Alistair Winter, on 24 July 2015 - 08:59 AM, said:
Well, you know how I feel about this. I'd definitely prefer if it was harder to get target locks at long distances. Especially against fast moving lights and mediums, it should be a bit tricky to get a target lock, just like it's hard to get target locks on fast and agile fighters in combat sims / space sims.
One of the reasons LRMs aren't used very much at higher levels of play, is because PGI has balanced them to be good weapons for people who don't have a good aim. They are
low skill, low reward weapons. Gauss weapons are high skill, high reward. I'm not saying you have to be Jesus to do ok with a Gauss Jager, but in terms of pure hand-eye coordination, gauss rifles are harder to hit with, LRMs are easy to hit with.
If PGI made it harder to get target locks, they could improve the stats of LRMs without making them overpowered. Specifically, they could do things like increasing missile speed. Low missile speed and long travel times is why LRMs are often completely ineffective against good players who use cover. Low missile speed is also why dumb firing LRMs against ECM-protected target is almost always doomed to fail.
If PGI increased projectile speed to the point where LRMs were a nice damage boost to lasers at long range, the same way that SRMs are a nice damage boost to lasers at short range, I would be a happy man. Imagine seeing lots of Timber Wolves and Summoners that actually kept their LRM launchers equipped to complement their lasers at long range. That would be pretty cool.
And as a bonus, making it more difficult to get target locks would probably dampen some of the QQ when people get killed by LRMs.

I'd prefer that missiles travel about twice as fast as they do now and try to lead the target like modern missile systems, but that they have a limited turn radius, so that a full deflection shot against a 150 kph light is likely to miss entirely. Much like in flight simulators, when you fire a missile at an enemy aircraft heading perpendicular to you at relatively close range, the missile can't turn hard enough to get the appropriate lead and will miss the target. From farther away, the missile has more room to turn and gain an appropriate lead on the target, so it has a much better chance to hit from a simple fire control geometry point of view.
This would allow us to remove the min range damage falloff. Missiles deal full damage at any range more than about 20 meters (this is only to prevent face hugging LRM 60 splat attacks). For IS, to incorporate the min range penalties, we could reduce or eliminate their turning capability within the first 50, or 100, or 200m.
As for locking onto the target, the mech should have to paint the target mech (reticle on the actual hitboxes of the other mech) for a duration of time to achieve lock, and then the lock remains similar to how it is now, where you only have to be close. Think of it almost like an invisible TAG laser you need to use to designate the target for your on-board fire control systems. Your designation of the target, once lock is achieved, can become a bit more loose, because your targetting computer has coordinated with your other sensors to focus on that particular mech and read it's position based on updates from all of your sensors. To do indirect fire, you need someone to use TAG or NARC to get that lock, which would require only being close (similar to when you already have a lock).
Consider that a laser designator should be almost weightless and tiny, but TAG requires 1 slot and 1 ton because it's the communications uplink to the friendly fire control systems that requires sufficient equipment - this system has to have anti-jamming, encryption, multiple redundancies, shockproof, etc. so that's where it gets more expensive in terms of size and weight. So the mech firing its own missiles has it's own invisible laser designator used to acquire LOS targets, and it costs nothing and weighs nothing. But to do this for another mech, you need TAG or NARC, each of which have the communications capabilities to do so at the appropriate cost in space and weight.
Edited by Dino Might, 24 July 2015 - 09:14 AM.