|
Post by nerd1000 on Sept 28, 2016 0:52:02 GMT
What are you using for point defense? I have found lasers near-useless (multiple optimized 13MW lasers usually getting 1-2 kills before flak missiles enter their danger space), and the only kinetic weapons I have made with an effective range usefully greater than the danger space abused the integrator bug. I agree that the guidance needs help; I was referring to the missile/decoy balance. (I have had decent results from lining up precise interceptions and reserving a large DV for the terminal phase, but it should not be that hard.) It is not hard to make a 5km/s, 8G missile at much lower mass than a 100MW flare. Missiles (especially those controlled by the AI) don't use evasive maneuvers, so you can tell your guns to ignore range and they'll be able to hit the missiles anyway.
|
|
joker
New Member
Posts: 8
|
Post by joker on Sept 28, 2016 4:32:46 GMT
What are you using for point defense? I have found lasers near-useless (multiple optimized 13MW lasers usually getting 1-2 kills before flak missiles enter their danger space), and the only kinetic weapons I have made with an effective range usefully greater than the danger space abused the integrator bug. I manually guide Beam drone to intercept enemy missile and drone. It's insanely effective, as drone can attack from "behind", and no ship can have armor on engine.
|
|
|
Post by Etgfrog on Sept 28, 2016 7:12:29 GMT
The problem with tracking targets comes from the ship's speed does not get added to where the missile is suppose to aim at. That I suspect is why the missiles always hits the back part of the ships or misses completely if you use a little bit of deltaV. When it comes to acceleration you need to add half of the distance / time, so if something is at X: 500, has a speed of X+= 20/s and accelerating X++= 10/s*2 then if you only calculate every second then after 1 second it should aim at 525. For a 3d game you just need to do this 3 times. Anyways, since computers calculate several times per second it can get confusing, so if your calculating 60 times per second the same would result in 500 + (20/60) + (10/120) meaning it would aim at 500.416 after 1 tick.
|
|
|
Post by Etgfrog on Sept 28, 2016 7:36:02 GMT
I forgot to include corrective maneuvering to put the missile back on course. You take the time to reach the target the multiply that by the missiles current speed on the 3 axis to give you the closest passby point. Then you take that point and find the distance on each axis to the target then add that to the estimated target location after the tick. Now that is the actual location the missile should head to.
|
|
|
Post by nerd1000 on Sept 28, 2016 10:04:12 GMT
I forgot to include corrective maneuvering to put the missile back on course. You take the time to reach the target the multiply that by the missiles current speed on the 3 axis to give you the closest passby point. Then you take that point and find the distance on each axis to the target then add that to the estimated target location after the tick. Now that is the actual location the missile should head to. You can actually achieve a collision course (though not necessarily an optimal one) by exploiting a geometry trick: if the bearing to the target is constant while the range is decreasing you are on a collision course. Therefore you can get a hit by watching how the target moves across your field of view- for example if the target moves to the left, you adjust your course left until the target isn't moving across your FoV any more. Similarly if it moves right you turn right. Do the same for the vertical axis and you have a proportional navigation homing system- it works extremely well, and can be implemented with 1950s' vacuum tube electronics (this is how the original AIM-9B sidewinder worked).
|
|
acatalepsy
Junior Member
Not Currently In Space
Posts: 97
|
Post by acatalepsy on Sept 28, 2016 12:43:50 GMT
So my views on how this shakes out are still evolving somewhat. My initial impressions of missile effectiveness have been revised upward; twenty to fifty flak missiles is a small price to pay for crippling any combat ship, as they tend to do if they intercept the target. Also, my estimations of laser effectiveness must be downgraded somewhat - my Advanced Gunship, firing its nine 60 MW lasers, could only take down twenty or so from a standing start. Also, heavy laser use largely precludes effective decoys.
On the other hand, there is still a great deal I'm unsure of regarding lasers. The unreasonable effectiveness of flak missiles against lasers is, I think, related to their tapering and wonder if the same interception logic that renders them capable of intercepting a maneuvering target also renders them vulnerable to laser defenses. Alternatively, a "buddy system" and mutual coverage might work just as well. I'm also still unsure about how lasers shake out in gun duels. My current experience suggests that they can do a great deal of damage from a long range, but this might be because their targets didn't properly armor their components or evade (roll). If lasers aren't as effective as I think in gun duels, then laser-centric defenses could become a mass tax for missile defense.
Finally, there are other PD options I haven't explored in depth yet, including drones, countermissiles, flak cannons, dedicated defense railguns, nukes (!) and more. I'm not willing to say we need decoys as a solution to missile spam just yet.
|
|
|
Post by blothorn on Sept 28, 2016 16:58:40 GMT
I have not tried flak/nukes as point defense yet---I suspect that flak is not going to help much because hit rates will be relatively low and missiles tend to be quite well-armored from ahead (although as you say, this will be reduced by target prediction---but it is extremely difficult to hit a missile off-bore while it is thrusting, even in a straight line). Nukes would be relying primarily on radiation heating, and given that most missiles are primarily armored against thermal damage I suspect that is tough going.
Defense railguns work, but suffer heavily from flight time (and scale very poorly with ship size). Even with 100% kill rates, a 10km/s railgun will only kill a missile every two seconds at 20km, and below 10km you start taking radiator damage if they are well-aimed. Distributed point defense (multiple ships) is advantageous here too, because they may choose different targets; a bunch of railgun drones with a brief "scatter" at the start of the engagement are probably even better. I think I mentioned it elsewhere, but it would be really nice if point-defense guns selected distinct targets.
And yes, laser component sniping seems incredibly powerful ATM---I should build a ship with proper anti-laser armoring (diamond? I wish that you could put composite armor on turrets...), and see how much better that holds up than the stock components (which I recall being largely metal-armored).
|
|
acatalepsy
Junior Member
Not Currently In Space
Posts: 97
|
Post by acatalepsy on Sept 28, 2016 18:05:45 GMT
Nukes are a different scale of "thermal damage", I think - it depends both on intercept logic (since the countermissile could interpose itself between the missile and target, which would be pretty hard to counter) and the yield/proximity of the nuke. You might be able to make use of these in the terminal phase, as a sort of final backstop. Just an idea at this point, of course.
(Of course, maybe one could put flares/decoys on the missiles as well? Unlike a ship, where I think some pretty basic logic could keep a missile swarm on track, decoys really would massively complicate missile defenses.)
I'm also looking into flak behavior in general. It seems like most don't miss by much more than 100 meters, and most ships aren't more than 200 meters long. Similarly, it seems that most of the flak kills I've gotten have bored a ten meter or more hole through the target ship, which seems....excessive, to say the least. A superior solution may be to set the detonation so that at the intercept velocity (2+ km/s) the cone of fragments is 300 meters wide at its "base" where it intersects the target ship. Sure, most of the fragments will be wasted that way, but it may do a better job at hitting in the first place and doing damage to radiators. It would also make it easier to do a "scatter" order, so that there's fragments coming in from multiple angles (better odds of an up the kilt shot on the engines).
Also maybe taking advantage of the (impossibly?) good mini nuclear rockets Tuna has cooked up, though that may just be taking things a bit too far.
|
|
arren
New Member
Posts: 2
|
Post by arren on Sept 30, 2016 4:10:13 GMT
Just chipping in my 2 cents as I struggle with Vesta overkill...
- The best defense against a missile intercept is the same as a modern fighter. Notch on release then maneuver to defeat their E.Generally speaking recalculating a high speed intercept requires more dV than changing an orbit a bit. Shifting out of plane slightly works wonders.
- Lasers vs Decoys.... if you have the power for lasers, your decoys are gonna be big and possibly ineffective... but you need a fair few lasers to defeat a missile swarm... as a player this can be mitigated somewhat by letting your missiles coast unguided and turning on maximum homing at the last moment (the decoys will be close enough to the enemy ships that it makes little difference) and perhaps the AI can be tweaked to try this sort of tactic too...? Hard to say. But I don't think either are unbalanced against the other per se...
- Counter missiles. With the recent fix for nuke blast calculations these are surprisingly effective. Run a counter-intercept at range (ie before the missiles intercept your ship) and blow them manually (even fly through the swarm unguided for extra funtime) and you'll knock more than a few out, and maybe even chain react some of them... It normalyl doesn't kill them all unless oyu launch a buttload, but it can reduce the numbers your PD has to cope with.
Anyone had any luck with long range high tracking railguns yet? Part of me feels a very lightweight railgun would be the counter to missiles - less thermal worries so decoys still work, and stock missiles are armoured to resist heat more than anything else, though those pointed noses would be an issue...
|
|
acatalepsy
Junior Member
Not Currently In Space
Posts: 97
|
Post by acatalepsy on Sept 30, 2016 4:34:32 GMT
Anyone had any luck with long range high tracking railguns yet? Part of me feels a very lightweight railgun would be the counter to missiles - less thermal worries so decoys still work, and stock missiles are armoured to resist heat more than anything else, though those pointed noses would be an issue... If you can *hit* them, you're probably golden; the catch is getting the hit. Drones are probably the best option here, in that you can get a couple of cheap drones in close to do the actual shooting. The problem with railguns is that the range is dependent on cross section, and missiles are a tiny target. Nuclear bombs as point defense/countermissiles is probably worth exploring, yes, but I'd be worried about having sufficient velocity to intercept a meaningful distance from your ship.
|
|
|
Post by quarkster on Sept 30, 2016 4:52:10 GMT
What are you using for point defense? I have found lasers near-useless (multiple optimized 13MW lasers usually getting 1-2 kills before flak missiles enter their danger space), and the only kinetic weapons I have made with an effective range usefully greater than the danger space abused the integrator bug. I agree that the guidance needs help; I was referring to the missile/decoy balance. (I have had decent results from lining up precise interceptions and reserving a large DV for the terminal phase, but it should not be that hard.) It is not hard to make a 5km/s, 8G missile at much lower mass than a 100MW flare. Good lasers, the artisanally crafted kind. LaserModule 1.000 GW Nd:YAG Green Laser ArcLamp GasComposition Krypton EnvelopeComposition Fused Quartz PowerSupplied_W 1e+009 Radius_m 0.027 CavityWallComposition Gold CavityCoolantComposition Hydrogen CavitySemimajorAxis_m 2.2 CavitySemiminorAxis_m 2.1 GainMedium Nd:YAG OpticalNodes 20000000 LasingRodRadius_m 0.053 Mirror Composition Gold OutputCoupler Composition Fused Quartz CoolantTurbopump Composition Amorphous Carbon PumpRadius_m 0.51 RotationalSpeed_RPM 2400 CoolantInletTemperature_K 1300 FrequencyDoubler NonlinearOptic Composition Silver Gallium Selenide OpticLength_m 0.028 OpticRadius_m 0.026 ApertureRadius_m 4.7 FocusingMirror Composition Silver Turret InnerRadius_m 10 ArmorComposition Boron Nitride ArmorThickness_m 0.015 ReactionWheels Composition Polyethylene RotationalSpeed_RPM 160 EngagementRange_km 250 TargetsShips true TargetsShots true
|
|
|
Post by Wesreidau on Sept 30, 2016 13:52:39 GMT
My point defense is intercepting the missile swarm with a single Striker nuclear missile and targeting the engine of the center-most missile. If the incoming fifty missiles don't collide with one-another, they all get fried by my shot.
Amazingly, shooting down a bunch of little missiles with one big missile was the first solution my 7 year old daughter suggested when I tried to make shoulder-watching a teachable moment, so she's either Petra from Ender's Game or the AI doesn't remember what it shot the missiles at strategically, only what the biggest heat signature is tactically. Its a compelling case for human weapons officers at least, and seven year olds are mass and volume optimized.
|
|
|
Post by RA2lover on Oct 1, 2016 21:32:41 GMT
ROFL.
In any case, although proportional navigation is much better than your generic pure pursuit homing(which is so terribad it shouldn't be used at all), it's still not optimized for spacecraft. For one, it assumes your missile has a constant lateral acceleration to its velocity vector, meaning you're losing a lot of delta-v while doing correction turns.
A thing i haven't seen taken advantage of here are nuclear decoys - a ship with a big, inefficient reactor whose only task is powering its massive cooling system.
|
|
|
Post by quarkster on Oct 1, 2016 23:01:48 GMT
ROFL. In any case, although proportional navigation is much better than your generic pure pursuit homing(which is so terribad it shouldn't be used at all), it's still not optimized for spacecraft. For one, it assumes your missile has a constant lateral acceleration to its velocity vector, meaning you're losing a lot of delta-v while doing correction turns. Proportional navigation in general has no such assumption, and is fantastic for terminal guidance in space, probably moreso than anywhere else.
|
|
|
Post by blothorn on Oct 1, 2016 23:39:11 GMT
I have thought about that--by setting output temperature very high (try tungsten/tungsten-rhenium and go up to 3000K?), you can even put out a lot of heat without a lot of radiators. The problem is that your decoy is still expendable, and I have not made a sufficiently compact reactor to compete with flares.
|
|