Jump to content

Projectile Count Vs Hit Detection


6 replies to this topic

#1 TheFlyingScotsman

    Member

  • PipPipPipPipPipPipPip
  • 639 posts
  • LocationCA

Posted 30 November 2013 - 11:46 AM

TLDR: There are too many individual projectiles in play thanks to MGs, LRMs, and most impportantly, Laser weapons. These cause a large amount of hits to deal no damage because of the processing needs. This can be fixed by reducing projectile count and increasing individual projectile damage (and for LRMs, health also, obviously)

So, as does anyone who bothers to tally their own hits and whether they deal damage, I have for a long (very) time noticed that many hits do not land. While this is most obvious when using, say, an AC20, PPC or other high-damage singe hit weapon, the fact that this happens to all weapons is apparent.

But, there are also "good games" and "bad games". Some games have a higher rate of hit detection, as do some periods of gameplay. This post is to describe the situations in which hit-non-registration happen most often. The biggest culprit of hit-non-registration seems to be the number of projectiles in the game at any given time.

For example, a shot from, say, an AC20, is not as likely to be properly recorded when the server is also parsing shots from a 4xMG spider or 6xMG jager. Likewise, it is also less likely to detect shots when the air is full of LRM volleys.

Likewise, weapons with higher projectile counts (MGs, LRMs, all laser weapons) have a bad habit of producing sub-par damage when actually used in a live match. I believe this is due to the high number of projectiles they themselves fire. This fact is accentuated by perceptible lag spikes when/while these weapons are firing from anywhere on the map.

This server overload is possibly among the largest factors that reduce the number of solid hits being actually detected and applied, and therefore the best way to reduce these unfortunate numbers of "hits that do nothing" is to seriously reconsider the mechanics of them.

For example, MGs could be made to fire at 1/4th the current rate and deal 4x damage. Lasers could have the number of individual "pings" of damage within the beam also reduced and increased in individual strength. I don't readily see a solution for LRMs, though reducing the processing needs for tracking other weapons would probably be enough to bring everything >closer< to hand.

Just a thought, PGI. Simpler programs are easier to maintain and ensure proper functions of, and most of the mechanics behind these weapons are arbitrary flash and wasted processing power, with unfortunate results.

Edited by TheFlyingScotsman, 30 November 2013 - 11:47 AM.


#2 Troutmonkey

    Member

  • PipPipPipPipPipPipPipPipPip
  • Moderate Giver
  • Moderate Giver
  • 3,776 posts
  • LocationAdelaide, Australia

Posted 30 November 2013 - 05:46 PM

Any decent game engine should be able to maintain, update, draw and calculate thousands of collisions at a time.

The main cause of hits failing to register is HSR not working as intended.

Reducing the amount of projectiles is only a bandaid fix. A much better fix would be to create more efficient code that's much better at handling larger numbers of calculations. As a programmer I know there are many ways to do this, but you have to be clever about it and it's not always easy. Brute force code is always the easiest to write, making it effecient is a bit harder.

#3 stjobe

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 9,498 posts
  • LocationOn your six, chipping away at your rear armour.

Posted 01 December 2013 - 03:07 AM

View PostTheFlyingScotsman, on 30 November 2013 - 11:46 AM, said:

I have for a long (very) time noticed that many hits do not land.

L2Aim :)

No, but seriously, how do you know if a "hit" was a hit and not a miss? Did that MG burst really hit with all projectiles? Were you really on-target with the whole LL beam? Are you 100% certain that your AC/20 shot in fact connected? I don't think very many people are able to tell the difference in most cases; there's a lot of graphical effects in the game and it's also a rather fast-paced game with lots of things happening at once.

There's issues with hit detection, there's no denying that. There's issues with HSR too. And there's a lot of misses that people think were hits.

Until we can separate the first two from the last it's a bit premature to lambaste the devs for not being able to write code to track all the projectiles in the game.

Unless you have some actual evidence, of course. But I didn't see any in your OP.

#4 Warrax the Chaos Warrior

    Member

  • PipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 925 posts
  • LocationMyrror

Posted 01 December 2013 - 09:51 AM

View Poststjobe, on 01 December 2013 - 03:07 AM, said:


There's issues with hit detection, there's no denying that. There's issues with HSR too. And there's a lot of misses that people think were hits.

Until we can separate the first two from the last it's a bit premature to lambaste the devs for not being able to write code to track all the projectiles in the game.


You might be correct, but there may also be some validity to the OP's idea. We have absolutely no way of knowing for sure.

Hit detection gets slightly better after a patch, and then degrades again quickly. The difference was more pronounced before HSR, but it's still there. There has to be a reason for that. This could be worth looking in to.

#5 YueFei

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,184 posts

Posted 01 December 2013 - 09:56 AM

View Poststjobe, on 01 December 2013 - 03:07 AM, said:

L2Aim :)

No, but seriously, how do you know if a "hit" was a hit and not a miss? Did that MG burst really hit with all projectiles? Were you really on-target with the whole LL beam? Are you 100% certain that your AC/20 shot in fact connected? I don't think very many people are able to tell the difference in most cases; there's a lot of graphical effects in the game and it's also a rather fast-paced game with lots of things happening at once.

There's issues with hit detection, there's no denying that. There's issues with HSR too. And there's a lot of misses that people think were hits.

Until we can separate the first two from the last it's a bit premature to lambaste the devs for not being able to write code to track all the projectiles in the game.

Unless you have some actual evidence, of course. But I didn't see any in your OP.



There are definitely occasional hitreg issues. I've been on the giving and receiving end. There are times when I've been SLAMMED by weapons in my Jenner, and absolutely nothing happened. And other times when I don't feel anything, but mysteriously I will be missing all the armor from a leg or CT.

There have also been times when I plonk myself down, stationary, firing at the back of a stationary target, with a couple shots, and nothing happens.

With that being said, hitreg works most of the time, and works well enough that gameplay isn't crippled by the non-regs. They still ought to look into and fix the non-regs.

#6 Murphy7

    Member

  • PipPipPipPipPipPipPipPip
  • Elite Founder
  • Elite Founder
  • 1,553 posts
  • LocationAttleboro, MA

Posted 01 December 2013 - 10:04 AM

Interesting conjecture, OP.

This would suggest an interesting and potentially awful test.... carry MG's and fire them continuously when being fired at by a large projectile, such as AC/10, AC/20, or Gauss..

Can you flood the system such that it loses track of the big round? Defacto anti-projectile defense by confusing the system into losing the incoming enemy projectile?

#7 thesleepyslam

    Member

  • PipPipPipPipPip
  • Big Brother
  • 112 posts

Posted 01 December 2013 - 07:50 PM

i run dual gauss. i have more hit registration issues than any other person. no round is easier to spot a hit, this game still has serious registration issues, not hit detection, i can see the rounds hit in my screen just fine. i usually run around 33 ping.

Edited by thesleepyslam, 01 December 2013 - 07:52 PM.






1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users