Original title: “Recoil Control: Horizontal Recoil (and 3)”, 29.11.2016

Author: leptis

*Part 1:* Recoil Control: A pattern of control (1, BF4)

*Part 2:* Recoil Control: A pattern of control, more weapons (2, BF4)

The analysis of the horizontal recoil control is more complex than the vertical recoil one for two reasons:

1.- The Recoil H has a completely random pattern so it is not possible to compare the No Control pattern with the controlled bursts.

2.- The player’s response is not continuous. He only acts when he appreciates an excessive cumulative deviation (by events type" right-right-right…" or by the cumulative deviation with weapons with asymmetric recoil).

Due to this difficulty the analysis methodology will be different.

We will analyze the weapons: ACE-52, ACW-R, AEK-97, AUG-A3, LSAT, MG4 and TYPE 88 (24 total bursts).

We know that in an uncontrolled burst the extreme values are increasing statistically with the number of bullets.

In an uncontrolled burst of 10 bullets the statistical distribution of the barrel positions (coordinate X) follows a perfectly normal distribution with a maximum extension that referred to the initial parameters is 3.8 times larger. We know what we can expect in 10 uncontrolled bullets.

Here the result position of the barrel in 5,000 burts (10 bullets) without control (weapon of REC H LEFT and RIGHT = 0.200):

### Spoiler

The measured distributions are of the type:

### Spoiler

The total span value measured on average is on the order of 64% of expected: So we can say that the SPECIMEN 0 player controls the Horizontal Recoil by 36%. This would be sufficient for a fast simulation by multiplying the parameters by 0.64 for the recoil controlled scenario. In this weapon, (AUG) the value is 67%.

Not all bursts are so regular. Weapons with strong asymmetry and high horizontal recoil provide distributions like this:

### Spoiler

Asymmetric and with worse relative control (74%). The irregularity in the peak is due to the small number of bursts measured.

The procedure that we will use to cover the relative differences is to fit a Normal curve that has the same horizontal extension as the measured distributions with the parameters obtained by correlation: Parameter of scale of Normal distribution referred to Total Rec H (Rec left+Rec right) and center of the simulated distribution referred to initial center of Recoil H:

### Spoiler

As you see there are very good correlation that permit modelate any weapon.

With that adjustement of simulated recoil H is:

### Spoiler

It has been a very hard project, much more than expected due to small number of measured burts. Now to rest for a while reading posts from other authors.