231 replies to this topic

[Cyborne Elemental]

Yeah Monkey, this is exactly what is happening for fast moving objects in the game, the hitbox of that spider has already moved enough out of the way to completely miss those shots.

Not even its arms would have taken damage because its hitbox was completely desynced and leading its direction of movement by about 2 mech lengths there, which is a close call, but just enough to clear all shots.

Notice that he does take the laser damage from your teamate up higher on the hill because at that point he is at a near deadstop where the hitbox and his geometry are close enough together to allow some reg.

The geometry you see there is not what really registers damage, its the invisible hitbox which pulls that geometry around as though its on a rubberband is what will need to be hit to for damage to register.

I play alot of Left-4-Dead and other Halflife-2 SOURCE engine based games, and its very similar in how physics collisions work, bump into another mech and you kinda jiggle around and teleport around eachother, or Melee such as in SOURCE games is a hitbox vs hitbox only issue.

With Melee weapons in SOURCE, you always have to lead a target at where its hitbox is for anything to register, you'll still get clientside soundseffects and even blood/sparks and all the telltale signs of a successful hit (false positives), but for your target to take damage, it must literally hit that invisible hitbox for anything to happen.

Difference is, SOURCE has a pretty decent hitscan detection for bullets (except there is no physics applied, its just lasertag tbh), better than most.

Now some people may not agree, but IMO Battlefield-4 is the only multiplay game I've ever played that has really good hitregistry all the time, not perfect, but better than the rest, and the majority of the weapons involved are all ballistic, with range drop physics and travel time as well.

Battlefield-4 uses Host-state-rewind afaik.

This is how things work in SOURCE engine games, and it seems to be so similar to whats happening here, I thought I'd post it.

Edited by Mister D, 25 September 2014 - 09:59 PM.

[zortesh]

Take a tag and attempt to tag a light, laser hit detection problems become very obvious then.

The narc has wonky hit detection too sadly.

[Mordin Ashe]

Laser hitreg kind of sucks, I found out when I switched to pure laser configuration on several Mechs. Sometimes it is ok but then I alpha a standing Mech that should take 20dmg to kill and even after 50 pinpoint alpha precisely where it should be he is still standing and laughing.
Hope this gets fixed.

[Kiyoshi Amaya]

All we can do is keep posting videos and try to keep this thread alive. Hopefully at some point, Russ will look into it.

[BARBAR0SSA]

He did ask for proof..

[Cyborne Elemental]

Proof will just require some test sessions, and should be fairly easy to obtain.

[Andi Nagasia]

back when i had some premium time i did some tests,
(sadly the video was corrupted, and im out of premium time to retest)

here's what i found out, test was KFX-S(-PL +ER) vs COM-B(-SRM +SL), one of each laser type
now for the test the moving COM-B was looking towards my stationary KFX-S, then repeat switched,


KFX-S vs moving COM-B,
moving towards me from 100m away,
test 97kpm, LL, 70% hit detection
test 97kpm, ML, 75% hit detection
test 97kpm, SL, 75% hit detection

side movement parallel 100m away,
test 97kpm, LL, 70% hit detection
test 97kpm, ML, 75% hit detection
test 97kpm, SL, 70% hit detection

moving away from me from 100m away,
test 97kpm, LL, 65 hit detection
test 97kpm, ML, 65% hit detection
test 97kpm, SL, 70% hit detection


COM-B vs moving KFX-S,
moving towards me from 100m away,
test 97kpm, LL, 75% hit detection
test 97kpm, ML, 75% hit detection
test 97kpm, SL, 80% hit detection

side movement parallel 100m away,
test 97kpm, LL, 70% hit detection
test 97kpm, ML, 70% hit detection
test 97kpm, SL, 75% hit detection

moving away from me from 100m away,
test 97kpm, LL, 65% hit detection
test 97kpm, ML, 70% hit detection
test 97kpm, SL, 65% hit detection


hopefully this helps, it seems the Laser is 25% off,

this was a simple raw damage test, not a hit test location test,
ran match fired 3 times with weapon then exited for damage total,
example IS ML should have done 15 damage, 11 was dealt,

Edit- spelling

Edited by Andi Nagasia, 26 September 2014 - 06:45 PM.

[Zoid]

zortesh, on 25 September 2014 - 09:45 PM, said:

Take a tag and attempt to tag a light, laser hit detection problems become very obvious then.

The narc has wonky hit detection too sadly.

Yep. I've started running TAG on my double ERLL KFX partially for this reason. I don't want to waste the heat if I know the hitreg won't let me connect anyway, so if I TAG that ECM spider and still can't lock on, I lead it more before I fire the actual guns.

Edited by Zoid, 26 September 2014 - 07:58 PM.

[Ghogiel]

Aww man I had the best example of laser hit reg fail yesterday. I hadn't seen it that clear for a while. I imediately started up banditcam after, so hopefully I'll grab a few clips that show the lead I need. It's pretty clearly not regging on lights circling at 25ms around my mech, the pixel differentation is quite large then. ie shoot the trailing side or arm and nothing will reg.

Edited by Ghogiel, 27 September 2014 - 02:05 AM.

[Kiyoshi Amaya]

A few days ago I had the perfect example. I'd carefully walked up behind a camping stalker that was take pot shots at my team mates, lined up the shot carefully and alpha'd him in the rear CT at about 30m with 2 x srm4, 2 x srm6, 1 x med laser and 1 x MG... not once, but twice before he turned to see what was shooting his rear. And guess what damage I did... That's right, I just scratched the paint. He still had most of his rear armour. I managed to get a third alpha in but that didn't register either. So, after the match, I went to watch the video and, that's right, murphy's bleedin' law, I'd forgotten to press record. I mentally took myself outside and gave myself a good kicking for that one.

[LiGhtningFF13]

What?! Beam weapons are more that fine, just had a round against a Stalker that always hit critical torso parts no matter what situation, really strange.

[Kiyoshi Amaya]

LiGhtningFF13, on 27 September 2014 - 02:46 AM, said:

What?! Beam weapons are more that fine, just had a round against a Stalker that always hit critical torso parts no matter what situation, really strange.

For you maybe, but obviously not for everyone. Just out of interest, where are you? It might be a location thing.

[Jonny Taco]

LiGhtningFF13, on 27 September 2014 - 02:46 AM, said:

What?! Beam weapons are more that fine, just had a round against a Stalker that always hit critical torso parts no matter what situation, really strange.

So hit detection is fine because some dude shot you is your argument?

[Marauder IIC]

Yeah its feels like closed beta without hsr. I have to lead the shot to hit. (Lasers) Balistics seems ok. Srms only have about a 65% hit reg.

[Jonny Taco]

Marauder IIC, on 27 September 2014 - 05:06 AM, said:

Yeah its feels like closed beta without hsr. I have to lead the shot to hit. (Lasers) Balistics seems ok. Srms only have about a 65% hit reg.

I've noticed that If I fire srms at a target while laser vomit is burning that the srms sometimes do 0 damage.

I'm wondering if too much is going on and the server just drops some of the damage in the transition. Lasers being a very common component within many of these complaints leads me to believe this even more. The reason of course being the number of "hits" calculated for each trigger pull. In the case of large numbers of lasers, there could be maybe 500+ instances of damage being applied per alpha to a hostile mech. Combine this with SRMS which also require large numbers of individual calculations. Reducing the number the total "hits" within the beam duration of each laser may be a very good experiment and a reasonable first step.

Edited by lartfor, 27 September 2014 - 05:56 AM.

[LiGhtningFF13]

lartfor, on 27 September 2014 - 04:27 AM, said:

So hit detection is fine because some dude shot you is your argument?

No, that is just one of a view! Trust me I regualry use beam weapons. Work fine for me and my location is central Europe/Ger.

Edited by LiGhtningFF13, 27 September 2014 - 07:16 AM.

[Cyborne Elemental]

How good is it for you against mediums and lights?

[Ghogiel]

LiGhtningFF13, on 27 September 2014 - 07:15 AM, said:

No, that is just one of a view! Trust me I regualry use beam weapons. Work fine for me and my location is central Europe/Ger.

You're the guy that said that an odd STK was lasering the squishy parts... well my STK stats are probably in the top 1% of players and I'm just not going to trust your laser ability over mine.

[Thorn Hallis]

It's not just lasers, PPCs seem also be affected.

[MechWarrior5152251]

Most pilots touch someone with a laser beam and expect a bunch of damage. Lasers don't work that way. You have to hold them on target for their full duration to get full damage. Its pretty clear how little of Clan UAC5 ammo actually lands on most of my targets.

Has someone had hit reg issues in the training ground while stationary? That I would be interested in hearing.

Why can't they move hit reg to client side then have the damage sent to server? I know this would open things up for hacks, and low ping clients could cause havoc. Damage would always land, but it may take a second for the mech to die.

Edited by 911 Inside Job, 28 September 2014 - 04:33 AM.