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

Wednesday, December 7th 2016, 8:12pm

BF1 shotguns: Analysis of Damage (2). Moving & Hipfiring

I prefer to tell this as a story because the result surprises. I will call "Dispersion" to the pellets distribution and "Spread" to the variation in the effective aiming line to avoid confusion.

When I computed the results of the ADS MOVING - HIPFIRING STATIONARY scenarios (both are equal: adding an spread of 0.4 degrees to every shotguns) and HIPFIRING MOVING (adding a spread of 0.6 degrees) I to be compared with the previous ADS STATIONARY I found this:

Spoiler Spoiler



No significant changes. Almost overlapping lines. ?????????

First step: The simulator is wrong --> Full review = everything was correct.

Second Step: Analysis of concrete results. Like this (M 10 A - Comparison ADS STAT and HIP MOV):

Spoiler Spoiler



As can be seen there is a change in the shape of the distribution that affects the peak frequency (but this is always within the body up to very far) and an extension of the distribution but without more significant changes in the distribution. Dotted vertical lines in distributions show the size of the body and head at 10 m.

As the differences are in the sensitivity range of the simulator (approximately 1% for 2,000 shots) and the tails of distribution are also very sensitive to the number of simulations, I have generated a serie of 20,000 shots to further analyze changes in distributions:

Spoiler Spoiler



The blue line is the distribution (frequency of the "X" coordinate) of the dispersion of the pellets without spread, the green one is the distribution of the added spread and the red one the combined action of both. The difference between pellet dispersion (ADS STAT) and the combined curve (HIP MOV) has differences of the order of 2-3% (FACTORY) to 4% (HUNTER) in the tails of the distributions and in a small area. Black dotted lines show de difference.

A different thing would have happened if the spread of BF4 for HIP-MOV (1.75) had been used (blue line Spread =0.6, red line Spread=1.75)

Spoiler Spoiler



The result is that there are no significant differences between the three scenarios and the curves shown at first were correct. Not important changes.

An example of numerical results of the average damage value can be seen here with the relative difference with ADS STATIONARY (at 50 m, where peak values are affected the differences are already sensitive):

Spoiler Spoiler



CONCLUSION: In BF1, shotguns HIPFIRING and MOVING have a very small penalty that does not significantly influence the results. Feel free to run ... but remember that you need a very accurate aim !!!.


NOTE: During these tests two things have been found:

- An analytical formulation for the uniformly distributed Spread that would allow to calculate the results of all the weapons (not only shotguns) analytically WITHOUT simulation. The formulation is related to the equation of the circle R ^ 2 = X ^ 2 + Y ^ 2.

- The combined value of spread and dispersion can be replaced by a virtual spread with extension (SPREAD + CONE) and an exponent of bias = 0.69 * (SPREAD / CONE) ^ - 0.147 (this can be refined). This can be useful for simulate Recoil Control (CIRCLE OF ACCURACY+SPREAD) or any combined action of spread.

But those are other (and tougher) issues...

This post has been edited 1 times, last edit by "leptis" (Dec 7th 2016, 8:20pm)


Miffyli

Symthic Developer

(6,927)

Posts: 3,742

Date of registration
: Mar 21st 2013

Platform: PC

Location: __main__, Finland

Reputation modifier: 17

  • Send private message

2

Wednesday, December 7th 2016, 8:34pm

Hmm I think the actual spread while moving is bit higher, see: Moving ADS spread values (likely still applies to BF1). Sorry if the variable names are confusing :s
But anyways, even if the actual spread is higher it shouldn't have too big of an impact to damage in light of these results.

How about "one shot OHK" probabilities though? Average damage stays more or less but now the cone deviates from the aimpoint, and we know you have to be accurate with shotguns to do lots of damage.
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)
  • InterimAegis - Weapon comparisons/scoring.
  • leptis - Analysis of shotguns, recoil, recoil control and air drag.
  • Veritable - Scoring of BF4/BF1 firearms in terms of usability, firing and other mechanics.
  • pmax - Statistical analysis of BF4 players/games.
  • 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

Posts: 1,535

Date of registration
: Sep 7th 2016

Platform: PC

Location: Toronto

Battlelog:

Reputation modifier: 5

  • Send private message

3

Wednesday, December 7th 2016, 8:38pm

How about "one shot OHK" probabilities though?

Something akin to the 95% confidence level used in statistics?

Miffyli

Symthic Developer

(6,927)

Posts: 3,742

Date of registration
: Mar 21st 2013

Platform: PC

Location: __main__, Finland

Reputation modifier: 17

  • Send private message

4

Wednesday, December 7th 2016, 8:52pm

How about "one shot OHK" probabilities though?

Something akin to the 95% confidence level used in statistics?

Kind of but simpler. I was curious how "probability of dmg > 100" chart from here works with spread values.
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)
  • InterimAegis - Weapon comparisons/scoring.
  • leptis - Analysis of shotguns, recoil, recoil control and air drag.
  • Veritable - Scoring of BF4/BF1 firearms in terms of usability, firing and other mechanics.
  • pmax - Statistical analysis of BF4 players/games.
  • 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

C0llis

Up and down. Bounce all around

(3,334)

Posts: 3,100

Date of registration
: Apr 15th 2013

Platform: PC

Location: Sweden

Battlelog:

Reputation modifier: 15

  • Send private message

5

Wednesday, December 7th 2016, 11:52pm

This seems odd, it has already been established that shotguns are very sensitive to shifts in the point of aim (bad aim <=> spread here), these results seem to disagree with that.

Maybe some other measure is warranted? It's reasonable that the pellet distribution evens out _over a large amount of shots_ (pellet cone is a bigger factor than spread). When using a shotgun it isn't a great comfort that you get similar performance over a large amount of shots, because shotguns are more about doing something with very few (preferably 1) shots.

Things people said

And reading Youtube comments still gives me Turbo Cancer.

It really is quite frustrating when Helen Keller sets up her LMG in the only doorway in/out of an area.

What kind of question is that? Since when is cheese ever a bad idea?

Hardline is a fun and sometimes silly Cops and Robbers sorta thing and I think that's great. Or it would be if it didn't suck.

This post has been edited 1 times, last edit by "C0llis" (Dec 8th 2016, 1:03am)


  • "leptis" started this thread

Posts: 282

Date of registration
: Mar 17th 2015

Platform: PC

Battlelog:

Reputation modifier: 10

  • Send private message

6

Thursday, December 8th 2016, 9:37pm

Hmm I think the actual spread while moving is bit higher, see: Moving ADS spread values (likely still applies to BF1). Sorry if the variable names are confusing :s
But anyways, even if the actual spread is higher it shouldn't have too big of an impact to damage in light of these results.

How about "one shot OHK" probabilities though? Average damage stays more or less but now the cone deviates from the aimpoint, and we know you have to be accurate with shotguns to do lots of damage.


According to the WoopsyYaya formula the ACTUAL SPREAD is (modifiers by attachments and bullpup = 1):

ACTUAL SPREAD= ( SPREAD MOVING - SPREAD STATIONARY) x MOD + SPREAD STATIONARY

This means that SPREAD varies according to the speed of movement with theoretical values of the modifier MOD between "0" (you do not move) and "1" (I have never known to what type of movement this value applies). The common values of the modifier are Forward = 0.5 and Strafing and Backward = 0.5 * 0.75. The SPREAD value is always between the minimum (SpreadBase not moving) and the maximum (listed value for moving). The Actual Spread Moving value is never bigger than the listed value for moving

I have applied the maximun listed values, but I agrre than the common values (for instance strafing) will be SMALLER (not bigger) than listed values for moving. More support of the small importance of Spread.

I have not given the graphs because they only indicate that there are no significant variations . Probability DMG>100 (rough graphic):





Something akin to the 95% confidence level used in statistics?


The confidence levels is a way to mean how variable can be a value you give. For instance for the Average Values of Damage the confidence levels (assuming that the distribution is Normal, which is not true for the whole range, since at a short distance the maximum value is bounded higher by the Number of pellets x maximum unit value, regardless of the value of the standard deviation and have to be demostrated in other distances.) could be like these:



The normal distribution shown is invented to illustrate what levels mean. With an average value of 50 at 15 m, you have a probability of 99.99% to get DMG=20 but only a 0.1% to get 80. That is Average=50 WITH A CONFIDENCE OF 99.98% that the value is AVERAGE +/- 30.

The Probability DMG>100 Curve can be taken as the cut of Life Line with the different levels of confidence. Itself haven`t level of confidence: It is a measure of them. ( It can have them I we analize the systematic error of the model ,... but this is other question)

This seems odd, it has already been established that shotguns are very sensitive to shifts in the point of aim (bad aim <=> spread here), these results seem to disagree with that.

Maybe some other measure is warranted? It's reasonable that the pellet distribution evens out _over a large amount of shots_ (pellet cone is a bigger factor than spread). When using a shotgun it isn't a great comfort that you get similar performance over a large amount of shots, because shotguns are more about doing something with very few (preferably 1) shots.


I agree. I expected other results. That's why I've done all those checks.

But the clue is while in BF4 we have a secuence of Spread ADS STAT=0.1 / ADS MOV=0.4 /HIP STAT= 1.25 /HIP MOV=1.75, in BF1 is 0 /0.4 / 0.4 / 0.6. Hipfiring is not penalized.

And Spread=0.4 gives a maximun (maximun!!!!) deviation of 6.3 cm at 10 m, 12.6 at 20 m, etc...Very small and only bad in transverse direction, because up and down is not so important.

Out of curiosity I have calculated the probability that the spread moves the shot 0.2 degree to the right (3.1 cm) and it is only 19.38%.

This post has been edited 4 times, last edit by "leptis" (Dec 8th 2016, 10:01pm)