8 replies to this topic

[Greyhart]

does all the damage have to be put on one pixel?

a technical and a philosophical question.

A lot of the convergence / cone of fire stuff is about moving from one pixel to another but does damage need a single pixel?

[LordNothing]

no, mainly because collision detection algorithms dont care about pixels. their much more accurate than that.

[Monkey Lover]

Maybe server load might take a very big hit.

[Quicksilver Aberration]

LordNothing, on 11 February 2016 - 05:53 AM, said:

no, mainly because collision detection algorithms dont care about pixels. their much more accurate than that.

In some cases (PPCs) they are actually "less accurate" (simply because of overly big projectiles), the number of times I've seen a PPC collide with terrain while a laser makes it through the same area is disappointing....

Edited by Quicksilver Kalasa, 11 February 2016 - 08:27 AM.

[Ultimax]

Is damage actually hitting a single pixel?

All of the time, with all weapons?

Really?

Please, demonstrate proof of this assertion.


Good luck.

Edited by Ultimax, 11 February 2016 - 10:42 AM.

[Sandpit]

Greyhart, on 11 February 2016 - 05:02 AM, said:

does all the damage have to be put on one pixel?

a technical and a philosophical question.

A lot of the convergence / cone of fire stuff is about moving from one pixel to another but does damage need a single pixel?

No
Damage is based on hit location and hit boxes. If you hit anywhere inside of a hitbox, you'll do damage to that specific area of the mech.

However, with less convergence, it's a lot easier to spread damage as well as prevent every weapon from firing at the exact same spot instantly. There's a few different ways it could be handled, but it's the single worst issue when it comes to damage balancing, always has been.

It's what started off the whole Ballistics are OP, buff lasers debate going all the way back to CB a few years ago.

That's what some of us mean by "band-aids". PGI has tried several things to "adjust" this dynamic over the years but they've never worked. Mainly because they continue ignoring the underlying issues that create the imbalance in the first place in favor of easier "fixes" that don't really fix the issue, just help cover it up.

[LordNothing]

Quicksilver Kalasa, on 11 February 2016 - 08:27 AM, said:

In some cases (PPCs) they are actually "less accurate" (simply because of overly big projectiles), the number of times I've seen a PPC collide with terrain while a laser makes it through the same area is disappointing....

this game probibly has one of the worst collision detection schemes ive ever seen. its very sensitive to rounding errors, and thats why mechs cant go faster than 200kph and why projectiles cant go faster than 2000kph. it invites rounding errors. it reeks of single precision floating point values. it doesn't slow things down to use double precision, even on the old x87 fpu. doubles are very cheap on modern cpus, even 64-bit potatoes.

lasers are all based on ray-box collisions, which are stupid easy and hard to screw up, and thats why lasers never screw up. sphere-box collisions arent that hard either but i guess it doesnt work when velocity is high. throw in hsr and you get more rounding errors. i presume simply making the ppc collision sphere bigger or throwing in a little bounding box overlap in the mech's boxes. its not just ppcs, ive seen it happen on ac2s and narcs and at least once with the gauss rifle. but there are lots of things you can do to reduce the effects without touching code.

Edited by LordNothing, 11 February 2016 - 08:41 PM.

[Quicksilver Aberration]

LordNothing, on 11 February 2016 - 08:39 PM, said:

thats why mechs cant go faster than 200kph and why projectiles cant go faster than 2000kph.

I'm pretty sure projectiles not going faster than 2000kph is by design choice not because of engine limitations. This doesn't smell of single precision problems, it smells of low server tick values and perhaps some HSR edge cases that aren't accounted for.

[LordNothing]

Quicksilver Kalasa, on 11 February 2016 - 08:54 PM, said:

I'm pretty sure projectiles not going faster than 2000kph is by design choice not because of engine limitations. This doesn't smell of single precision problems, it smells of low server tick values and perhaps some HSR edge cases that aren't accounted for.

perhaps. reguardless debugging required.

still reworking collision boxes and resizing the collision sphere for the weapon might be viable work arounds.