31 replies to this topic

[MightyBolamite]

Cest7, on 26 February 2013 - 05:10 PM, said:

I've noticed front to back but not so much back to front.

I've seen this too.
Hit a K2 with an SRM6 dead on and his back armor blinked. Annoyed me so much.

[Venraa]

I think I have had this too. In a HBK-4G and lost my AC/20 in the first shot I took of the match. Wasn't sure exactly what weapon hit me but it was a ballistic of some kind from a phract. Given that a phract cannot stack enough ballistic damage to get through my front torso armor in a single shot it must have got transfered to the back, then critted and destroyed the AC/20.

My friend in his raven had the same thing happen earlier today. Took a single PPC hit which transfered to the back, critted, ammo explosion, dead raven.

[Kaeb Odellas]

At least twice I've been killed by a light mech will running from it in my Trebuchet. Still had back armor, and I'm fairly certain he was directly behind me the entire time.

I've only ever noticed it in Trebuchets though.

[Ari Dian]

Just got killed by a Jenner with this bug.
My back armor was open, so i turned around. I got killed with the next salvo. My front armor was still fine, but side torso destroyed by damage in the back. He was the last remaining mech. So there was no other who could have shot me in the back.

[Corvus Antaka]

I was fighting an Atlas today whos side torsos had both been ripped off...on side from behind, and he still showed armour on his front side torso for that side, because I hit his completely red CT (or so i thought) only to see the missing side torso light up and take damage. very wierd.

[Melcyna]

sycocys, on 28 February 2013 - 06:26 AM, said:

From their explanation the net-rewind code will do something along the lines of keeping a server side log per match of mech locations that the weapons code can poll to determine exactly where a mech was at the time it received a fire command.

My guess is that this will hurt high ping players more than help them. And it will incredibly help low ping players because they will better be able to predict shot locations on faster/high ping targets. Will as I said probably lead to mystery head shots from the perspective of higher ping players.

If the rewind is well coded and works properly that's not supposed to happen

Spoiler

say someone is playing with 500ms latency, if we assume that to be round trip latency then everything he sees is therefore 250ms late and everything he executes reaches the server 250ms late as well

when he sees a moving target and aim at it, he sees what's essentially a ghost 250ms behind the real one, now let's say he fires at it... this shot takes 250ms to reach the server

Depending on how it's coded several things can happen: (we assume the system including the client time stamp all the message or packets sent to the server including firing commands)

A. the server can just look at the log and determine where things were when he actually fired without accounting for the time it takes for the latency and thus what he actually sees
[in this case, it cuts the effect of latency for his command, but not what he sees]

B. the server can factor in the latency delay to the shooter into where the players in the game ACTUALLY were as according to his screen when he issued the command
[in which case what he sees in his screen when he fired is used as what the server calculates to determine his shot miss or hit, note that it wouldn't be exact but a close approximation since latency is NOT completely fixed]

it's not exactly rocket science, the server knows the latency to the player, all it has to do is add or rather subtract this latency delay time into the timeline to see what it looks like from the player's point of view since most players latency are not exactly fluctuating by large margin (if the player did something to cause large latency in short spike then that's his own fault) and therefore this should be a mostly accurate rendition of what he saw back then

Spoiler

Someone also previously mistaken rubber banding with this state rewind... which is not accurate,

What happens with rubber banding is quite simple, the player's movement and firing effect (visual and audio) are executed client side... that is to say you don't seems to be waiting for the server confirmation before you actually move or fire and this is done for obvious reason since it's extremely UNPLEASANT to play a game where even your movement and action is delayed by latency especially in real time action game

The action however is confirmed with the server, and when it comes to movement this means that thanks to the current flawed system the client may well render you move through certain things or path which is not supposed to be passable (the client doesn't seems to check these properly in some cases) resulting in the player seemingly rewinded to his previous position or more correctly to use the term, he rubber banded back to where he was before because the confirmation with the server would then indicate that the player's new position is invalid based on what he was doing and not possible and thus not accepted while his old position is used instead (except he keeps trying to move there)

Normally the client should be indicating the proper response to this that is to say it should recognize that the action was invalid (obstacle in the way or what not) rather than needing to wait for the server confirmation (and thus displaying wrong visual effect to the player) but for whatever reason this is not always the case right now with MWO

Edited by Melcyna, 03 March 2013 - 01:10 AM.

[Ari Dian]

And again. was running away from a raven because my front armor was open. Got hit by SSRM2, dead. Screen say back armor is still there, but CT internal is gone.

This bug is really getting annoying.

[Metalice99]

This is obviously still a problem. Last night I shot a centurion in the front CT from close range with an SRM cat. Although his frontal armour exceeded the alpha strike of the missiles, and he had received no previous damage, he popped like a very expensive, nuclear reactor powered balloon. Poor guy.
Hopefully if there's a patch on Tuesday, it will go some way to solving the issue. Fingers crossed!

[Aratan Aenor]

Happened to me today, my K2 took 6 SRM4 from an A1 about 100 meters in front of me. All of the damage was applied to my rear torso sections. I was left with no armor on my right rear torso section and my rear CT armor was in the red. My front armor was completely untouched.

[Vedic]

I opened a ticket last week and gave the explanation I posted here - along with a link to the thread. I got this response:

"The problem you have encountered is a known issue and we hope to have it fixed soon."

So we're not crazy. It would be nice if they'd list it on the official "Known Issues" thread so we know it's actually being addressed...

[Mawai]

ohtochooseaname, on 25 February 2013 - 02:27 PM, said:

I've been in a Raven, and taken a double gauss to the face with full armor. It went straight to the back and turned my ct orange without damaging my front armor.

If I had to guess ... I would say that the shots ducked past the front armor hitbox ... went through the model ... and hit your rear armor hit box from the inside. Which eliminated your rear armor and transferred damage to your internals ... though that would depend on the implementation of hit boxes and weapon tracking and assumes that internal structure doesn't have its own hitbox but relies on damage transfer from the armor hitboxes for damage propagation.

I could see this sort of thing happening with moving mechs ... depending on the spatial resolution applied in weapon tracking code ... it might be possible for a mech to step into a ballistic trajectory in such a way that the calculated impact point ends up on the rear armor.

Edited by Mawai, 05 March 2013 - 08:03 AM.

[ohtochooseaname]

Mawai, on 05 March 2013 - 07:59 AM, said:

If I had to guess ... I would say that the shots ducked past the front armor hitbox ... went through the model ... and hit your rear armor hit box from the inside. Which eliminated your rear armor and transferred damage to your internals ... though that would depend on the implementation of hit boxes and weapon tracking and assumes that internal structure doesn't have its own hitbox but relies on damage transfer from the armor hitboxes for damage propagation.

I could see this sort of thing happening with moving mechs ... depending on the spatial resolution applied in weapon tracking code ... it might be possible for a mech to step into a ballistic trajectory in such a way that the calculated impact point ends up on the rear armor.

I was stationary...but yeah, hit detection appears to occur at the skins (after all no reason ever to have anything but hollow shells, which is why, if you get your cockpit into the drop ship on river city, you can see through it. In a rendering sense, it's usually not just shells, but also the objects only appear on one side.

I killed an atlas last night. Turned his front ct red via an AC2. He turned his back to me and was solidly facing away with full armor. My ac2 went straight through and killed him.