Jump to content

Increasing Problem With Immune Mechs


11 replies to this topic

#1 Recon Don

    Rookie

  • Philanthropist
  • 3 posts

Posted 17 December 2024 - 11:07 AM

Myself and many other players (guild mates also) are noticing an increasing frequency of enemy mechs becoming seemingly immune (temporarily) during drops. We have seen this for years on light mechs and chalked it up to latency (and desycning) but over the past two-three months have noticed it on larger - slower moving mechs also.

Few nights ago it reached a very noticeable point with a Fafnir - both myself and several lancemates noticed that direct hits (red reticule) on the barn door of a mech were not being reflected at all on the paperdoll - this was noted by spectators who saw the same thing through our cockpits. The Fafnir went on to tank 4 mechs and kill each, not taking a single point of damage.

Getting this every other match almost now, mech size not mattering at all anymore - is anyone else having this problem? Is there a fix in the works?

#2 BlueDevilspawn

    Member

  • PipPipPipPipPip
  • CS 2023 Top 12 Qualifier
  • CS 2023 Top 12 Qualifier
  • 164 posts

Posted 17 December 2024 - 11:10 AM

View PostRecon Don, on 17 December 2024 - 11:07 AM, said:

Myself and many other players (guild mates also) are noticing an increasing frequency of enemy mechs becoming seemingly immune (temporarily) during drops. We have seen this for years on light mechs and chalked it up to latency (and desycning) but over the past two-three months have noticed it on larger - slower moving mechs also.

Few nights ago it reached a very noticeable point with a Fafnir - both myself and several lancemates noticed that direct hits (red reticule) on the barn door of a mech were not being reflected at all on the paperdoll - this was noted by spectators who saw the same thing through our cockpits. The Fafnir went on to tank 4 mechs and kill each, not taking a single point of damage.

Getting this every other match almost now, mech size not mattering at all anymore - is anyone else having this problem? Is there a fix in the works?


Were you playing event queue? During bonus armor events, the damage is NOT shown until the bonus armor is shot away. Only then does the paperdoll show damage.

#3 Recon Don

    Rookie

  • Philanthropist
  • 3 posts

Posted 17 December 2024 - 11:12 AM

View PostBlueDevilspawn, on 17 December 2024 - 11:10 AM, said:


Were you playing event queue? During bonus armor events, the damage is NOT shown until the bonus armor is shot away. Only then does the paperdoll show damage.



These were all quickplay drops.

#4 LordNothing

    Member

  • PipPipPipPipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 17,224 posts

Posted 17 December 2024 - 11:33 AM

i was doing some drops in my dual ac20, dual snub bullshark, and it seems like sometimes damage would register, but not quite a 60-point ppfld which is what i was packing. when you hit a mech square in the ct, its a medium, and it turns slightly yellow, you know you didnt make full contact. the splash from the snubs seems to register. and i can explain the ac20s for being widely spread and likely to throw off the convergence, eg converge around the target rather than on it. but this particular shot looked like it landed square in the ct.

maybe i walked in front of a twig and it split my shot. ac20s went around and snubs connected. if a snub hit each side torso, splash may have converged on the ct and appeared as if it hit there. then again the very same build got two headshots and legged a few squirrels way closer than the medium in question (i beleive it was a griffon). i had the aim gods' favor yesterday. so this just seems a bit off. i usually just write it off as an fpu glitch and fire again. these kind of occurrences are pretty rare.

i think it could be a bug in collision response. a hit is registered but the game gets confused about what to do about it and ignores it.

most of the time there is video, people are looking at the tracers and not the projectile itself. those are just purely decorative, the projectiles matter. if the server disagrees with what you see, the server wins.

Edited by LordNothing, 17 December 2024 - 11:36 AM.


#5 BlueDevilspawn

    Member

  • PipPipPipPipPip
  • CS 2023 Top 12 Qualifier
  • CS 2023 Top 12 Qualifier
  • 164 posts

Posted 17 December 2024 - 11:46 AM

There may be a couple of things. First of all, if you're shooting at a light mech, it's likely convergence and not hitreg. Given the way convergence is coded only part of your alpha may be landing on the mech. Against a large slow assault like a Fafnir, it could be a server connection/desync issue which rarely happens at least in my experience. I have unloaded against an assault with lasvom before, and didn't get a hitreg. Again, that's quite rare.

Best thing you can do is share a recording of the incident. Got any vods or clips?

#6 Meep Meep

    Member

  • PipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 2,982 posts
  • LocationBehind You

Posted 17 December 2024 - 12:26 PM

The only thing I've noticed that's a bit more pronounced are IS er ppc seem to pass through mechs more often now than they used to do and I've hit frontal armor only to have it register as an internal or rear armor hit with them. Slower ppc seem fine.

#7 l33tworks

    Member

  • PipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 1,293 posts
  • LocationSydney, Australia

Posted 17 December 2024 - 04:29 PM

View PostMeep Meep, on 17 December 2024 - 12:26 PM, said:

The only thing I've noticed that's a bit more pronounced are IS er ppc seem to pass through mechs more often now than they used to do and I've hit frontal armor only to have it register as an internal or rear armor hit with them. Slower ppc seem fine.


This happens mostly because ppcs have a trail and only the leading front of the projectile clips and registers damage. So when you barely miss a PPC shot, and then the target walks in the direction of the PPC shot right after, you see the ppc trail clipping through their mech and you think your shot went through the target, when in fact the damage causing part of it (the front) it did not. With slower ppcs the enemy has more time to walk into the shot and actually get hit. PPC velocity has increased so you have to lead less these days than you had to.

I actually think the rollback code and moving mech detection is pretty good, but not as good as it used to be when they first patched in rollback netcode, however there is a problem with total damage calculation even on mechs standing still. I have unloaded 2x lb20x 5smr4s and 2 plasmas into all sorts of mechs over and over even ones standing still at close with no convergence issues range and they do not take the damage they are meant to take. This is the weird one to me.

A lot of the damage simply turns into vapourware and not because of a miss. Maybe I am not taking skill trees between players into account correctly but some games the damage registers fine and others its just not adding up as much.

Edited by l33tworks, 17 December 2024 - 04:31 PM.


#8 D V Devnull

    Member

  • PipPipPipPipPipPipPipPipPip
  • 4,329 posts

Posted 17 December 2024 - 08:13 PM

View PostRecon Don, on 17 December 2024 - 11:07 AM, said:

Myself and many other players (guild mates also) are noticing an increasing frequency of enemy mechs becoming seemingly immune (temporarily) during drops.

<<< snip >>>

Getting this every other match almost now, mech size not mattering at all anymore - is anyone else having this problem? Is there a fix in the works?

Hello there, 'Recon Don'... You're sure having one heck of a glitch there. I suggest you run the MWO Repair Tool (or Steam's File Verify if that is how you're calling MWO up to play) and make sure that MWO's files are not messed up. If you are indeed using the MWO Portal access method (and not Steam for such) to launch, then ensure that the MWO Repair Tool only has those last two checkboxes ("fetch files from other MWO installs on this computer", and "double-check hash mismatch") enabled in its' internal options so other things are not being messed up. :o

After doing so, or maybe even before, go get whatever Time Synchronization Program makes you comfortable. This will help ensure that your computer's internal clock is close as it can get to real time. Someone else introduced me to this option, and it has saved me some grief with MWO gameplay in general. Yes, really, it happens that MWO is very sensitive to an improperly tuned system clock. Do not trust the one in Windows itself for this part, because it may not be properly doing the job right. I learned this one the hard way due to issues loading in and random lag troubles during gameplay. :huh:

One program which I can seriously recommend is "NetTime 3.14" @ https://timesynctool.com/ and happens to be what I actively use daily now. The webpage is very basic, likely because of that author's life, at least given what I've been able to determine, but they still make a good program to use. It has kept me from giving up on my MWO sessions more times than I can count now, and has many configurable options to utilize should the need arise. Make sure to always resync however many times it takes after system startup, resume from sleep, and whenever else needed in order that your computer's clock will never be more than mere milliseconds (less than 100) off of real time. Share it with your guildmates if this solution works for the computer which you use! ^_^

EDIT (About 13 Minutes Later) :: I almost forgot... When you go to resync time with any tool, actively avoid MWO's Game Client being in RAM on your computer at that time. It would mess with the connection in an unwanted manner! :blush:

Other than that, I wish you well on the battlefields. Perhaps we might end up on the same Team at some future point by random chance? B)

~D. V. "trying to help 'Recon Don' put an end on Hit Registry Issues they're having" Devnull

Edited by D V Devnull, 17 December 2024 - 08:26 PM.


#9 Saved By The Bell

    Member

  • PipPipPipPipPipPipPip
  • 738 posts
  • LocationJapan

Posted 17 December 2024 - 09:10 PM

I can confirm, shooting standing light-medium with ac20 + 2 snubs doesnt kill him, in some times barely any damage.

Like I shoot 3 times with it in cicada, which was standing and looking at me. Still standing and shooting. It must be more 120 damage, no?

I understand, it was moving, but on my screen it was just standing and shooting, not moving torso.

It must be some exploit.

#10 LordNothing

    Member

  • PipPipPipPipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 17,224 posts

Posted 17 December 2024 - 09:20 PM

View PostMeep Meep, on 17 December 2024 - 12:26 PM, said:

The only thing I've noticed that's a bit more pronounced are IS er ppc seem to pass through mechs more often now than they used to do and I've hit frontal armor only to have it register as an internal or rear armor hit with them. Slower ppc seem fine.


when i was doing quake modding i never made anything's velocity more than 2000. anything higher than that broke things. that game ran entirely on fp32. fp64 is pretty fast now, but i bet the engine still uses fp32. this engine is old enough where that would come with some performance gains. on most modern compilers i use (gcc, visual studio) sizeof(float) returns 8 bytes in its default configuration. however whatever they used to compile the blobs might have been something different. keep in mind we are talking the size of the data type, not the instruction size (you can still use a 32 bit type on a 64 bit build).

multiplayer games often prefer smaller types for transmission that dont bog down the internet connection as much, trading precision for performance. also the server tick is 30/s and most of the frames you see are actually based on interpolated position data.

Edited by LordNothing, 17 December 2024 - 09:29 PM.


#11 Novakaine

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Veteran Founder
  • Veteran Founder
  • 5,731 posts
  • LocationThe Republic of Texas

Posted 18 December 2024 - 08:26 AM

View PostRecon Don, on 17 December 2024 - 11:07 AM, said:

Myself and many other players (guild mates also) are noticing an increasing frequency of enemy mechs becoming seemingly immune (temporarily) during drops. We have seen this for years on light mechs and chalked it up to latency (and desycning) but over the past two-three months have noticed it on larger - slower moving mechs also.

Few nights ago it reached a very noticeable point with a Fafnir - both myself and several lancemates noticed that direct hits (red reticule) on the barn door of a mech were not being reflected at all on the paperdoll - this was noted by spectators who saw the same thing through our cockpits. The Fafnir went on to tank 4 mechs and kill each, not taking a single point of damage.

Getting this every other match almost now, mech size not mattering at all anymore - is anyone else having this problem? Is there a fix in the works?


See I told you plot armor.

#12 RockmachinE

    Member

  • PipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 1,186 posts

Posted 18 December 2024 - 11:52 AM

Haven't noticed anything. Seems normal on my end.





2 user(s) are reading this topic

0 members, 2 guests, 0 anonymous users