Range vs Accuracy: Let's improve it.

[Iluvalar]

Reload times can be adjusted. I did some playtesting (with speed increased...),

How do you plan to adjust it ?! You broke the other parameter with which we can tune it.

Playing with speed increased doesn't count for 2 reasons :
A) We are talking about the "feel" of the game, since you demonstrated that the balance should be kinda similar.
B) NullBot happens to kill me, even at normal speed once in a while. I don't want to insult your skill, but I doubt you can play a competitive match in speed game. You can't test the balance with weakened AI or by cheating.

[Iluvalar]

Double post sorry.

Situation :
A) A tank is coming, your lancer could kill it in 1 shot but your accuracy is 2/3 of chance to hit. There is 1/3 of chance that he survive 1 round and 1/9 more chance that it survive another one after... You can expect to lose half a round of damage against him.

B) A tank is coming. BOOM you have 100% chance to kill him. No damage (despite the fact you will need to reload longer, but who cares ?).

How this 2 situations are supposed to be the same ?

[Stratadrake]

right now it target at 2 ticks per inaccuracy in a 360° circle around the enemy. a 70% accuracy weapon have 30% of innacuracy so the projectile will lend 60 ticks away from the CENTER of the target.

What unit of measure is a tick?

Otherwise, are you saying that the projectile's target location is picked based on adding a (random-angle) distance vector to the target location? I think misses would be more realistic if they were deviations from the turret firing angle, not target location.

By they way, what exactly those [short/medium/far] buttons did?

I explained that in an earlier post. Short means Short. Long means Long. "Optimum" does NOT mean "somewhere inbetween" or "medium", it means "decide for yourself either Short or Long" - which is a very subtle cheat (in your favor) btw. Really only rocket weaponry gets "optimum" = "long range"(another reason why Lancers are so awesome), for most other weapons (cannons/etc.) "optimum" = "short range".

[Iluvalar]

right now it target at 2 ticks per inaccuracy in a 360° circle around the enemy. a 70% accuracy weapon have 30% of innacuracy so the projectile will lend 60 ticks away from the CENTER of the target.

What unit of measure is a tick?

Otherwise, are you saying that the projectile's target location is picked based on adding a (random-angle) distance vector to the target location? I think misses would be more realistic if they were deviations from the turret firing angle, not target location.

The smallest unit of measure, how do we call it ? the one there is 128 in one tile used in the pie files .

Yeah, I think that's what I am saying. And yes, basing the miss proportional to the distance like in my patch make it much more like an angular devation. At least on X/Y axis it does.

[Per]

How do you plan to adjust it ?! You broke the other parameter with which we can tune it.

I didn't break it, I put a knife in its back...

As usual, it can be adjusted by changing damage. Yes, this requires a lot of re-balancing. I'm not saying it will be easy, only that it is possible. This is in any case not meant to be a mere short-term quick-fix.

Situation :
A) A tank is coming, your lancer could kill it in 1 shot but your accuracy is 2/3 of chance to hit. There is 1/3 of chance that he survive 1 round and 1/9 more chance that it survive another one after... You can expect to lose half a round of damage against him.

B) A tank is coming. BOOM you have 100% chance to kill him. No damage (despite the fact you will need to reload longer, but who cares ?).

How this 2 situations are supposed to be the same ?

Most of the time you require a lot of shots to kill a tank, and randomness will be averaged out. So in practice this is not much of a problem.

[Iluvalar]

How do you plan to adjust it ?! You broke the other parameter with which we can tune it.

I didn't break it, I put a knife in its back...

As usual, it can be adjusted by changing damage. Yes, this requires a lot of re-balancing. I'm not saying it will be easy, only that it is possible. This is in any case not meant to be a mere short-term quick-fix.

And at the end, not a single weapon would be recognizable as a warzone2100 weapon.

How about starting anew with a game about poneys and goblins ? Would be easier .

Most of the time you require a lot of shots to kill a tank, and randomness will be averaged out. So in practice this is not much of a problem.

That's not quite true. Rockets can kill in very few shots. Even if it's 3 or 4 it still mess up the balance.

[JJjopando]

I am a war veteran and weapons do not always function properly nor aim perfectly. Especially in a wet marsh in VIETNAM.
Stick with some element of random malfunction. That is what happens in war.

[Stratadrake]

right now it target at 2 ticks per inaccuracy in a 360° circle around the enemy. a 70% accuracy weapon have 30% of innacuracy so the projectile will lend 60 ticks away from the CENTER of the target.

What unit of measure is a tick?

Okay, I read you now. So the question is, if you have a weapon whose rated accuracy is close to 100%, the shot deviation is still within the target's hitbox (as derived from body statistics and/or PIE geometry), the projectile may land inside the hitbox and inflict a "hit" regardless of diceroll?

Okay, if that's the current behavior - hitbox overriding diceroll - that is definitely a bug, but how severe depends on metrics we don't exactly have quantized.

[Iluvalar]

hitbox problems are solved assuming the units are a simple cylinder. Based on PIE data.

I was hoping we could solve the bug with that patch without delving into the code to find out how exactly the radius of the unit is calculated...

But we just saw that about half the weapons on the game can trigger on the python body with or without a certain amount of upgrade.

[Andrie]

This is very interesting, what is a canons accuracy?

[Iluvalar]

Can' t you click on the FAQ tab on top of the page and find the turrets information ?

[iap]

Hi, please consider this suggestion to solve tje accuracy problem:
Instead of predetermined dice role use the accuracy value as "spread " , I mean in the sense of deviation from the correct angle.
This way you take into account the distance from the target and the body size without even calclating them. A close target will have more chance to get hit, and this is logical like in real life!

The formula I'm suggesting is this:
Consider Angle is the correct hit angle and Accuracy is the accuracy defined to this weapon by the distance (like now).

Angle = Angle - random((100 - Accuracy ) * BaseSpread ) -BaseSpread / 2

BaseSpread can be an arbitrary angle, like 90 but Maybe even a researchable item that will affect all the weapons.

In any case I wouldn't have suggest this if you didn't mention that there is already a hit test on top of the accuracy.

[Iluvalar]

Angle = Angle - random((100 - Accuracy ) * BaseSpread ) -BaseSpread / 2

iap, please, understand that this real life equation also imply the range of the weapon. That would make a small cyborg difficult to hit beyond 4 tile range, but the factory hitable up to 35 tiles away. And we don't want this to happen.

Please read this topic : viewtopic.php?f=42&t=9645

[iap]

Ok, I could have say "yes, but..." but there is no need for this now. Is there any way I could mod this change without recompiling the C code?

(I do have real arguments for the "yes, but")

[Iluvalar]

No, we need to edit C code.

ps: note that my patch use something better than your formula to draw the missed shot, so at least it look better.