle velocity will match reality. At least one mod mod does try to adjust for ammunition type by use of tables and extended configuration files - but that system is difficult to maintain due to difficulties in collecting accurate data. As a result, errors are bound to slip in now and then. If the tables were the results of actual calculations the performance would be about the same, but the developers would need good data about powder composition and geometry, primer type and geometry, barrel wear, and data on any suppressors/muzzle brakes or the geometry of the barrel crown otherwise.

Exterior ballistics needs to account for subsonic, trans-sonic, and super-sonic speeds were applicable. Pistol rounds are typically slow enough that one only needs to deal with subsonic values. Everything else should be handling all 3 domains since the transition through mach-1 causes a significant change in flight behavior. The range at where a round drops to subsonic speeds occurs is generally accepted at the limit for accurate fire by markman. For indirect fire, one of the 6 degree-of-freedom models is a must since those rounds encounter environments that require extra calculations if the point of impact is be determined with any accuracy.

The round impacting with it's target involves terminal ballistics, which is vastly more complex than what has required up to this point. Materials like steel, aluminum and concrete use radically different models to characterise how each is damaged or penetrated. For wound results - the research by Fackler is commonly cited, but it's not universally accepted.

The gist of all this is that ArmA is using very simple assumptions on weapons behavior. It or mods could do much to improve on the stock program. But at some point you will have to draw a line on detail if you wish the simulation to not bog down during a large firefight.

Share this post


Link to post
Share on other sites
From all the games i have played so far (i doubt i missed any proper FPS in the last 10-15 years), arma is by far the most accurate. As previously said, the barrel lenght is configed in a simple way (muzzle velocity = range). The game does not calculate this by itself taking into acount the geometry or other special LOD from the model.

Ah, I guess I misinterpreted that then. Constants for muzzle velocities (and the barrel lengths which they would derive from) are certainly an efficient and sufficient model for damage over range in my opinion. Really, the only disadvantage to that system that I could imagine would be if there was a very large number of weapons, some would still end up with identical stats, and of course such a model would also fail to account for how a growing understanding of ballistics has led to modern cartridges having much better performance than older ones. Although, that's only problem if the firearms in the game represent a broad historical range, but even then it's not a huge issue at all. Perhaps then the next step for ArmA's damage model is to have a more realistic representation of penetration instead of bothering with the meticulous specifics of external ballistics.

Yes, that is something i agree with you on. But the issues is this sort of thing would need to be simplified somewhat for the game, since simulating it in real time would be overkill (things like ricochet, bullet fragmenting, rotation, angle of penetration in relation with point of entry, body armor type, bones and organs hit and so on) would be a bit too much for a game that is already CPU intensive.

Indeed, obviously memory resources are always an issue and for a series devoted to realistic simulation like ArmA probably the primary limiting factor. At least simpler models do have some advantages- every time you introduce a new variable, the damage "feels" more random, which may be seen as frustrating for those who want reliable kills in a multiplayer context. In this way, there are sacrifices that go along with more "organic" damage models.

That said, at this very moment, a bullet does NOT go through bodies (as in killing 2 guys with the same bullet is impossible even at close ranges with a 50cal). It does go through different types of materials of the environment.

Hmm... that's interesting, one would think that if bullets could penetrate soft cover they could also penetrate tissue; of course not having such a feature doesn't hold the game back, but it could be nice to have (within reasonable technical limits) and simulating penetration would kill two birds in one stone. It sounds like if ArmA's model has any room where improvement would be both practical and significant, it's penetration.

---------- Post added at 03:52 PM ---------- Previous post was at 03:50 PM ----------

Various barrel lengths and their effects can be easily scripted, and is implemented like that in ACE; http://ace.dev-heaven.net/wagn/Muzzle_Velocity

Penetration etc is modeled to some extent in the Engine, in ACE we try to raise the bar somewhat; http://ace.dev-heaven.net/wagn/Wounding_System

More hitpoints (being able to find out where a bullet hit on the body other than just "Overall", "Head', "Body", "Legs", "Hands") could offer more accurate / interesting wound modelling.

Ah... nice work. I love to see the community take games to the next level.

---------- Post added at 04:18 PM ---------- Previous post was at 03:52 PM ----------

One thing to understand - in ballistics, there is no one true formula. There are well-accepted formulas that model various aspects of projectiles as well as popular simplifications. The most controversial section is terminal ballistics - where a projectile interacts with another object.

Indeed, I am very aware of the fact that this is a contentious issue, which is why I avoided proposing a specific equation of my own and only focused on which variables were being factored in.

Muzzle velocity is the topic of interior ballistics. Stock Arma only allows for muzzle-velocity to vary by weapon - meaning that it's very limited in the situations where muzzle velocity will match reality. At least one mod mod does try to adjust for ammunition type by use of tables and extended configuration files - but that system is difficult to maintain due to difficulties in collecting accurate data. As a result, errors are bound to slip in now and then. If the tables were the results of actual calculations the performance would be about the same, but the developers would need good data about powder composition and geometry, primer type and geometry, barrel wear, and data on any suppressors/muzzle brakes or the geometry of the barrel crown otherwise.

Exterior ballistics needs to account for subsonic, trans-sonic, and super-sonic speeds were applicable. Pistol rounds are typically slow enough that one only needs to deal with subsonic values. Everything else should be handling all 3 domains since the transition through mach-1 causes a significant change in flight behavior. The range at where a round drops to subsonic speeds occurs is generally accepted at the limit for accurate fire by markman. For indirect fire, one of the 6 degree-of-freedom models is a must since those rounds encounter environments that require extra calculations if the point of impact is be determined with any accuracy.

The round impacting with it's target involves terminal ballistics, which is vastly more complex than what has required up to this point. Materials like steel, aluminum and concrete use radically different models to characterise how each is damaged or penetrated. For wound results - the research by Fackler is commonly cited, but it's not universally accepted.

The gist of all this is that ArmA is using very simple assumptions on weapons behavior. It or mods could do much to improve on the stock program. But at some point you will have to draw a line on detail if you wish the simulation to not bog down during a large firefight.

Indeed. Great post, I feel significantly more informed now. I will say though that it's my belief that applying ballistics to a video game is a bit simpler than you present it to be. Of course, there's no doubt that ballistics in real life is very complicated, and each aspect (internal/interior, external/exterior, and terminal) is equally important, but for the purposes of a video game that is not the case. In real life the two main applications of ballistics (at least the ones I can think of at the moment) are military ends and forensic science.

Forensic ballistics requires one to work backward to the source, which in the context of the video games is largely irrelevant. Alternatively, for the sole purpose of calculating stopping power/damage/injury, interior ballistics are only relevant insomuch as it affects external ballistics, and external ballistics is in turn only significant in its effect on terminal ballistics.

In this way the theoretical details of internal and external ballistics can be ignored, and instead the relevant quantities can be measured experimentally, which takes out a lot of unnecessary computations (unless of course the game in question is more concerned with ballistics theory than real firearm performance). If internal and external ballistics can just be summarized and represented by simple sets of values, then the only sub-field left for a game like ArmA 3 to excel in is terminal ballistics, as ominous as it is. Indeed, most mainstream FPS's ignore terminal ballistics altogether under the clearly wrong assumption that external ballistics accurately predict damage/injury. This is why terminal ballistics are in fact an area of great opportunity for a series like ArmA. Obviously certain things like yawing and fragmentation are likely overkill and too CPU intensive anyway, but I think one could make a good case that more realistic penetration models could really benefit the game.

Share this post


Link to post
Share on other sites
Obviously certain things like yawing and fragmentation are likely overkill and too CPU intensive anyway, but I think one could make a good case that more realistic penetration models could really benefit the game.

Anything and everything can be done by probability. If the engine can't handle rolling a 20 side die (or thereabouts) for damage, then it's not much of an engine.

Share this post


Link to post
Share on other sites

Hi, many times i've wondered what happens with the bullets once they reach a certain distance... now on the ArmA2, the tracers dissapear at a certain (unknown for me) distance, i think that's something like 600 to 800m, and i've my ViewDistance set at 3000m; do the damage effects of this "dissapearing bullets" work?, i mean... once the tracers dissapear, the bullet and the damage effects of it, dissapear too?.

Other thing... AFAIK two 7.62x51 bullets on the head should kill you, playing to the ArmA2, this isn't always a rule for some (unknown to me) reason, hitting the target's head at less than 100m and the ArmA2 don't simulates the presence of a BP helmet, overall because we're talking of two shots in the face, one under the nose the other on the eyebrows; i don't really see huge changes on the game's system/engine from the ArmA to the ArmA2 and to the OA (that i'd uninstalled).

An example of what i mean, if you shoot at a BRDM-2 with a M-107, to the driver, or passengers possition (always talking of the ArmA2) nothing happens... the bullet don't passes the armor of the BRDM-2; that the military say that if it receives a shot of a 12.7x99mm the bullet will enter for a side and will exit for the other on the 85% of the cases, IRL. They say that the chances of pierce (from side to side) the turret is like a 65% or so, still a 85% to pierce the front side if the vehicle is on a flat surface, offering just as deflection "protection" the natural angle of the turret and it's armour as resistance to the piercing. Sumarizing... a realistic damage system will require a powerfull hardware, but will require much more if it were simulating (that not recreating) the damage on the human body.

Other (last) thing before i forget it, a bullet very out of it's "flat" trayectory because of the wind can loose at leat 1/3 of it's force; AFAIK the ArmA3 don't gonna simulate the wind in any way. Don't gonna move the bullets from side to side, don't gonna increase or decrease the vehicles fuel consumption aswell their max speed, acceleration, maneuverability etc. IMO the damage system should have in consideration the wind speed change on the distance, the air density, the altitude, the moisture and air pressure for be realistic enough. But i'll be happy if it had in consideration at least the wind speed changes on the distance, the caliber, the length of the barrel & the number of grooves per inch. Let's C ya

Share this post


Link to post
Share on other sites

http://forums.bistudio.com/showthread.php?p=2016983#post2016983

Check that thread out, has covered almost everything to do with how this kind of stuff could be implemented easilly in arma 3, it would even work in arma 2.

It has been around for a while, but hasn't had heaps of replies in a while.

It talks about body armour as well, as you may or may not know, body armour can protect from up to 6 shots from a 7.62x51mm ball round shot at 50 metres with a muzzle velocity of 840m/s

Ofcourse, that would ruin the game, so there could be armour piercing ammo as well. This ammo does slightly less damage but has much better penetration, capable of penetrating body armour first round within 300 metres.

IMO there is no sense in having massive long calculations in formulas to decide how well a round will penetrate x material, when you can give it an arcadish value and get a similar result. That kind of stuff doesn't need to be super accurate, although it deffinitly needs to be based on real life data.

---------- Post added at 04:26 PM ---------- Previous post was at 03:41 PM ----------

Forgot to mention, one of the main ideas is that the damage value given to a bullet is based on its terminal ballistics. Throw that in a simple formula with bullet velocity as the only other main variable and you have a better than average system.

Throw in a few scripts, for example one to simulate fragmentation, and you have an awesome system. The formula would be very simple and stuff like expansion, tumbling and yawing, exit wounds should all be considered when making the SINGLE damage value.

Stuff like fragmentation, explosions, incindaries etc should be addition values. Whether or not they determain depends on chance and projectile velocity. For example:

5.56x45mm M855A1 ball round

damage = 7

bodyArmorDamage = 3

vehicleDamage = 1

fragments = 1

incindary = 0

explosive = 0

fragmentationMinVelocity = 650

penetrationMultiplier = 1.2

incapShots = 1

Another point this thread makes is that acculmative damage should not kill, but the wounds it produces should eventually kill. Bleeding, burning, shock (from explosions, collisions and falling), and headshots are what kills you.

Depending on the location of the hit, and the end damage result after all the numbers have been crunched will cause bleeding and stuff like pain (already in arma 2) instead of just death.

The more you bleed, the faster you will bleed out and then die.

Damage determains bleeding rates (which can accumulate with multiple wounds), it does not determain incapacitation.

Incapacitation is determained by a value in each bullet's (or magazine or whatever) config. If a bullet's incap value is 1, then it only takes 1 direct hit to incapacitate someone. A direct hit is any bullet that hits the unit and is not stopped by body armour.

People get technical on the topic of stopping power and what it takes to physically incapacitate someone. Blah blah central nervous system blah - F**k that. Pain is more than enough to put someone down.

Yeah yeah I know adrenaline, but the idea that adrenaline will numb ALL pain is bull. End of the day, a soldier cannot sprint around, accurately fire his weapon or fight, with a gut wound. People pass out from pain, people's bodies stop functioning properlly in some ways to keep the rest of the body alive etc.

The way the wounded are treated is similar to the current AIS module: soldiers can stop bleeding, medics can heal. Units can be dragged and carried and pulled out of vehicles.

Anyway I am blabbing on too much if you want to know more check out the thread I linked.

Share this post


Link to post
Share on other sites

Indeed, there are some other good threads discussing ballistics and wounding - some reaching back into ArmA2.

One thing that deserves clarification is what is meant by realistic. One definition implies getting exactly the same results for a given RL situation. Another is where any deviations are statistically insignificant, which allows for greatly relaxing the amount of number grinding required. Let's stick to the statistical variant to preserve game performance and sanity ( those ultra-realistic formula can get really messy ).

A .338 Lapau Magnum round is going to behave pretty much the same way inside any weapon designed to use it. So it would be easier to simply state the cartridge type along with barrel length and an optional fudge factor to correct for manufacturer differences to get the muzzle velocity instead of specifying a long list of obscure parameters that increase the risk of error creepage. The notion is to maximize precalculation in order to increase both performance and statistical realism.

External ballistics is fairly unforgiving, since you do need to get the correct value for hit/no-hit. Fortunately it requires a relatively small number of parameters and would be imbedded in the core simulation so all weapons are treated the same way.

Terminal ballistics - this is where individual opinion diverges. If you treat any armor as a homogeneous material ( not true IRL, but often close enough ) it is possible to compile a table of penetration velocities. One can take the residual velocity along with optional spall or scabbed armor and move on to the wounding calculations.

At first view, it seems that statistical methods are the simplest. I would keep the use of random numbers to bare minimum though. It's not to disagree with the posting by maturin, but merely due to the fact that a good random number generator is fairly expensive in terms of cpu cycles, nor are sections of code using random numbers typically easy to make multi-threaded for use on large-core processors. So I recommend avoiding use of random numbers where possible.

This brings up an important point - the simulation is worthless unless it can run the problem in a reasonable amount of time. Many games do not use any ballistics at all for shrapnel, instead using a technique called ray-casting to handle the cloud of projectiles flying around. This works because the relatively short range and sufficently high velocitiy of the fragments means that the assumption of straight-line movement used by ray-casting is close enough.

I've spent considerable time enhancing projectile ballistics as part a project not involving BI. Good ballistics do make for noticably better simulation, but must be coded carefully at the core level for maximum performance. The mod maker is left to focus on the terminal effects, being sure to ensure consistant results using well-researched data. Hopefully, ArmA3 is following a similar path. If so it will be magnificent simulation.

Share this post


Link to post
Share on other sites
It talks about body armour as well, as you may or may not know, body armour can protect from up to 6 shots from a 7.62x51mm ball round shot at 50 metres with a muzzle velocity of 840m/s

So me, today, being extremely effective with a silenced XM8 using SD ammo at 2-300 meters (maybe more?) against army personnel ("armored", compared to civ/ins units), isn't realistic?

C'mon, the people have spoken, they want this crap in. They don't care about realism, only about being more efficient in a game than real life. Don't get me wrong, I actually don't want realistic firefights wrt how many bullets vs kills, but the current efficiency is beyond insane.

Hey Evil_Echo, haven't seen you around for a while :)

Not sure about random numbers having to be so costly. Seed a generator at start and store say 100 numbers in an array and loop through those. Every couple of minutes or so the array is updated. 100 longword sized floats for 100 players stored on server memory wouldn't take that much space or time to generate, then just issue (yeah, I know) the seed number to each player and have him generate his own. Sync accomplished, not that many applications would have to be synched like this.

People get technical on the topic of stopping power and what it takes to physically incapacitate someone. Blah blah central nervous system blah - F**k that. Pain is more than enough to put someone down.

But not everything is painful. I've stepped on a nail (about 2 inches into my heel), which hurt quite a bit despite it's relatively narrow damage zone. I've also cut my foot on broken glass, where I didn't feel a thing for a while (it became a "weird" sensation stepping on my foot, but others pointed out I was painting the ground red, lol) probably because the nerves got cut and blocked. It took several hours before that actually started to hurt, of course then it hurt badly too :D Smash your shinbone into a table, and most cry like babies (or at least invent some new words). Level of hurt varies, which is why I think randomness is a good thing.

Edited by CarlGustaffa

Share this post


Link to post
Share on other sites

Yeah yeah I know adrenaline, but the idea that adrenaline will numb ALL pain is bull.

It's enough to make the effects of some wounds not kick in for a while. Adrenaline will cause dopamine to flood the synaptic cleft and use up the majority of the receptors in the associated dendrite. If the receptors are occupied then then effects of other neurotransmitters will be limited. It's the same with some drugs. I've read stories of Afghan insurgents needing a helluva lot more rounds put into them to bring them down - it's the same science but much more exaggerated.

It won't stop all pain, but it will certainly nullify a large proportion of it.

Share this post


Link to post
Share on other sites

Hi, about the stopping power... penetration... adrenaline effects... there's a video on youtube that you can find if you look for: I shot myself! Original Upload!. Have strong words and shows the .45 ACP entry and exit holes at point blank, is an instructive video; but maybe could bring some light to the thread. Judge by yourselves. Let's C ya

Share this post


Link to post
Share on other sites

Hehe, sorry, but I lol'ed hard on his reaction to the event :) Good video though, and nice to see warnings like this get posted despite the fear of getting ridiculed for it.

Share this post


Link to post
Share on other sites
Anything and everything can be done by probability. If the engine can't handle rolling a 20 side die (or thereabouts) for damage, then it's not much of an engine.

True, I hadn't even considered the possibility of using randomized dice to simulate terminal ballistics. I think with different dice for each ammo type that would do a more than sufficient job of modeling fragmentation and yawing while keeping things as simple as possible. Penetration probably couldn't be represented by a die though.

An example of what i mean, if you shoot at a BRDM-2 with a M-107, to the driver, or passengers possition (always talking of the ArmA2) nothing happens... the bullet don't passes the armor of the BRDM-2; that the military say that if it receives a shot of a 12.7x99mm the bullet will enter for a side and will exit for the other on the 85% of the cases, IRL. They say that the chances of pierce (from side to side) the turret is like a 65% or so, still a 85% to pierce the front side if the vehicle is on a flat surface, offering just as deflection "protection" the natural angle of the turret and it's armour as resistance to the piercing. Sumarizing... a realistic damage system will require a powerfull hardware, but will require much more if it were simulating (that not recreating) the damage on the human body.

Well, I think penetration (both in vehicles and soft targets) is a must for any simulator. In both cases I think the best way would be a simple version of what some here have called a component-driven damage model. I'm not asking for ArmA to simulate the locations of individual vital organs and bones and what not, but there's nothing too terribly limiting about different damage multipliers for the limbs, torso, and head. Also, the maiming thing (having slower movement when shot in the leg, having considerably more recoil or less accuracy when shot in the arm) and bleedout potential do go a long way towards making combat more immersive.

Other (last) thing before i forget it, a bullet very out of it's "flat" trayectory because of the wind can loose at leat 1/3 of it's force; AFAIK the ArmA3 don't gonna simulate the wind in any way. Don't gonna move the bullets from side to side, don't gonna increase or decrease the vehicles fuel consumption aswell their max speed, acceleration, maneuverability etc. IMO the damage system should have in consideration the wind speed change on the distance, the air density, the altitude, the moisture and air pressure for be realistic enough. But i'll be happy if it had in consideration at least the wind speed changes on the distance, the caliber, the length of the barrel & the number of grooves per inch.

Hmm... yeah, wind is something I'd love to see in a simulator, but generally it doesn't have a big impact on bullet damage (unless shooting into the wind would slow the bullet down significantly). I'm thinking about making another thread that would summarize factors such as recoil, bullet drop, wind, and any other things pertaining to accuracy over range much in the same way that this one deals exclusively with wounding/stopping power.

It talks about body armour as well, as you may or may not know, body armour can protect from up to 6 shots from a 7.62x51mm ball round shot at 50 metres with a muzzle velocity of 840m/s

Of course, that would ruin the game, so there could be armour piercing ammo as well. This ammo does slightly less damage but has much better penetration, capable of penetrating body armour first round within 300 metres.

IMO there is no sense in having massive long calculations in formulas to decide how well a round will penetrate x material, when you can give it an arcadish value and get a similar result. That kind of stuff doesn't need to be super accurate, although it definitely needs to be based on real life data.

I really like your idea about having different types of ammo for each caliber. In fact, the distinction between the damage of normal rounds and the lower damage was one of my main reasons for wanting penetration simulated (in one way or another, whether it be a simple value programmed into bullet statistics or actually calculated in-game), and has everything to do with the overpenetration that I am talking about in this thread. Although I'm not sure the hollow points would go over too well with the strictest of the realism crowd, I think a tradeoff between standard ball and AP rounds would really add to depth.

Forgot to mention, one of the main ideas is that the damage value given to a bullet is based on its terminal ballistics. Throw that in a simple formula with bullet velocity as the only other main variable and you have a better than average system.

Throw in a few scripts, for example one to simulate fragmentation, and you have an awesome system. The formula would be very simple and stuff like expansion, tumbling and yawing, exit wounds should all be considered when making the SINGLE damage value.

Stuff like fragmentation, explosions, incendiaries, etc should be addition values. Whether or not they determine depends on chance and projectile velocity. For example:

5.56x45mm M855A1 ball round

damage = 7

bodyArmorDamage = 3

vehicleDamage = 1

fragments = 1

incindary = 0

explosive = 0

fragmentationMinVelocity = 650

penetrationMultiplier = 1.2

incapShots = 1

Looks good, I really like it! Yeah, let me clarify that I'm not saying the game has to have this hugely complicated formula that it has to go through for every single bullet fired in the game- models that reduce these variables to single values that can either be measured experimentally outside of a game or simulated by a random die are certainly sufficient in my opinion. It's just that I think these variables cannot be ignored in a "military sim" like ArmA.

Another point this thread makes is that accumulative damage should not kill, but the wounds it produces should eventually kill. Bleeding, burning, shock (from explosions, collisions and falling), and headshots are what kills you.

Depending on the location of the hit, and the end damage result after all the numbers have been crunched will cause bleeding and stuff like pain (already in arma 2) instead of just death.

The more you bleed, the faster you will bleed out and then die.

Damage determines bleeding rates (which can accumulate with multiple wounds), it does not determine incapacitation.

Incapacitation is determined by a value in each bullet's (or magazine or whatever) config. If a bullet's incap value is 1, then it only takes 1 direct hit to incapacitate someone. A direct hit is any bullet that hits the unit and is not stopped by body armour.

Indeed, instead of completely eschewing an upfront numerical damage model altogether to focus on a completely organic wounding system, I like your idea of having a dual model; bullets do a certain amount of damage upfront, and cause bleeding which deals damage over time. Bleeding is another aspect in which a good penetration model is beneficial; a big question is whether there would be an exit wound that would cause additional bleeding.

The way the wounded are treated is similar to the current AIS module: soldiers can stop bleeding, medics can heal. Units can be dragged and carried and pulled out of vehicles.

Well, that system sounds good to me. A system that distinguishes between stopping bleedout and actually healing would also be great in my opinion. Thanks for the info and ideas; I tried searching for threads before making this one but nothing terribly redundant showed up with the keywords I used.

---------- Post added at 05:47 PM ---------- Previous post was at 05:28 PM ----------

One thing that deserves clarification is what is meant by realistic. One definition implies getting exactly the same results for a given RL situation. Another is where any deviations are statistically insignificant, which allows for greatly relaxing the amount of number grinding required. Let's stick to the statistical variant to preserve game performance and sanity ( those ultra-realistic formula can get really messy ).

Indeed, the issue of what realism actually entails is a huge discussion in itself, for sometimes real life can be seemingly random, so does a certain amount of randomness then belong in a simulator or should a game like ArmA stick to strictly theoretical models? There's no perfect answer.

A .338 Lapau Magnum round is going to behave pretty much the same way inside any weapon designed to use it. So it would be easier to simply state the cartridge type along with barrel length and an optional fudge factor to correct for manufacturer differences to get the muzzle velocity instead of specifying a long list of obscure parameters that increase the risk of error creepage. The notion is to maximize precalculation in order to increase both performance and statistical realism.

Indeed the barrel length/muzzle velocity is in most cases the only variable between guns that is relevant to how much damage bullets do, or close enough to being that anyway. You're right that even the "fudge factor" could be a single value, but at least to me I think I might rather see a little bit of chance in that or not bother having such a variable at all.

Terminal ballistics - this is where individual opinion diverges. If you treat any armor as a homogeneous material ( not true IRL, but often close enough ) it is possible to compile a table of penetration velocities. One can take the residual velocity along with optional spall or scabbed armor and move on to the wounding calculations.

Yeah, that's along the lines of what I'm going for, and a binary model of penetration failure or success is sufficient for most materials, but I think the game could really benefit from treating soft targets and materiel (including vehicles) with a little more detail, making it possible for a bullet to either be absorbed/deflected by armor, penetrate partially and eventually come to rest somewhere inside the target, or fully penetrate and create an exit wound. The same would go for vehicles, and how much armor each individual part of the vehicle has should also be factored in.

Share this post


Link to post
Share on other sites
Hi, about the stopping power... penetration... adrenaline effects... there's a video on youtube that you can find if you look for: I shot myself! Original Upload!. Have strong words and shows the .45 ACP entry and exit holes at point blank, is an instructive video; but maybe could bring some light to the thread. Judge by yourselves. Let's C ya

45ACP%20230gr%20FMJ.jpg

:D (http://www.firearmstactical.com/wound.htm)

Share this post


Link to post
Share on other sites
It talks about body armour as well, as you may or may not know, body armour can protect from up to 6 shots from a 7.62x51mm ball round shot at 50 metres with a muzzle velocity of 840m/s

Only Class IV body armour would protect against 7.62 NATO at such a short distance, almost nobody is issued Class IV body armour. FLAK jackets are more like Class II (will stop 9mm Para and 9mm Mak from a distance). Both the FLAK jackets and PASGT helmets will provide absolutely no protection from 5.45mm or 5.56mm in ranges under 400 meters, 7.62mm will zap straight through it from more than 500 meters. I.e., body armour should be able to "reduce" lethality in the game due to less penetration, but the player should in every situation in which he is shot with anything other than a pistol round be severely injured, and die from wounds if not treated.

Pistol rounds shouldn't work on center of mass hits on an armoured opponent though.

Another thing that bugs me is the ability to shoot through the canopy of gunships and fighter jets. Almost all of them are rated bullet proof up to 12.7mm ball, so without armour piercing rounds, nothing much would happen if you try to shoot the pilot. Right now you can even shoot him wit 9mm Para!

Share this post


Link to post
Share on other sites
So me, today, being extremely effective with a silenced XM8 using SD ammo at 2-300 meters (maybe more?) against army personnel ("armored", compared to civ/ins units), isn't realistic?

C'mon, the people have spoken, they want this crap in. They don't care about realism, only about being more efficient in a game than real life. Don't get me wrong, I actually don't want realistic firefights wrt how many bullets vs kills, but the current efficiency is beyond insane.

Hey Evil_Echo, haven't seen you around for a while :)

Not sure about random numbers having to be so costly. Seed a generator at start and store say 100 numbers in an array and loop through those. Every couple of minutes or so the array is updated. 100 longword sized floats for 100 players stored on server memory wouldn't take that much space or time to generate, then just issue (yeah, I know) the seed number to each player and have him generate his own. Sync accomplished, not that many applications would have to be synched like this.

But not everything is painful. I've stepped on a nail (about 2 inches into my heel), which hurt quite a bit despite it's relatively narrow damage zone. I've also cut my foot on broken glass, where I didn't feel a thing for a while (it became a "weird" sensation stepping on my foot, but others pointed out I was painting the ground red, lol) probably because the nerves got cut and blocked. It took several hours before that actually started to hurt, of course then it hurt badly too :D Smash your shinbone into a table, and most cry like babies (or at least invent some new words). Level of hurt varies, which is why I think randomness is a good thing.

I have also cut myself quite seriously on glass where i couldn't feel it for a while because i had cut the nerves off.

I also fell into the dishwasher and the big carving knife went right into, being cms away from my liver. Tell you what I couldn't move, I couldn't open my mouth to screem in pain... I had a lot of adrenaline when it happened because I paniced and thought I was going to die as there was a lot of blood. I felt every bit of it.

A big gaping hole from a rifle round is far more than enough to put someone down. And as for the SD ammo... You should be getting headshots anyway, otherwise you run the risk of the enemy screaming in pain. But the standard ammo soldiers will be using by the time they are fighting wars where both sides have body armour will be armour piercing ammo, but guerillas may only be able to obtain ball or hollow point rounds.

And remember you can create the mission how ever you like.

---------- Post added at 04:52 PM ---------- Previous post was at 04:43 PM ----------

Only Class IV body armour would protect against 7.62 NATO at such a short distance, almost nobody is issued Class IV body armour. FLAK jackets are more like Class II (will stop 9mm Para and 9mm Mak from a distance). Both the FLAK jackets and PASGT helmets will provide absolutely no protection from 5.45mm or 5.56mm in ranges under 400 meters, 7.62mm will zap straight through it from more than 500 meters. I.e., body armour should be able to "reduce" lethality in the game due to less penetration, but the player should in every situation in which he is shot with anything other than a pistol round be severely injured, and die from wounds if not treated.

Pistol rounds shouldn't work on center of mass hits on an armoured opponent though.

Another thing that bugs me is the ability to shoot through the canopy of gunships and fighter jets. Almost all of them are rated bullet proof up to 12.7mm ball, so without armour piercing rounds, nothing much would happen if you try to shoot the pilot. Right now you can even shoot him wit 9mm Para!

You're wrong, I took all that info from the class III armour classification, also, US army are issued class 4 inserts now, I am 99% sure that NATO all have atleast class III armour today. Also, newish 5.56x45mm rounds(M855a1) (in service) have better penetration when compared to the 7.62x51mm ball round (m80) Check out the link i posted. The ideas on that thread are based ENTIRELY off extensive research, not just two second wikipedia checks :)

Ofcourse the main reason for body armour and incapacitating wounds isn't to please those super realism fans, but it will make the flow of combat more realistic if done properlly while still keeping the game fun for everyone. One dead is one man down, one wounded man is 2 or 3 men down. (2 because another person must get him to safety and attempt to stabalise him. 3 men because a medic may need help, and i assume it would take 2 men to carry a stretcher, which were a great, yet non compulserary feature of ace mod.)

Modern ammo is designed to wound not kill outright because it not only demoralises the enemy, it slows them down much faster by putting more than one man out of the fight at a time. And even if the casualty survives, he will most likely be a burdon to the team. In arma this could be implemented by allowing medics to heal while soldiers can only stabalise (stop bleeding), but even medics can't fix broken legs and only field hospitals and ambulances can (module, as this will be shit in SP and will piss off casuals) and this would be a great addition to the game and I am already thinking of some scenarios where while on a mission, a man is injured and we have to extract, while being chased by enemy, and this casualty is proving to be a major burdon because he has to be carried everywhere.

I would like it if all of that was just included in the AIS modules, although I think the AIS module should automatically be synced to everyone when active, kind of like how the ace mod's wounds module is always active for everyone once the module is placed in the editor.

2 things that deffinitly need to be in the vanilla game even without modules:

Body armour and ammunition types to go with it.

No more of this system where once a certain amount of damage points are received, even from just leg shots.

Edited by Tom1

Share this post


Link to post
Share on other sites

I am all for body armor As long as it's realistic.

Also, if body armor is implented I really don't want to see people taking a shot or two and not even flinching because they are wearing armor. Wearing armor or not, getting shot is going to Hurt like a bitch and most likly send you into shock, and pain for a fair bit of time. This needs to be implemented in game.

Share this post


Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
Sign in to follow this  

×