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.

Hey! If this is your first visit on symthic.com, also check out our weapon damage charts.
Currently we have charts for Battlefield 3, Call of Duty: Black Ops 2, Medal of Honor: Warfighter and Call of Duty: Modern Warfare 3

  • "leptis" started this thread

Posts: 282

Date of registration
: Mar 17th 2015

Platform: PC

Battlelog:

Reputation modifier: 10

  • Send private message

1

Saturday, February 6th 2016, 8:31pm

Sensitivity of Accuracy to Spread and Recoil

The question is to know what spread (or recoil) means, not in terms of parameters, but in terms of results (mainly accuracy and damage). So when we say that HB reduces the spread but recoil increases ... what does it mean in terms of accuracy (does it reduce... or increase)?.

I have analyzed the variability of results (accuracy and damage) in the SPREAD-RECOIL space, specifically when an SPREAD BASE and RECOIL UP vary in the range of the weapons in the game do.

For that I have prepared a weapon with the average parameters of all full-auto weapons in the game, where only the parameters SPREAD BASE and RECOIL UP will vary. Horizontal Recoil, FSM, and SpreadInc are fixed in the average values of the game. A further sensitivity analysis shows that the variation in results between using Mean or use Mean +/- Stand. Dev. provides errors of less than 2% for the SpreadInc and 4.5% for FSM.

RECOIL UP varies from 0.000 to 0.600, in 10 steps (11 values), even though there aren't weapons in the game below 0.180, to have a complete "map"(or to know what happens if you are able to correct the RECOIL to "0").

SPREAD varies from 0.000 to 3.000 in 20 steps (21 values) to cover most of the chances in the game (no weapons under 0.200... or 0.100 using some some attachment).

The position of the actual weapons in this "map" can be seen here (a factor of 0,5 is used for moving). The grey area is the "space" analyzed. Position of axes required for subsequent graphics. Yellow dots: mean values : RECOIL UP= 0.336 for all cases and SPREAD= 0.301 (ADS stat.), 0.635 (ADS Moving =,5) and 2.411 (Hipfire stat.)


Spoiler Spoiler



With that we have a set of 231 weapons that I process with the first module of ILETH "engine" (results against target) at all distances, without any comparison (perfect aiming to center mass of torso, 2.000 bursts 10 bullets/distance, continous burst, etc...).

The quantitative results for plain accuracy (number of bullets on target) can be seen here (10 m):

Spoiler Spoiler




Clarifying that, the "map" of isolines is, for 10 m:

Spoiler Spoiler



For 20 m:

Spoiler Spoiler



INITIAL CONCLUSIONS

- The "map" of space SPREAD-RECOIL UP is highly anisotropic and non-linear. To say "use that atachment" may be correct in a part of this space ... but a mistake in other one.

- The isolines are, in fact, parabolic third grade that trend to be straight lines with high values. That means the relationship between SPREAD and RECOIL UP varies depending on the point of the "map", but in the area where most of the weapons are (yellow segment), the relation is 1:10 (0,1 RECOIL UP produces the same result than 1 SPREAD). The relation 1:5 may be taken as average in the "map".

The resuts in terms of damage have a curious feature. 10 m:

Spoiler Spoiler



and 20 m (yellow line =DMG 100)

Spoiler Spoiler



The best results (DMG) at 10 m are not in the minimum SPREAD-RECOIL !!!. A bit of RECOIL is good for better results (it is logic... more headshot in the bursts). And this happen (10 m map) in the area where mean values of RECOIL are (the game is cleverly designed). Due to anisotropy you have to think differentially with Low RECOIL weapons than High RECOIL ones.

We are going to do an exercise. To apply HB to some points of the "map" (it does't matter if a weapons exists in it or not). This is to increase RECOIL UP 30% and reduce SPREAD 50%. I know that SPREAD INC also changes, but we have already established that the effect is small:

Spoiler Spoiler



The effects seen in CBQ are very different. Low RECOIL weapons improve but Middle and High RECOIL weapons much worse ... if you don't correct RECOIL (if you can). For these weapons in ADS stationary you only create RECOIL that you will have to correct later without any other improvement. High SPREAD weapons are mainly Hipfiring ones and HB doesn`t work with them, so only the two upper rows of arrows represent actual weapons. HB has other virtues at range...but not at close distances. Many other conclusions can be drawn from this.

This exercise can be done with other attachments that modify SPREAD or RECOIL.

This post has been edited 9 times, last edit by "leptis" (Feb 7th 2016, 12:38am)


  • "leptis" started this thread

Posts: 282

Date of registration
: Mar 17th 2015

Platform: PC

Battlelog:

Reputation modifier: 10

  • Send private message

2

Sunday, February 7th 2016, 1:04pm

To be more clear, I have calculated how damage varies using HB for two cases of RECOIL: 0.180 and 0.420 (extreme cases).

Obviously when HB is used RECOIL and SPREAD values change, but the graphics are referred to the initial values.

Areas where there are real weapons are shown (ADS-Stat. and ADS-Mov. 0.5). The Hipfire area is not shown because HB has no effect in that case.





As you can see weapons worse in ADS-Stat (always continuous burst and no RECOIL correction), and there is only a small area of improvement for ADS-Mov. 0.5.

Changes are more important with low RECOIL than with High RECOIL and they increase with the initial SPREAD. And biger close than far.

Unfortunately, the area where there could be a substantial improvement (high initial spread) are fast movement (factor >0.5) or hipfire area, but DICE design wanted in these area HB hadn`t any effect.

This post has been edited 2 times, last edit by "leptis" (Feb 7th 2016, 6:12pm)


Posts: 293

Date of registration
: Jun 9th 2012

Platform: PC

Battlelog:

Reputation modifier: 10

  • Send private message

3

Sunday, February 7th 2016, 2:58pm

Very detailed analysis, thanks. But Vrecoil is much more controllable than say Hrecoil. Sometimes I find that after putting on the HB to lower initial spread but ended up getting thrown off by Hrecoil. In that regard I think Hrecoil is acting somewhat like a pseudo-spread. Best example is PDW-R HB. I dont really find that I am much more accurate after putting on the HB, at least it feels much less than the stat would suggest.

  • "leptis" started this thread

Posts: 282

Date of registration
: Mar 17th 2015

Platform: PC

Battlelog:

Reputation modifier: 10

  • Send private message

4

Sunday, February 7th 2016, 6:09pm

Very detailed analysis, thanks. But Vrecoil is much more controllable than say Hrecoil. Sometimes I find that after putting on the HB to lower initial spread but ended up getting thrown off by Hrecoil. In that regard I think Hrecoil is acting somewhat like a pseudo-spread. Best example is PDW-R HB. I dont really find that I am much more accurate after putting on the HB, at least it feels much less than the stat would suggest.


You're right. RECOIL UP is more controllable ... because it is predictable: always up and in a fixed amount (note that you can`t "feel" the spread). The reason to have a different feeling when HB is equipped is that the proportion of both RECOIL changes. Total deflection feels more vertical... and therefore it seems (only seems) more controllable.


In any case, the effect of HRECOIL is much less important than the vertical one, on average. The average value of RECOIL UP in the game (full auto weapons) is 0.366 (beside FSM) and is IT IS ADITTIVE, while the average value of max. deflection of HRECOIL is 0.261 and IT IS SELF-CORRECTING (statistically). Only the deflection of center of HRECOIL accumulates, and this is on average 0.011, ie only 3% of the RECOIL UP.

Posts: 293

Date of registration
: Jun 9th 2012

Platform: PC

Battlelog:

Reputation modifier: 10

  • Send private message

5

Monday, February 8th 2016, 7:11am

I agree that the Hrecoil is less important than Vrecoil as the average values of Hrecoil is and Hrecoil self corrects. From my understanding, your first post is making a point in the effects of Vrecoil and spread on accuracy at different ranges, and the effects of HB as an example. The point I would like to add is that in some cases HB could be cutting down a lot of the initial spread (like PDW-R), while Hrecoil takes over spread as the source of inaccuracy because its pseudo-random.

  • "leptis" started this thread

Posts: 282

Date of registration
: Mar 17th 2015

Platform: PC

Battlelog:

Reputation modifier: 10

  • Send private message

6

Monday, February 8th 2016, 12:24pm


I agree that the Hrecoil is less important than Vrecoil as the average values of Hrecoil is and Hrecoil self corrects. From my understanding, your first post is making a point in the effects of Vrecoil and spread on accuracy at different ranges, and the effects of HB as an example. The point I would like to add is that in some cases HB could be cutting down a lot of the initial spread (like PDW-R), while Hrecoil takes over spread as the source of inaccuracy because its pseudo-random.



I agree. But only a small remark. The RECOIL H is not pseudo-random ... it's totally random.

Just to play with numbers.

PDWR at 10 m produces DMG: 265.069, bullets on target: 8.025 and effective bursts (DMG>100): 100.00% . Mean values. 10 bullets fired.

PDWR + HB produces DMG: 244.720, bullets on target: 7.821 and effective bursts (DMG> 100): 99.65% .

HB moves the results outside the sweet spot (stationary ADS) ... although at this distance is not important to the effectiveness (many bullets in the target).

The importance of individual parameters on the accuracy can be estimated "switching off" the variables (by setting the value to "0"), no rigorous method because you have to take into account the synergistic effect ... one is important because there is another.

Simple accuracy (bullets on target), the 7,821 bullets become

Spread inc = 0 ----> 8,051 (2.9% better)
SpreadBase = 0 ---> 8,008 (2.34% better)
Recoil H = 0 -------> 9,220 (17.8% better)
Recoil V = 0 --------> 9,782 (25.1% better)

Note that the ratio Recoil V / Spread is 10 (approx.) which is the number that I have given as relative importance.

HB is good at range, when the spread circle become bigger than the body, to ensure the firts shot, perhaps the second one and luckily the third one, in repited micro-bursts, not in CQB.

This post has been edited 1 times, last edit by "leptis" (Feb 8th 2016, 12:41pm)