Jump to content


Accuracy rework proposal.


  • Please log in to reply
119 replies to this topic

brumbarr #1 Posted 14 January 2018 - 12:55 PM

    Lieutenant General

  • Player
  • 38626 battles
  • 6,326
  • Member since:
    07-30-2012

After reading some threads that some people find accuracy RNG to be the most anoying, I started thinking of reworking the accuracy distribution and values to remove some of that frustration. 

Before I make proposal however, I need to know what the goals would have to be, what should the perfect accuracy in this game look like?

 

What I am thinkinf for now is something allong these lines:

1) Make accuracy  more distinct, more difference between inaccurate and accurate guns.

2) Keep current % of shots hit on the move.

3) Increase the chance of shots hitting around where you aimed.

 

 

What would be the goals of a good accuracy to you? Opinions needed.



Jigabachi #2 Posted 14 January 2018 - 01:14 PM

    Field Marshal

  • Player
  • 17923 battles
  • 19,011
  • [T-D-U] T-D-U
  • Member since:
    08-12-2011
Just reworking accuracy won't help, especially since many of the current problems in the game have a big impact on it. E.g., on bigger and more open maps, you can't just preaim the obvious bottleneck, so good accuracy would be worth much more.
Besides that, changing accuracy should go hand in hand with changing gun stats, otherwise it would be the kind of one-sided "fixing" that caused all the problems.

But if you just want to play around a bit:
1. Massive misses that hit the dirt in front or a poor bird above the target are annoying as hell.
2. The accuracy should always allow active play, especially for light tanks. But right now there are way too many accurate high damage guns.
3. We need active skills that alter how accuracy works, so that there is more than just accurate or inaccurate guns. More variation needed.

Homer_J #3 Posted 14 January 2018 - 01:17 PM

    Field Marshal

  • Beta Tester
  • 28692 battles
  • 29,991
  • [WJDE] WJDE
  • Member since:
    09-03-2010

You can't make more shots hit centre and make more difference between accurate and inaccurate guns.  They tried making more shots hit dead centre and it was bad.

 

What's wrong with the current system anyway?

 

View PostJigabachi, on 14 January 2018 - 12:14 PM, said:


1. Massive misses that hit the dirt in front or a poor bird above the target are annoying as hell.
 

I'm pretty certain that's just user error.  If your circle is over the target then you will hit it.  If your circle is bigger than your target then it's your gamble.



slitth #4 Posted 14 January 2018 - 01:22 PM

    Second Lieutenant

  • Player
  • 11952 battles
  • 1,008
  • Member since:
    06-23-2011

I think what kind of accuracy you would like to improve.

If you improve the overall accuracy then it would feel like you are just shooting a rifle and not a big cannon.

And we are shooting big cannons that are hard to aim.

 

If we are talking long range accuracy, then the easiest way to improve this is to let the aim circle shrink more on more accurate guns if you are in sniper mode.

This would mimic the gunner focus and the reward of the added time on aiming.



brumbarr #5 Posted 14 January 2018 - 01:29 PM

    Lieutenant General

  • Player
  • 38626 battles
  • 6,326
  • Member since:
    07-30-2012

View PostJigabachi, on 14 January 2018 - 01:14 PM, said:

Just reworking accuracy won't help, especially since many of the current problems in the game have a big impact on it. E.g., on bigger and more open maps, you can't just preaim the obvious bottleneck, so good accuracy would be worth much more.
Besides that, changing accuracy should go hand in hand with changing gun stats, otherwise it would be the kind of one-sided "fixing" that caused all the problems.

But if you just want to play around a bit:
1. Massive misses that hit the dirt in front or a poor bird above the target are annoying as hell.
2. The accuracy should always allow active play, especially for light tanks. But right now there are way too many accurate high damage guns.
3. We need active skills that alter how accuracy works, so that there is more than just accurate or inaccurate guns. More variation needed.

 

Ofcourse changing the accuracy distribution and stats go hand in hand. 

vasilinhorulezz #6 Posted 14 January 2018 - 01:30 PM

    Second Lieutenant

  • Player
  • 22813 battles
  • 1,109
  • Member since:
    09-26-2014

View PostHomer_J, on 14 January 2018 - 01:17 PM, said:

You can't make more shots hit centre and make more difference between accurate and inaccurate guns.  They tried making more shots hit dead centre and it was bad.

 

What's wrong with the current system anyway?

The problem is there isn't a distinction in between accurate and derp guns now,

because they didn't compensate for those changes, when it comes to accurate guns.

 

I don't know if it's possible with the current engine, but I think there should be, different shot spreads for each gun, and, of course, for when it's fully aimed or not.



brumbarr #7 Posted 14 January 2018 - 01:31 PM

    Lieutenant General

  • Player
  • 38626 battles
  • 6,326
  • Member since:
    07-30-2012

View PostHomer_J, on 14 January 2018 - 01:17 PM, said:

You can't make more shots hit centre and make more difference between accurate and inaccurate guns.  They tried making more shots hit dead centre and it was bad.

 

What's wrong with the current system anyway?

 

I'm pretty certain that's just user error.  If your circle is over the target then you will hit it.  If your circle is bigger than your target then it's your gamble.

 

I worded it badly, I meant  that there would be less outliers, so shots wouldnt go very low or high, and it is possible, you just have to mess with the stats and distribution.

 

I dont mind the current system, but people dont seem to like it. This is more  for me to have some fun  playing around with distributions and variables



AliceUnchained #8 Posted 14 January 2018 - 01:39 PM

    General

  • Player
  • 38414 battles
  • 8,928
  • [322] 322
  • Member since:
    10-18-2011

View PostHomer_J, on 14 January 2018 - 01:17 PM, said:

I'm pretty certain that's just user error.  If your circle is over the target then you will hit it.  If your circle is bigger than your target then it's your gamble.

 

Wrong assessment really. As there's more to this game than just hitting the target, penetrating and damaging matters. Even when you do hit, you can bounce, ricochet because you failed to hit where you were aiming. Aiming for a lower plate you can even miss the target completely with shots dispersing below the hull; there's not a single gun in WoT which when fully aimed will be filled entirely by a lower place, or cupola for example, unless at very close distance. Calling that user error or gamble is inaccurate, flat out wrong even.

 

@brumbarr: Introducing precision next to accuracy could be an idea. Inverse distribution for inaccurate guns is another; e.g. normally 68,2% falls within +/- 1 sd, with inverse only 31,8% would disperse within the same range. Just an example, numbers could be tweaked by modifying the inverse. Would be easy to calculate for the server and wouldn't put much more strain, if any, on it I think (but anyone,  feel free to correct me on this).

 

 



Gkirmathal #9 Posted 14 January 2018 - 01:59 PM

    Lieutenant

  • Player
  • 8125 battles
  • 1,509
  • Member since:
    01-14-2013

View PostHomer_J, on 14 January 2018 - 12:17 PM, said:

You can't make more shots hit centre and make more difference between accurate and inaccurate guns.  They tried making more shots hit dead centre and it was bad.

 

 

Yes you can make it possible, in my opinion, but not how accuracy is currently implemented in BigWorld/Encore. IMO it's a too inflexible mechanic to properly balance and configure. The 2.0 sigma showed this flaw pretty well!



Bucifel #10 Posted 14 January 2018 - 02:35 PM

    Second Lieutenant

  • Player
  • 30911 battles
  • 1,373
  • [B4DD] B4DD
  • Member since:
    03-18-2013

first, yes...should be more differences between accurate guns and durps

then i think shots distribution within aim circle should increase more when you move and be more accurate when you fully aim

and last...only way to make snipers good is to improve accuracy a lot...on maximum 0,2 dispersion and less for tanks what are suposed to be accurate.

 

*honorable mentions

aim time should be totally reworked (that stupid formura wich is just wrong)

dispersion values should be first, shown in garage


Edited by Bucifel, 14 January 2018 - 02:44 PM.


Dr_Oolen #11 Posted 14 January 2018 - 02:53 PM

    Lieutenant

  • Player
  • 21813 battles
  • 1,600
  • [IDEAL] IDEAL
  • Member since:
    03-13-2012

There are imo basically 3 ways how to do it (make accuracy matter and make some sense). 

 

- The fastest solution that would require the least amount of changes is to simply  widen the range of accuracies on guns by a lot. Rather than 0.3 - 0.4, which is where 98% of all accuracies are on t10 it should be more like 0.15 - 0.5. This in practice would though only help differentiate between accurate and inaccurate tanks (brawlers/snipers) which while a good step forward (along with adding significant pen drop at distance for brawlers and no pen drop for snipers) wouldnt really fully solve the issue.

 

- 2nd option which again would not be sufficient would be keeping the accuracies more or less as they are but using 2 or more (but best be 2 for simplicitys sake and for it to be understandable for players) different distributions. When again tonks that are supposed to be "brawlers" would get more ebola distribution than they have now while "snipers" would get distribution slightly better than what it was at the point f being the best. But then the accuracy itself would be even more arbitrary stat than it is now and it would be kinda too arbitrary system for my taste.

 

- 3rd option, which imo would be the best would basically be the first option (much greater range of base accuracies) which can differentiate between "snipers/brawlers" with addition of shell distribution function which controls distribution of shells in the aiming circle based on current speed/traverse of the tank, which would serve to differentiate between tonks accurate on the move and inaccurate on the move. As imo the current dispersion stats dont do particularly well (especially since literally every tank that got buffed tha last year was basically put in the 0.08-0.14 range anyway). So basically rasha hover meds would get the same distibution at any speed (the way they are now), while tonks that arent supposed to be good at snapshots would get way worse distribution when moving, but would get superior distribution when firing stationary (or at speeds/traverses close to 0). Combination of these two changes would then allow to make any combination possible in a way that would actually work and be meaningful.


Edited by Dr_Oolen, 14 January 2018 - 02:57 PM.


Thuis001 #12 Posted 14 January 2018 - 03:23 PM

    Staff Sergeant

  • Player
  • 5967 battles
  • 466
  • [-SBN-] -SBN-
  • Member since:
    05-29-2015

View Postbrumbarr, on 14 January 2018 - 01:31 PM, said:

I dont mind the current system, but people dont seem to like it. This is more  for me to have some fun  playing around with distributions and variables

I think that what people hate about it is the fact that they can perfectly aim a shot, at a weakspot at a good distance. Fire, and then have the shell go to the other side of the tank, hitting that upper front plate instead of where you aimed. Just take a look at the video where Lemmingrush tried cupola sniping a T95 on about 400m distance, which is a fair engagement distance for a Leopard 1. He literally ran out of ammo before killing the T95. And that tank was stationary, let alone if he were to move and wiggle. You can't snipe in WoT. I mean sure, a tank can have .3 accuracy, but you still wouldn't really be able to snipe as RNG can just send your shots into oblivion for no reason. Also, something that I don't really know exactly: does, let's take .3 for the heck of it, mean that the diameter of the aim cirlcle is .3 (thus .15 to either side of the central dot) or does it mean it is .30 to either side, with a diameter of .6. In the first case WoT would  have to come up with something creative, in the second case, they could change it to the first one and see how it goes. Something else connected to this issue, and a huge difference with for example WT is the engagement range. Where in WT I basicly have a huge engagement range, with 1km being just a small part of the map. In WoT the largest map is 1.4km whide. If WoT wanted to make snipers a lot more usefull, aswell as fix a lot of other issues, then making maps that are much larger (2km or heck, even 5km) with larger teams (perhaps 20v20 on 2km up to 30v30 on 5km) on such maps you can make all classes usefull,and you can actually fight somewhere without getting butchered by top tiers. )



brumbarr #13 Posted 14 January 2018 - 03:38 PM

    Lieutenant General

  • Player
  • 38626 battles
  • 6,326
  • Member since:
    07-30-2012

View PostThuis001, on 14 January 2018 - 03:23 PM, said:

I think that what people hate about it is the fact that they can perfectly aim a shot, at a weakspot at a good distance. Fire, and then have the shell go to the other side of the tank, hitting that upper front plate instead of where you aimed. Just take a look at the video where Lemmingrush tried cupola sniping a T95 on about 400m distance, which is a fair engagement distance for a Leopard 1. He literally ran out of ammo before killing the T95. And that tank was stationary, let alone if he were to move and wiggle. You can't snipe in WoT. I mean sure, a tank can have .3 accuracy, but you still wouldn't really be able to snipe as RNG can just send your shots into oblivion for no reason. Also, something that I don't really know exactly: does, let's take .3 for the heck of it, mean that the diameter of the aim cirlcle is .3 (thus .15 to either side of the central dot) or does it mean it is .30 to either side, with a diameter of .6. In the first case WoT would  have to come up with something creative, in the second case, they could change it to the first one and see how it goes. Something else connected to this issue, and a huge difference with for example WT is the engagement range. Where in WT I basicly have a huge engagement range, with 1km being just a small part of the map. In WoT the largest map is 1.4km whide. If WoT wanted to make snipers a lot more usefull, aswell as fix a lot of other issues, then making maps that are much larger (2km or heck, even 5km) with larger teams (perhaps 20v20 on 2km up to 30v30 on 5km) on such maps you can make all classes usefull,and you can actually fight somewhere without getting butchered by top tiers. )

 

I  dont think any tank in the game should be able to reliably hit a small cupola at 400m though. That would be horrible for the game, it could be better but you certainly shouldnt expect that.

gunslingerXXX #14 Posted 14 January 2018 - 03:39 PM

    Lieutenant

  • Player
  • 11023 battles
  • 1,993
  • [GUNSL] GUNSL
  • Member since:
    11-16-2014
Personally I would like to see a bigger difference between snapshot and fully aimed. Snapshot should hit less often as I think the mechanism is not ok. 

brumbarr #15 Posted 14 January 2018 - 03:41 PM

    Lieutenant General

  • Player
  • 38626 battles
  • 6,326
  • Member since:
    07-30-2012

I have a question what is mean by snapshots? 

Do you mean shooting on the move at 60kph or do you mean  not fully aiming your shot but shooting a bit before its fully aimed?



HaZardeur #16 Posted 14 January 2018 - 03:49 PM

    Second Lieutenant

  • Beta Tester
  • 33294 battles
  • 1,180
  • Member since:
    08-14-2010

View Postbrumbarr, on 14 January 2018 - 12:55 PM, said:

After reading some threads that some people find accuracy RNG to be the most anoying, I started thinking of reworking the accuracy distribution and values to remove some of that frustration. 

Before I make proposal however, I need to know what the goals would have to be, what should the perfect accuracy in this game look like?

 

What I am thinkinf for now is something allong these lines:

1) Make accuracy  more distinct, more difference between inaccurate and accurate guns.

2) Keep current % of shots hit on the move.

3) Increase the chance of shots hitting around where you aimed.

 

 

What would be the goals of a good accuracy to you? Opinions needed.

 

The only way I can think of ( and maybe WG too ) would be a rework on crew skillz that influence positive accuracy RNG.

Edited by HaZardeur, 14 January 2018 - 03:50 PM.


Laatikkomafia #17 Posted 14 January 2018 - 03:54 PM

    Brigadier

  • Beta Tester
  • 21315 battles
  • 4,245
  • [ELC-P] ELC-P
  • Member since:
    12-27-2010
Why bother making guns more accurate when there are no actual weakspots to shoot anymore? Tapping 2-key is the best accuracy buff in the game.

Edited by Laatikkomafia, 14 January 2018 - 03:55 PM.


Hiisi #18 Posted 14 January 2018 - 03:56 PM

    Sergeant

  • Player
  • 14926 battles
  • 265
  • [-PJ-] -PJ-
  • Member since:
    03-21-2011

Accuracy diffecence is just too small. Leopard 1 is the most accurate medium and it has 0.30. Is-7 accuracy is 0.4 for example. The difference is way too small, it`s only 0.1. 

 

At minimium difference it should be 0.2. Leopard should have 0.25 accuracy and is-7 should have 0.45. 

 

 



Thuis001 #19 Posted 14 January 2018 - 04:07 PM

    Staff Sergeant

  • Player
  • 5967 battles
  • 466
  • [-SBN-] -SBN-
  • Member since:
    05-29-2015

View PostDr_Oolen, on 14 January 2018 - 02:53 PM, said:

 

- 3rd option, which imo would be the best would basically be the first option (much greater range of base accuracies) which can differentiate between "snipers/brawlers" with addition of shell distribution function which controls distribution of shells in the aiming circle based on current speed/traverse of the tank, which would serve to differentiate between tonks accurate on the move and inaccurate on the move. As imo the current dispersion stats dont do particularly well (especially since literally every tank that got buffed tha last year was basically put in the 0.08-0.14 range anyway). So basically rasha hover meds would get the same distibution at any speed (the way they are now), while tonks that arent supposed to be good at snapshots would get way worse distribution when moving, but would get superior distribution when firing stationary (or at speeds/traverses close to 0). Combination of these two changes would then allow to make any combination possible in a way that would actually work and be meaningful.

This means that WoT would actually implement a gun stabilization system. Such a system means that even if you drive fast your gun still points where you're aiming with little to no difference with a fully aimed shot as the gun is hold in place. To point out a few tanks that had this kind of mechanic in rl. Cent III and up (so Cent I with top gun and turret would get it, stock Cent not though) this would make tanks a lot more accurate on the move (perhaps, besides buffing the accuracy of most lights, giving them stabilizers on higher tiers would be really nice) here is a video from 1952 showing the Cent III with a fully stabilised gun https://www.britishpathe.com/video/bang-on-always the gun did literally have no dispersion values from then on, as the mechanism kept the gun exactly on target, negating any and all dispersion caused by moving, rotating and traversing. And while this is really nice, not that many tanks have it, making those that have it truly special. Tanks that could get it on higher tiers: Cent I and up (Cent I only with the top gun and turret) Obj 430, STB-1 and some others. (The Leo 1 could get one if they were to change it into a slightly more modern version) I really think this is worth considering tbh as it would be some really interesting mechanic.



_T_1_T_4_N_0_ #20 Posted 14 January 2018 - 04:08 PM

    Second Lieutenant

  • Player
  • 31028 battles
  • 1,114
  • [SQCI] SQCI
  • Member since:
    02-17-2015

Accuracy difference is too small but rather than change every tank just make dispersion non linear.
100m disp x1

200m disp x1.1

300m disp x 1.2

400m disp x 1.3

500m disp x 1.4

Doesn't have to be those values but something along those lines and go back to standard gausian tight center.
Different dispersion model for snapshots and a swap over point in aim from snapshot inverse gaussian to gaussian.

Wot is a compressed battlefield with compressed ranges so disp should also be non linear to act compressed also.
 


Edited by DumbNumpty, 14 January 2018 - 04:10 PM.





1 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users