[Coding] Realistic Deviation [WIP]

[nedlands1]

I have feeling the whole system is a little more complex still, what about the recoilforce modifiers (I think I have seen something with a name like that, cant remember it exactly though).

I think it could be that this determines the highest point the barrel climbs to, and that the other recoil modifier determines the position it stops at on the way back down. Although I am only guessing about that, I have not had a look at the system at all. It could be totally different.

It is my understanding that there are two recoil systems. One for handheld weapons (ObjectTemplate.createComponent SoldierBasedRecoilComp) and one for mounted weapons (ObjectTemplate.createComponent WeaponBasedRecoilComp). From my experience you can't mix and match the two groups of modifiers (eg ObjectTemplate.recoil.recoilForceUp CRD_UNIFORM/1.8/1.8/0 won't work on a mounted weapon). ObjectTemplate.recoil.recoilForce is probably the modifier that you are thinking of and it happens to be from the mounted weapon category. What it does is apply a set force in the opposite direction to the projectile.

[Wolfe]

In PR, infantry combat has 1 major problem: hand-eye coordination; the ability for a player to near-instantly move their mouse over the target and click.

You cannot use recoil as a means of preventing this. Sure, it prevents full-auto accuracy, but recoil has minimal effect in single-shot accuracy. In order to make it worth while, the recoil would have to be increased higher than a player's hand-eye coordination ability to return it to center which is nearly impossible; there are too many players who can too easily compensate for it. Moreover, such a large recoil doesn't fit will with low recoil weapons such as the M16.

The only solution is a system that is out of the player's control; one that forces each player (regardless of hand-eye coordination twitch skill) to pause, aim, then fire. Weapon sway is an excellent solution but unfortunately that ability doesn't currently exist so we have to use the next best thing: simulated weapon sway through turning, walking, and firing deviations. This is an excellent alternative because the only way to reduce deviation is through time. As a result, the deviation recovery times become the most critical factor; one that I'm surprised that nobody has even mentioned yet.

If the recovery times are too short, it allows exactly what we see in v.75 where the per-shot recovery time is 0.33 seconds. That's 3 bullets per second with perfect accuracy. That's insane. If the weapon is fired faster than .33 second per bullet, the recovery time only rises to a maximum of 0.67 seconds; 1/2 a second per shot with perfect accuracy. In v.70, the per-shot recovery time was only marginally better with min/max of 0.33/1.33 seconds where the third and fourth shot became less accurate.

Combine that with small cones of fire. At long ranges, the challenge to hit someone is minimal. At short ranges, where most combat occurs, the challenge to hit someone is... well.. there is no challenge. Just run and rapidly click your mouse. Whoever has the fastest ping and split-second reaction time wins. Some people call that skill. Maybe mouse skill.. but it's not even remotely close to any kind of realistic warefare skill that PR should promote; where victory goes to the team with better positioning, not the team with better twitch.

To prevent both long and short range twitch run'n gun, you need a wider cone of fire. If the cone of fire is too big you'll shoot your feet or the guy next to you. Nobody wants that, so the cone has to be narrow enough to hit in the general area where you're aiming but wide enough to prevent run'n gun. Whatever that cone is, people will still complain how "unrealistic" it looks which is exactly why a visual indicator is needed. If done correctly, a wider cone of fire is more acceptable because players can SEE it rather that guessing where it is. The HAT is a good example; nobody complains about the 20+ deviation because they can see the progress of their deviation and know exactly when it comes to a point. Small arms needs the same thing, except we have to use something more natural than a big, moving crosshair.

This brings us to the current state of this project: Longer recovery times, wider cones of fire, and a visual indicator that displays the two. The idea of slightly-blurred vision around your target that widens and narrows could work well and provide an excellent balance between realistic looks and function.

P.S.:
Enter the above values into the HAT, then test again. Pretend the HAT is the M16 and you'll see what I mean; the wider cone of fire looks and feels much better with a visual indicator.

[Drav]

I disagree Wolfe. I say the cone of fire needs to be narrower if anything, but the recovery time for perfect accuracy should be longer.

This would mean the weapons react acceptably at sub 50m ranges, where hitting a target really isnt very difficult, while needing to settle for longer shots.


Personally, I'd like to get some headcams and get UK_Force and a couple of the ABA guys to do a couple of advance and shoot runs on a 100m electronic target range. Then we could answer sooo many questions, how accurate are guns when you're running around, whats the recoil on automatic like etc. From my experience you can hit a man sized target at up to 50m with a very quick aim. So boys, up for it??

[CAS_117]

I have a solution. You know how tanks, when when at a high magnification are somewhat jumpy in their ability to aim? Compared to snipers they are much harder to control at long range. This is caused by a zoom modifier which when added, makes weapons have less movement, this being both recoil and the ability to compensate for it. Tanks and APCs however don't have this, which is why they're so hard to aim when zoomed. If zoom modifiers are removed from weapons, they will only be able to move in small, miliradian jumps. A better example would be to compare the recoil on the irons sight and scoped weapons. Scoped weapons have less recoil, and are less sensitive to control input, making compensation more or less the same as irons. But the difference is that its much harder to adjust for recoil at a higher sensitivity, even though you move the mouse the same distance on your computer; you have much more sensitivity and much more chance of screwing up.

[Drav]

I quite like that idea. Can you alter the zoom modifier so they arent quite as jumpy as the tanks, or is it on or off? And unfortunately the major flaw is the dorks with adjustable sensitivity mice can get around this easily.....

[Wolfe]

I say the cone of fire needs to be narrower if anything... This would mean the weapons react acceptably at sub 50m ranges, where hitting a target really isnt very difficult

But that's exactly the problem. Killing at short ranges is too easy. The result? Players run directly into combat so that any deviation is negated. This is exactly the definition of run'n gun and exactly what we're trying to avoid.

Personally, I'd like to get some headcams and get UK_Force and a couple of the ABA guys to do a couple of advance and shoot runs on a 100m electronic target range. Then we could answer sooo many questions, how accurate are guns when you're running around, whats the recoil on automatic like etc. From my experience you can hit a man sized target at up to 50m with a very quick aim. So boys, up for it??

I'm going to be very blunt here; not to be rude but to get the point across. It doesn't make a damn bit of difference what happens in the real world. You can have all the realism in the world but it doesn't matter if that realism allows crappy run'n gun gameplay. Gameplay MUST come first and molded as much as possible in realism but without sacrificing gameplay.

[Wolfe]

CAS, that isn't going to work for two reasons:

1) Increasing sensitivity only helps long range (beyond 100 meters), not short range where most infantry combat takes place.

2) People have mouse tools that allow them to modify sensitivity on the fly. This would create a huge and unfair gap between those who had such a mouse, and those who don't.

[Drav]

Ummm.....ye it does. If you jog across a field and target pops up, the time it takes to pull up your sights and fire into a man sized target in real life should be the same in PR. To suggest anything else is ridiculous.....

To deliberately make the shooter inaccurate at close range is very frustrating. Close range with an automatic weapon, I actually think I can shoot better than the soldiers in PR at present. By all means make the scopes more sensitive as CAS suggested, or make them take longer to shoulder, but I'd like my bullets to go roughly where they're pointed when I get in a position to fire. Of course this is different at long range and while moving, but at sub 50m ranges, weapon up and stopped, I think the deviation should be a bit less than it is now.

[Masaq]

he HAT is a good example; nobody complains about the 20+ deviation because they can see the progress of their deviation and know exactly when it comes to a point.

Wrong, the accept the 20+ deviation because they're aware that the long steady time is a representation of the time it takes to deploy wire-guided shoulder-launched anti-tank weapons.

They won't accept a 20+ deviation on the M16 because there is no possible way that a gun aimed perpendicularly to the floor will hit the ground 5m away from the soldier's feet.

Having an AT round blow up in your face because you've tried to fire it quickly is one thing; having a 5.56 fall out of the muzzle and flop to the ground is quite another.


Mosquil - apologies, it wasn't too clear it was sarcasm.

Everybody else - back on topic please.

[Wolfe]

Masaq;700770']They won't accept a 20+ deviation on the M16 because there is no possible way that a gun aimed perpendicularly to the floor will hit the ground 5m away from the soldier's feet.

Who said the M16 has a 20+ deviation? The max deviation for the M16 is about 6, and that's only if you are unscoped, strafing, quickly turning, and firing wildly. In the real world, any human performing that combination is going to have horrible aim. In v.75, that same combination guarantees you 90% accuracy, and that is unrealistic. And for the record, if you're scoped, the max deviation is 1.8, not 20+.

The HAT example was used simply to illustrate that a visual indicator helps players accept a wider cone of fire.

The purpose of this thread was to give people hope that run'n gun prone spam might come to a stop, and post my progress along the way. In addition, I have repeatedly stated this is a work in progress and that the latest deviation model is simply an outline, not the final product. Deviation is clearly a touchy subject for some and it may be necessary to continue development behind closed doors to prevent it from being annihilated before it has a chance to succeed.

[fuzzhead]

Please dont turn this into a ***** fest... this is really intersting and badly needed work that Wolfe and others are volunteering their time towards. Its about experimenting different ways in helping PR's firefights feel more natural and realistic. So please keep your posts positive and constructive instead of negative and negating possible avenues.

We can all agree here that soldiers in PR do not suffer from the same things that real soldiers suffer from: fatigue, "fog of war", manuvering a weapon in CQB, breathing (especially when tired), morale, etc etc. We can also agree that the firefights in PR are fun, but could definitely be improved upon.

We have to work through the confines of the BF2 engine to bring realistic gameplay. We want players to simulate real life tactics as much as possible and those tactics should be EFFECTIVE in game. This means players should NOT be acting like super humans with instant accuracy who charge across the battlefield completely fearless, with no thought towards their own virtual lives. So yall need to keep that in mind when handling the topic of weapon accuracy.

Please keep your minds open and be HELPFUL, not HURTFUL. You can be honest and have an opinion, but do not attack each other, that goes for all parties.

Thanks everyone.

[CAS_117]

CAS, that isn't going to work for two reasons:

1) Increasing sensitivity only helps long range (beyond 100 meters), not short range where most infantry combat takes place.

2) People have mouse tools that allow them to modify sensitivity on the fly. This would create a huge and unfair gap between those who had such a mouse, and those who don't.

Having higher sensitivity with a 4 power scope will have plenty of effect at close range. Also the hardware people use with their game is completely out of our control, and increasing the sensitivity will:

A) balance scopes and irons.

B) reduce the ability to stabilize on a target at short ranges.

If for no other reason than balancing the iron sights and scoped weapons this should be done regardless of the advantages/ disadvantages it will have on prone spam or run and gun or what have you.

[Wolfe]

the hardware people use with their game is completely out of our control

That is exactly the reason why deviation should not be based on it. There is already an unfair playing field with suppression effect shaders and to broaden that concept into deviation would be catastrophic.

In regard to suppression shaders, a possible solution is to decrease the duration 4x, but increase the effect 4x. By doing so, both fast and slow computers would mostly be suppressed by rapid/automatic fire which is how suppression should work, rapid fire, not individual ricochets that suppress for 6 seconds or longer.

[Drav]

CAS, I really like your idea, all the magnifying scopes I have used seem seem much more sensitive to movement than ironsights (obviously!) but I have to agree with Wolfe here. The advantage for those with adjustable sensitivity would be enormous, so much so that I could see it being buy a special mouse or stop playing situation....

Like the suppression idea too....

[Masaq]

Having higher sensitivity with a 4 power scope will have plenty of effect at close range. Also the hardware people use with their game is completely out of our control, and increasing the sensitivity will:

A) balance scopes and irons.

B) reduce the ability to stabilize on a target at short ranges.

If for no other reason than balancing the iron sights and scoped weapons this should be done regardless of the advantages/ disadvantages it will have on prone spam or run and gun or what have you.

And for £15 anyone can buy a mouse with adjustable dpi and simply slow the mouse right down so they can stay tracked on someone.

It's an even easier fix than splashing out £100 on a graphics card with enough power to turn a 3 second supression shader into a 0.3 second shader.


It'd be unfair and ultimately, pointless.

[Wolfe]

Updated original post with new deviation numbers.

[Conman51]

Necro Time!!


So what happened to this? was it already implemented?

The Op, Wolfe, hasn't logged on in a long time

hope this project isnt dead

Was this supposed to be a replacement for deviation?

[Nitneuc]

IIRC this thread is about 0.8 deviation, which is now replaced by 0.85-0.87 deviation and about to be replaced by 0.9 deviation.

[AnRK]

This level of detail was never implemented (at least in full from what I remember) because of the issue of tracers. If you search for "deviation" and "tracers" you'll find some answers, very unfortunate given all the work put in, although in the time since it was discussed they might well have found a way round the problem but I wouldn't get your hopes up.