Loading...
  OR  Zero-K Name:    Password:   

bug in range(s) declared as shperical

25 posts, 535 views
Post comment
Filter:    Player:  
Page of 2 (25 records)
sort
In one relatively old game 3 vs 3 i've build stinger first, but enemy built it in another height (as i remember, on lower ground), and it flawlessly fried my!
And we lost the whole game due to this develobstering. Otherwise my team would outporc, as map was tiny.
+0 / -0
Bug: spherical ranges are reported as spherical


The real problem here is that in Spring, units shoot at a specific point inside the target unit known as the midpoint, typically from another, different point, usually at the tip of their gun.

For Stinger, the gun is above the midpos; so it's geometrically quite simple to arrange two stingers in such a way - like on the picture - that one can reach the other Stinger's midpoint without having its midpoint reached in return.

quote:
And we lost the whole game due to this develobstering

Your team lost because you're lobsters who cannot into geometry.
+4 / -0
Many years ago i've read: if a user could not handle a program, then it is a guilty of it's programmer.

Where in-game info about mid-points? So "lobsters" should be addressed not to players in this topic.

Don't remember in which game, but there was an option "show damage boxes" or "... spheres".

Anyway, for ray guns hit points/areas should be adjusted, to exclude described bullshitty situations. May be as separate way of processing for ray gun vs ray gun situations, leaving some "default" damage points at an old places for any other combination of shot kinds.

About your demagogy - example from real world: boeing 737max crashed at least two times due to develobsters, not due to pilots who just did not know about develobstering.
+1 / -0
2 months ago
quote:
something something hitboxes and spheres


Alt+b for collision volumes, midpoints, aimpoints etc.

quote:
a million aimpoints


re-geometry the situation so that one and only one stinger can't touch the other's colvolume.
+2 / -0
quote:
May be as separate way of processing for ray gun vs ray gun situations, leaving some "default" damage points at an old places for any other combination of shot kinds.

If you add another target point for every unit, you will effectively double the amount of targets to consider for every targeting calculation, which will square the performance cost thereof, and then you will cry about how "develobsters" caused the game to have 0.2 FPS or something. Truly resignable!

What's worse, it will be for no gain, because you'll just end up with this situation:


quote:
Anyway, for ray guns hit points/areas should be adjusted, to exclude described bullshitty situations.

There are two way to do this:

1) Make every ray gun's target point be identical to their target point. This means that things whose guns are high enough may be immune to Fleas because the target point is too high up and flea has a ray gun. I guess you could also make Stinger lower to fix that. And make all of their collision volumes be spherical as well. And rename ZK to Sphere-K.

2) Have all Stingers' true range be like 200 more than it is, but hack it so that they won't fire at anything outside of their current range. Whenever a Stinger obtains a targeting solution on another Stinger, allow the target to retaliate with full range against that specific attacker. Because fuck physics.

This will work only after the attacking Stinger makes the first shot, so in a typical Stinger duel situation, it will change nothing.

quote:
About your demagogy

You know who else accused everyone of demagogy? HITLER.
+4 / -0

2 months ago
quote:
2) Have all Stingers' true range be like 200 more than it is, but hack it so that they won't fire at anything outside of their current range. Whenever a Stinger obtains a targeting solution on another Stinger, allow the target to retaliate with full range against that specific attacker. Because fuck physics.


We're already fairly close. It's disallowed to give Stinger manual attack commands near their maximum range, since this could be used to manually target the edge of the collision box. It's disallowed to give any unit attack commands above ground level, since this could abuse physics into firing beyond maximum unit range. Additionally, Stingers will retaliate against targets they can't technically see (outside LOS) if they get fired upon.
+3 / -0


2 months ago
quote:
Additionally, Stingers will retaliate against targets they can't technically see (outside LOS) if they get fired upon.

This isn't unique to Stingers though; any unit firing within AirLos is revealed (except Phantom. And subs outside sonar). That's okay.

quote:
We're already fairly close. It's disallowed to give Stinger manual attack commands near their maximum range, since this could be used to manually target the edge of the collision box.

Sad!

+2 / -0
Why double calculations? I wrote about another processing. Overhead would be only in form of "if","case", etc.

About re-making may be i did not wrote clearly: i meant that hit points for ray shoters shoud be the same as shot "sources", if ray-to-ray.

But about limitting of firing (initiating of firing) as was described by conversator(s) may be it is not bad. I need to test/spectate.
+1 / -0

2 months ago
This will make melee and short-ranged units unable to attack large units or turrets, even if they can touch them they wouldn't be able to shoot them.
+0 / -0
quote:
Why double calculations? I wrote about another processing. Overhead would be only in form of "if","case", etc.

Fair. That's slightly less bad, but still has issues with short-ranged units and ray origins being attached to rotating things.

quote:
About re-making may be i did not wrote clearly: i meant that hit points for ray shoters shoud be the same as shot "sources", if ray-to-ray.

Also guns should have zero length in that case, or this happens:

+10 / -0
2 months ago
Oi if we're talking about bs with stingers, let me air a grievence. Light radar sight range exceeds that of stinger and stinger attack range, so I often build aradar on the edge of stinger range to get vision. But, sometimes it dies to the stinger 30 seconds later. What's with such a delay? If enemy gets aradar dot for my radar, and it ocilates in acirtain way, stinger can fire at the max range radar dot. And sometimes when it fires, the Lazer shots through, overshooting max range and killing my radar. Super annoying that the radar death is a function of rng and time.
+0 / -0
2 months ago
#metoo

AI has once managed to outrange my Stinger with a Fencer :(
+0 / -0

2 months ago
Every turret should have LOS slightly over its range so they can't abuse radar dots. It's ridiculous how turrets get more range by making sure the enemy is NOT in LOS.
+2 / -0
2 months ago
Giving units more targeting points wouldn't solve the problem completely but would reduce it, would it really increase the CPU load that much? I could be wrong but I'd think the physics simulation side is way heavier than the targeting.
+0 / -0
Hmm what if guns had flag to use the midpoint of unit body for range calculation? That could fix edge case models like Stinger.

btw great ilustrations :D
+0 / -0
2 months ago
quote:
float targetBorder  default: 0.0 (1.0 for #Melee) 

Controls whether the weapon aims for the centre or the edge of its target's collision volume. Clamped between -1.0 - target the far border, and 1.0 - target the near border.


+0 / -0
*comment removed as already discussed*
+0 / -0

2 months ago
quote:
AI has once managed to outrange my Stinger with a Fencer :(

As far as I know Fencer has cylindrical range so this is entirely intended.

quote:
Many years ago i've read: if a user could not handle a program, then it is a guilty of it's programmer.

This is almost entirely inapplicable to computer games. If it does not take some time to learn to get good at playing a game it probably wasn't a very interesting game.
+2 / -0
2 months ago
Steel_Blue: Are you saying that units are already capable of targeting model borders if desired? If so, could the current behavior of stingers be adjusted, maybe?
+0 / -0

2 months ago
Any problem with changing the midpoint to a sphere or a cone (if you ignore y pos) depending on model instead of a point?

It would not completely solve the problem but make it far more rare.
+0 / -0
Page of 2 (25 records)