Jump to content

I Still Prefer The Convergence Idea More Than Laser Damage With A Lock


  • You cannot reply to this topic
26 replies to this topic

#21 Dracol

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Steadfast
  • The Steadfast
  • 2,539 posts
  • LocationSW Florida

Posted 21 October 2015 - 04:01 PM

View Postoldradagast, on 21 October 2015 - 03:17 AM, said:


How would it look? The weapons would not all hit at the same spot on a mech at long range, basically. A small, random element would affect their exact aim, so they'd be off by a bit, though sometimes you wouldn't even notice. Again, the goal is NOT to turn all weapons into shotguns that spray damage all over, but to end the ability to drop two Gauss rounds and a pile of lasers and PPC's (or a pile of AC rounds, etc.) perfectly on one hit location every time at along ranges. At mid to close range, yeah, you probably won't even notice the cone of fire.

I understand that is what happens when the trigger is pulled. However, large lasers have a significant burn time. Add zoom/adv. zoom, and a pilot will have the opportunity to preposition the laser.

So, after that initial trigger pull and the point of light is a little bit off center, does:
A. the pilot continues to control the laser like they do now when firing with lasers while ascending on jump jets, thereby giving the pilot a chance to re-position the laser onto the component they were initially aiming?

B. the laser continually moves on its own in the COF as the pilot re-positions the laser during the duration of the laser burn?

Which way do you see the laser working and why do you prefer that effect?

Edited by Dracol, 21 October 2015 - 04:04 PM.


#22 Hotthedd

    Member

  • PipPipPipPipPipPipPipPipPip
  • The People's Hero
  • 3,213 posts
  • LocationDixie

Posted 27 October 2015 - 09:10 AM

Honestly, if it's just ONE laser, there shouldn't be a need for a CoF. It should be precise. I think a CoF should come into play when:

Moving fast (Run speed or jumping)
Running too hot
Firing multiple weapons

That way the PILOT decides the possible spread.

#23 Hans Von Lohman

    Member

  • PipPipPipPipPipPipPipPip
  • Knight Errant
  • 1,466 posts

Posted 28 October 2015 - 10:47 AM

View PostTeam Chevy86, on 19 October 2015 - 08:04 AM, said:

I heard recently that it ate a ton of memory in game with 16 players all sending and receiving different convergence scenerios... Mechs were literally rubber banding like on an old dial-up internet connection. Maybe that's why they seem to be pussyfooting around the idea...


We're not talking about the convergence that was in the beta test, where your arms took time to re-focus on a new range. I get why they did that. They were trying to make light mechs harder to hit when running directly at and away from you by going faster than the convergence could keep up with them.

This is NOT that convergence. This is still instant focus on the range. In fact, there are fewer variables in that your range focus is not constantly being updated every few milliseconds as you move around. You just have two ranges. infinity, or your lock on range.

However, the thing is I would like to point out again that when you get a lock on your target, everything works as it already does today. You only lose out when you don't have a lock, and having the game play janky without a lock is sort of a good thing in a weird way. The worse it is without a lock, the better motivated people will be to want a lock.

#24 Piousflea

    Rookie

  • Bridesmaid
  • 8 posts

Posted 30 October 2015 - 01:58 PM

Wanderer's convergence idea makes target lock valuable; it makes targeting computers awesome and very easy for newbies to understand; faster targeting = your shots converge on the enemy. Targeting-dependent convergence would put a hard cap on the effectiveness of every pinpoint build regardless of whether it is lasers, ppc, Gauss, or AC.

You could still snipe without target lock by using only the weapons mounted on one arm - there would be no "cone of fire", they would all hit one spot. However, you'd have to compensate your aim based on the fact that the arm is off-center.

The current PTR system only nerfs lasers, it's extremely complicated and gives very poor feedback to the players... you can't really tell whether your lasers are doing full damage or not.

An "all or nothing" Convergence system is easy for players to understand. If your shots are failing to converge on your enemy then you need to target him. As long as you have your foe targeted, the aiming system works exactly the same as it does today.

Wanderer's Convergence idea also has the advantage of not having a bunch of variables to tweak like a "cone of fire" system. No one needs to worry about how big the cone of fire is, what changes the size of that cone, etc... either your weapons have 0% convergence or they have 100% convergence.

Edited by Piousflea, 30 October 2015 - 02:00 PM.


#25 ChapeL

    Member

  • PipPipPipPipPipPipPipPip
  • Elite Founder
  • Elite Founder
  • 1,363 posts

Posted 30 October 2015 - 02:55 PM

View PostHans Von Lohman, on 20 October 2015 - 05:59 PM, said:

Here is a simple picture of what I am talking about. You want people to care about information, and getting a lock on, then you have to make it affect combat directly.

I just don't think making lasers act weird is the way to go.

Posted Image


I'd go a step further and have everything mounted in the arms can converge and everything in the torso cannot.

Edited by ChapeL, 30 October 2015 - 02:57 PM.


#26 Fubbit

    Member

  • PipPipPip
  • Mercenary Rank 4
  • Mercenary Rank 4
  • 84 posts

Posted 30 October 2015 - 03:44 PM

I like the idea. (Seems) simple and fixes the pinpoint issue.
Treating arm and torso mounted weapons differently is also an interesting extension of the idea.

#27 Garheardt The Black

    Member

  • PipPip
  • Bad Company
  • 26 posts

Posted 01 November 2015 - 05:50 PM

Have any developers addressed this idea at all? Was it shot down or has it not been acknowledged?





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users