ST-CS-10-339-75
March 2002

4.  Recoil  (S)

A common recommendation to improve accuracy is that rather than holding down the trigger constantly a player should use burst fire, firing off short groups of rounds and waiting for the gun to settle before firing another burst. The idea is simply that it is better to have three rounds on target every two seconds than three rounds on target and then a large number of rounds off target.

The size of the crosshair when firing is intended to represent how unsettled (accurate - or in this case, inaccurate) the gun is at any given time.

It was found that after firing off enough rounds for the crosshair to reach maximum size and then pausing for about 0.5 seconds, the first shot fired in the next burst would be exactly on target but that the next rounds in the burst would immediately return to the high degree of inaccuracy previously achieved by the prolonged burst of fire.

To be succinct; after firing off some shots, if a player then fires another burst while the weapon is still unsettled, the first bullet fired in the second burst will be exactly accurate. After that, the gun returns to the expected high level of inaccuracy.

In the footage shown, the M249 first fires a contiguous burst of ten rounds on the left hand side of the wall. Aim is then shifted to the right hand side and after about a 0.5 second pause, fire is resumed, with another ten round burst being fired.

It can be seen that the very first bullet in the second burst falls exactly where the crosshair points - despite the crosshair being significantly enlarged from the first burst.

This behaviour is consistent across all automatic weapons.

Note that all crosshair movement is due to gunfire jerk and offset. Once the second burst of ten rounds is fired off, the crosshair naturally falls back to the original point of aim and it can be seen the first round lies there, while the other nine rounds all fell at the height expected from a M249 firing at full-auto.

Further experimentation revealed this behaviour is part of a more general recoil behaviour which displayed non-intuitive properties.

Crosshair dilation is supposed to relate to the accuracy of the gun. When the crosshair is fully dilated, the gun is least accurate. When the crosshair is fully contracted, the gun is most accurate. When the crosshair is 50 percent dilated, the gun is 50 percent accurate.

Intuitively, one would expect that for a given dilation (25 percent, 50 percent, etc) the gun will always have the same accuracy. For example, consider two MP5s. The first is in the process of firing a long burst and the crosshair has just reached 50 percent dilation. The second has fired off a long burst and is now no longer firing; the crosshair on this gun is in the process of contracting (having reached maximum dilation during the burst of fire) and is also just at the 50 percent mark.

It would be expected that if the second MP5 fired at that point, the accuracy of the bullet fired would be the same as that of the bullet fired by the first MP5 when it was at 50 percent dilation.

This is not the case.

To understand what actually occurs, first consider the behaviour of an assult rifle firing a long burst from minimal crosshair dilation. The first shot is exactly on target, the next few shots form a vertical column and the shots after this form a band at the top of the column as the crosshair shifts from side to side - the behaviour shown in the first burst of ten rounds from the M249 in the video footage.

The crucial fact is that when a gun is fired, it always produces this pattern. The difference the current degree of recoil makes is how quickly the gun goes from the first bullet being on target to the point where it is producing the side to side band.

With zero percent recoil, the gun will manage one shot on target, six or seven rounds in the vertical column and then go side to side.

With 50 percent recoil, it will manage three rounds - rather than the normal six or seven - in the vertical column before going side-to-side. The three rounds are equally spaced in the vertical column. The first round is exactly accurate, the second half way and the third at the top of the column.

If the crosshair is highly dilated (near 100 percent) the gun will only manage one shot on target before returning to the side-to-side pattern. This shot, however, is exactly on target.

If we consider our two MP5s (one at 50 percent accuracy in the middle of a long burst and the other at 50 percent accuracy during recovery) we will find that the first MP5 firing will produce a bullet which is at the beginning of the side to side band, whereas the second will produce a bullet which is exactly on target.

All in all, the recoil model is flawed. Firstly, as described in the critique, it does not accurately model gunfire. Secondly, even as it is in its flawed state, it permits players with highly unsettled weapons to fire exactly accurate shots.

Complex systems are unpredictable. Changes - even small changes - may or may not produce the desired effect, and are likely to produce a number of unpredicted effects, some of which may well outweigh the desired effect. Counter-Strike is a complex system.

The author speculates recoil was implimented in this was to make burst fire effective. It achieves this goal well enough. The unintended side-effect is that players can go full-auto and yet then be instantly accurate, avoiding the effects of recoil.