ST-CS-10-339-75
March 2002

2.  Gunfire Critique  (S)

The lack of relationship between consecutive bullets is completely unlike real weapon behaviour and has very serious consequences for Counter-Strike and is the cause of the very high frequency of headshots found in the current version.

In the bullet dispertion video, each frame lasts for 0.05 seconds. Bullets, once fired from a weapon, follow a straight line until they hit an object. This means that the barrel of the weapon, at the point of firing, must be pointing at the location where the bullet impacts.

It can be seen that the positions of consecutive bullet impact points change by such a large amount that the barrel of the weapon is changing angle so quickly (in the 0.05 seconds between each round) that several laws of physics relating to mass, acceleration and inertia are being broken.

It is physically impossible for the barrel of the gun to change angle that much in 0.05 seconds.

However, this is not real life; this is a computer simulation and so the barrel of the gun moves without regard to the laws of physics and interia.

The consequence of this behaviour is that all weapons, after their first few accurate shots, are like shotguns; they uniformly cover the area centered on the crosshair with bullets. This means that whether or not the player aims accurately at a target, bullets will still hit that target, because the target is still inside the large bullet cloud area, the area over which bullets randomly hit.

Note that randomness implies uniformity. If thirty shots are fired, and each shot is randomly offset from the crosshair, then each point around the crosshair has an equal chance of being hit - so the end result, after firing all thirty rounds, is a nice even uniform covering of bullets.

It is useful at this point to consider and compare this behaviour with the actual behaviour of a weapon. A real automatic weapon, when firing, produces a stream of bullets, so that each bullet fired impacts at a point fairly close to the point where the previous bullet fell.

Rather than a uniform random smearing of bullets all over the area in front of the weapon, a line of holes, following the movement of the weapon, is stiched into the wall in front of the weapon.

In Counter-Strike, every point in the area in front of the weapon has an equal chance of being hit by each bullet. Realistically, there is no chance of this occuring whatsoever; what would happen is that the small part of wall near where the current bullet fell will recieve the next bullet - the rest of the wall simply cannot be hit because there is not time to move the barrel of the weapon that far.

The consequences of this are severe. First and foremost, large numbers of headshot kills occur. This is because no matter where the weapon is aimed, as long as it is in the general vicinity of a target, the random spread of bullets usually hit the target and often lead to a headshot.

This is why players can jump out of cover, fire a quick burst of three or four shots - which are totally unaimed and where the crosshair is nowhere near the head - and yet score a headshot.

What happens is that because the player is jumping, the randomness factor applied to the bullets fired is huge. So although the players aim is hopeless (jumping, and no time to aim) and so miles away from any target, the bullets go all over the place - basically, the player has three or four random chances of scoring a headshot.

In reality, the bullets would go where he pointed his gun; e.g. into short stiched line of three or four bullets which in almost all cases would end up in a wall, since there is no time to aim.

Changes in recent versions of Counter-Strike which increased the degree of randomness of gunfire of certain weapons (SMGs while jumping, AK47 all the time) have served only to increase the frequency of headshots caused by these weapons, by ensuring they reach the point of maximum randomness more rapidly.

The following footage clearly demonstrates this problem.

The final position of the crosshair and distance between it and the sparks of the headshot are telling. It is clear that there was no aiming for the head; only the random bullet dispertion led to the headshot occuring.

The author finds it impossible to accept that at a range of one meter, the recoil from the AK47 is such that despite pointing the gun at the shins of the target, a headshot occured; the behaviour seen is an artefact of the Counter-Strike game engine.

When a player fires wildly, his weapon currently covers the entire area before him with bullets. As such, he can hit targets which were never anywhere near where he pointed his weapon. If a stream of bullets were emitted, however, and the bullets fired formed a line as if his gun were a pen and the wall a paper, then his wild firing would have one result; he would miss completely, and in the footage shown above, the headshot would never have occured.

In the author's opinion, the best way to reduce the number of unaimed headshots is to make the weapons more accurate, thus permitting gunfire which missed to actually miss.