One of the most misunderstood and misrepresented elements in RPG design is the effects of firearms in the real world. In fact, few in the real world understand the subject and myths abound everywhere.
On the surface, one would think that shooting someone would be a well understood matter. The truth is far different. The area has seen countless studies, vast amounts of work, and there is the experience of millions of deaths to draw on. And yet, only the most basic conclusions can be reached.
And sometimes what are considered firm conclusions are overturned.
In the 70's the FBI used a measure of weapon effectiveness called the Relative Incapacitation Index (RII) which in effect gave a heavy weight to the Kinetic Energy of a bullet (KE = M * V ^2). This help lead to a rush of 9mm handguns becoming the standard for police nationwide (and even the military). In the 80's, the error of this was revealed in spades and today most law enforcement agencies (and elite military teams) have moved to the .40 S&W and (back to) the .45 ACP.
Given the complexities of the subject, no role-playing game has even come close to modeling realistic firearm combat.
How Complex is it?
Lets consider the classic example of the 'tumbling' bullet:
The M16 uses a 5.56mm bullet moving at high velocity (nearly 3000 fps in its latest standard military design) that is commonly thought to 'tumble' in its victim resulting in huge damaging wounds.
The fact is a little different than the common thought.
Its been know for a very long time that any bullet is subject to 'tumbling' if its velocity is high enough. This is a result of the relationship of velocity to spin against the resistance of the medium being penetrated.
The actual effect is not the full endless tumbling that comes to mind. It is a single flip of the bullet 180 degrees as the heavier bottom rotates to point in the direction of motion.
During this flip (depending upon the structural strength of the bullet and the medium being traveled through) the bullet will often shed its jacketing and fragment into multiple pieces. At high enough velocities, this can be as violent as if the bullet exploded. This creates a truly ugly wound. At low velocities, no fragmentation will occur.
If the velocity was too high, the bullet will fragment too soon causing an injury too shallow to seriously affect the target. If the velocity is too low, the bullet will never flip or will do so too late in the penetration of body to cause major damage.
Now add in the fact that a bullet begins losing velocity as it leaves the barrel.
The M16 bullet will only tumble if it strikes it's victim at a range of 100 yards or less as beyond that it has dropped to less than 2000 fps. Within that range, it may flip too soon or too late to cause a significantly more deadly injury.
And finally, if used against a dense target, like say a deer instead of a human, it may flip too soon causing a shallow wound whose best chance of causing death is from infection. For these reasons, the 5.56mm is an illegal round to use against deer in many states (those with the larger animals) in the US.
For a realistic RPG would have to model this, it have to vary the damage by both range and target with individual results for each design of bullet in use. Talk about chart heavy.
And that's just one factor. Even worse is attempting to model the the effect of a hit on person. Real life shootings produce an incredible wide range of responses ranging from a victim fainting at the sound of the gunshot to those who have been hit dozens (yes dozens) of times without apparent effect.
Those wishing further information on the real world effects of firearms are advised to check out http://www.firearmstactical.com/ and/or contact the IWBA for a vast array of eye opening information. This FBI document is especially valuable for a beginning overview as is this list of ballistic myths. In the past some have recommended Evan Marshall's data, I find his studies to be flawed and untrustworthy for reasons you'll find on my recommended sites.
Opposed to the questions of realism is the fact that most games are designed to simulate a genre (Superheroes, Hong Kong Action, Wild West) rather than reality. Additionally few would find completely realistic gunfights to be any fun at all; for many players, it is more important to recreate the action from a book or movie than to recreate a real life police shootout.
Any RPG design must balance the concerns of realism and genre to achieve their goals. However it is my opinion that very few goals require one to completely reject reality; if a system should do so it should be so noted in the designer notes. Too many people are taking RPG representations as actual facts these days.
While complete realism would damage a game by sheer complexity and is impossible due to lack of solid data to boot, Suspension of Disbelief (SOD) is important. At least some basic real world concepts should apply unless one is specifically wanting truly wild results.
The following are the known key facts about real world firearm wounds that I think should be displayed in all but the most unrealistic games:
Most people when shot are incapacitated (down and useless)
Most people are not killed when shot if medical attention is prompt
Those who are not incapacitated by a injury often don't notice the wound, let alone suffer additional negative effects beyond the those one would expect from the local area of the wound itself (limping, inability to stand on a broken leg, etc.).
Multiple gunshot wounds seem to have no cumulative effect on performance beyond bleeding
Besides the location of the injury, bullet diameter is the most important factor in damage followed by bullet fragmentation. If penetration is too low, damage is reduced (over penetration has little effect).
Hydrostatic Shock (the big gun of the discredited RII) has no impact on damage except for brain and liver hits.
Most combatants have a very difficult time hitting their target. The typical police officer hits with only one out of six shots at ranges under 20 feet!
Many of the above points are counter intuitive to many people (especially the lack of cumulative effect) and may actually cause SOD problems despite the fact that they're firmly grounded in reality. Education is the best counter to that problem.
Applying this information to a RPG design is a tricky matter. There are issues of game balance, limits of useable complexity, genre elements, and pure character survival to take into account. Out of the gate, most game systems are correct in upping the chance of hitting your target in order to avoid lengthy and boring battles.
Still many games could use a bit of tweaking, the most important thing is to give the appearance of being real- and this can be done without actually detailing every little point and element.