Welcome to symthic forums! We would love if you'd register!
You don't have to be expert in bit baking, everyone is more than welcome to join our community.

You are not logged in.

  • "leptis" started this thread

Posts: 282

Date of registration
: Mar 17th 2015

Platform: PC

Battlelog:

Reputation modifier: 8

  • Send private message

1

Sunday, November 27th 2016, 9:43pm

Recoil Control: A pattern of control (2). More Weapons.

In a previous post

Recoil Control: A pattern of control (1)

a Recoil Control pattern based in one weapon was found, finding that this was done in three domains differentiated: no control, adaptation and stable control. (X always miliseconds and Y always degrees)

Spoiler Spoiler



Subsequently, more records based on memory reading, from series of controlled bursts (3 or 4 per weapon. 24 bursts in total) of the same player SPECIMEN 0 have been available.

Weapons analyzed are: ACE-52 CQB, ACW-R, AEK-97, AUG-A3, LSAT, MG4 and TYPE 88. That weapons ranging from Recoil Up = 0.260 to 0.560.

The methodology of analysis has been the same as in the previous post arriving at the following medium bursts, in which the three differentiated domains that had been detected in the previous post are appreciated. We analyze the range from 0 to 1200 milliseconds that covers the definitive range (0-800 milliseconds approx.) in which the 10 bullets to be used in the simulator must be produced. There are other effects in long bursts - 20-30 bullets - (oscillations, loss and recovery of control, fatigue, etc ...) that are not analyzed here.

Spoiler Spoiler



You can see the difference between high and low Recoil weapons and the three phases. In the low recoil weapons control is easier and earlier, and even the control tends to worsen very slightly in the adaptation phase. With high Recoil the control starts after the third bullet (on average) and with low Recoil after the second, and sometimes after the first. Remember that high Recoil corresponds in general with high ROF.

With this we need four parameters to characterize the control:

- Reaction Time: at which the control starts (no control phase).

- Adaptation Time: time in which a stable level of average control is achieved (in the studied range).

- Stable Level: average level of the barrel position in the stable control phase.

- Standard Deviation: actual position of the barrel relative to the average position.

It has been found that there is a strong correlation between the FSMxRECOIL UP parameter and the previous parameters, which allows to model these parameters for each weapon (and consequently for any level of Recoil):

Spoiler Spoiler



This allows to simulate recoil control (at the level of SPECIMEN 0) for any weapon. Below you can see the simulation made for two weapons (high and low recoil). The model used is the calculation of the mean level of control from parameters (red dashed line) and simulate the variability by applying an uniform random model between +/- standard deviation which generate the solid gray line, which is different for any simulation. However, we do not need to know the position of the barrel at any the time of the process but only the position in which the barrel will be at the time of each shot being the red dots of the graph (simulated shots which also change in each burst).

Spoiler Spoiler



In the simulator two models will be implemented: uniform variation between +/- the standard deviation and variation biased to the center between +/- 2x standard deviation to simulate the position only at the time of the shots.

A more complex line of analysis has been found, because the player does not fight against a QUANTITY of Recoil, but against a QUANTITY in a certain TIME, which opens the way for a dynamic analysis in which ROF plays a fundamental role and the speed of adaptation to the stimulus as well. In that line the main questión is the angular speed (pitch/milisecond). That is, you can control ANY quantity of recoil, but you need ANY time to do it... and you haven't this... You can imagine the line.

Conclusion: A model of Vertical Recoil Control has been found to be implemented in the simulator.

Next posts: Control of the Horizontal Recoil to close the question and one of Consequences for Players (without graphs nor standard deviations :rolleyes: ).

Miffyli

Symthic Developer

(6,543)

Posts: 3,705

Date of registration
: Mar 21st 2013

Platform: PC

Location: __main__, Finland

Reputation modifier: 16

  • Send private message

2

Sunday, November 27th 2016, 11:44pm

Hmm especially the longer bursts are not common (?) in BF4 because most weapons share similar "micro-burst / tapfire is most accurate and deadly". Wonder how this applies to BF1 though where LMGs' optimal burst is ~7 rounds or so (spread starts high but decreases over time). I figure this applies quite well since recoil/spread mechanics are mostly the same, except recoil decrease is different (but now there's zero recoil decrease while shooting, versus 0.1x recoil decrease in BF4 while shooting).
Links to users' thread list who have made analytical/statistical/mathematical/cool posts on Symthic:
  • 3VerstsNorth - Analysis of game mechanics in BF4 (tickrates, effects of tickrate, etc)
  • leptis - Analysis of shotguns, recoil, recoil control and air drag.
  • Veritable - Scoring of BF4/BF1 firearms in terms of usability, firing and other mechanics.
  • Miffyli - Random statistical analysis of BF4 battlereports/players and kill-distances. (list is cluttered with other threads).
Sorry if your name wasn't on the list, I honestly can't recall all names : ( . Nudge me if you want to be included

  • "leptis" started this thread

Posts: 282

Date of registration
: Mar 17th 2015

Platform: PC

Battlelog:

Reputation modifier: 8

  • Send private message

3

Monday, November 28th 2016, 1:35am

Hmm especially the longer bursts are not common (?) in BF4 because most weapons share similar "micro-burst / tapfire is most accurate and deadly". Wonder how this applies to BF1 though where LMGs' optimal burst is ~7 rounds or so (spread starts high but decreases over time). I figure this applies quite well since recoil/spread mechanics are mostly the same, except recoil decrease is different (but now there's zero recoil decrease while shooting, versus 0.1x recoil decrease in BF4 while shooting).


I have given much thought to the length of the bursts.

Why are micro-bursts used?

You start shooting, you see two or three hitmarks and then nothing. You also see (or "fell") the sight outside the target. You stop, re-aim (forced or helped with the recoil decrease) and re-fire 2 or 3 microburtss until you get the 6-7-8 bullets you need to kill.

It's a question of angles. With a high recoil weapon the recoil of the third bullet (which we have already seen that is not easilly controlled) places the barrel with a pitch of about 2 degrees that will place the sight above the head to about 25 meters and that if the aim is perfect to the chest. If you try to correct the recoil the sight will not return to the chest until the 5 or 6 bullet. and that geometrically because the spread makes the probability of hitting very low. The feeling of the player is that with 8 or 10 bullets he has only 3 or 4 hits.. Better microbursts.

But you don't do that at 3 meters. The size of the target is huge.

I have calculated the distances (related to the probability of killing) WITHOUT RECOIL CONTROL (I have hundreds of calculations that I have not posted):



As you can see there are weapons that can only be used with microburts except if you are within 5 meters of target or you have an exceptional recoil control.

I'll tell you what happens with a certain degree of control. The distances will be longer, but I do not think much more. The spread in those cases is definitive to ruin the 100% probability..

I have not yet worked on these issues in BF1. But the optimal length of the burst has to be refered to the distance: THE SIGHT HAS TO BE WITHIN THE TARGET (at least), and that happens at a specific distance (longer with recoil control and shorter without it).

At 100 meters it is only possible to place the first bullet with or without control, and this only with the probability that the spread circle gives.

In any case for simulator the number of bullets has to be enough to not bound the results (as happen with Benet-Mercie Storm at close distance) because we know that you will stop the burts when you kill and the definitive data is the number of the bullet that has got the kill.