|
|
| Author |
Thread Statistics | Show CCP posts - 4 post(s) |
Stefan Stahl
Seituoda Taskforce Command
813
   |
Posted - 2014.10.30 18:03:00 -
[1] - Quote
CCP Rattati wrote:Have you experienced a change in Swarm efficiency, now that we nerfed their turn radius by another 10%? There is an issue with 'observing swarm behavior' from a pilot's perspective. If you can see them coming you're likely hovering and thus not being evasive. If you're running from them you can't see them.
Can you pull data from the game on how often swarm missiles actually reach their target? Are those numbers different for HAVs and DS?
On the topic itself:
Personally I don't see a use-case for evading swarms as a pilot. Either we nerf swarms until they have the maneuverability of a MLT DS (those hit often enough!) or swarms will always be able to hit their target.
If I cover behind a building I do it to break the lock on the third volley. However that strategy is only possible within ~2-3 s of the first swarm being launched, thus only if I spotted the launch itself or if I know that the swarmer is very close for some other reason. Swarms that are in the air are expected to hit 100% from my point of view.
Anyways, since my gunship Myron can tank three volleys from a single source (boosters!  ) I'll try to keep an eye on swarm missile behavior this weekend.
Do you accept invoices for costs incurred during data acquisition? |
Stefan Stahl
Seituoda Taskforce Command
817
|
Posted - 2014.10.30 18:40:00 -
[2] - Quote
manboar thunder fist wrote:Flying close to buildings to avoid swarms is begging to be thrown about by their impact into the building leading to certain death I love it when - due to lag - the missiles locally pass under you and then push you in the direction of the swarmer rather than away from it. That always screws with me.
That's a consequence of hit detection being server side but knockback being computed locally. It's a good reason not to try your luck swarms by covering behind buildings. |
Stefan Stahl
Seituoda Taskforce Command
817
|
Posted - 2014.10.31 15:14:00 -
[3] - Quote
CCP Rattati wrote:Have you experienced a change in Swarm efficiency, now that we nerfed their turn radius by another 10%? Do you have any information as to how the "maximum swarm turn angle" should be interpreted?
Earlier I was trying to check whether the value you modified could be interpreted as "swarm missile turn angle per second", so I excel'ed out the approach of a swarm missile on a DS that is closing in on the swarmer (head on) at 50 m/s at an altitude of 75 meters (here's a a graph). Turns out in the given scenario the missile would've been trying to turn at over 90-¦/s at one point during the ascend. On the other hand I've never seen a swarm missile have trouble hitting a dropship neither before hotfix delta nor after it.
I have two explanations for this:
- The "maximum swarm turn angle" is a different parameter entirely.
- DS can out-maneuver swarm missiles, but in scenarios where that is effective (e.g. running at the swarmer) the missile can afterwards outpace the DS once it's heading in the right direction. Pilots can effectively buy themselves a second or two of time before impact by making the missile go wide, but that doesn't help unless those 1 or 2 seconds make the missile run out of time or distance.
I'll try to make some hands-on observations if I get a chance to go online today. |
Stefan Stahl
Seituoda Taskforce Command
818
|
Posted - 2014.10.31 19:04:00 -
[4] - Quote
Stefan Stahl wrote:I'll try to make some hands-on observations if I get a chance to go online today. Apparently there are less pilots about when I'm using a SL than when I'm using a DS... 
Anyway, this looked like a nice problem and I'm on my day off, so I pulled matlab on this.
Under the assumption that we're talking about the "Swarm Launcher missile maximum turning angle per second" I wanted to compare how much of an advantage we get by turning this from 90-¦/s to 70-¦/s. I have a similar concept in on e of my projects so I wanted to look into it.
As the scenario I chose to look - again - at a DS that is going at full speed (50 m/s) towards a Swarmer at a constant height of 75 m. As soon as within range the Swarmer launches a single volley at the DS. The missiles then attempt to home in on the DS as it passes over the position where they were launched. Missile acceleration and maximum turning angle per second are both considered.
Here's an updated graphic, now at a simulation speed of 10 ticks per second. The dropship starts from the right (red line), the Missile starts from the center (blue line). Regardless of maximum turning angle per second the missile soon catches up to the DS and hits it. Reducing the maximum turning per second from 90 to 70-¦/s buys the DS pilot 0.7 s of time.
I'm not going to tweak this until I don't get any hits anymore because the DS isn't being evasive in this example. If anyone can show me what an evasive maneuver looks like for a DS and provide descriptive numbers I can incorporate this. The math I used is 3d, so anything goes. 
P.S.: Don't forget that we might be talking about a different variable entirely. This is mostly for my amusement. Though it would explain why we haven't seen much of a difference in the Swarm's ability to hit dropships. |
Stefan Stahl
Seituoda Taskforce Command
820
|
Posted - 2014.10.31 21:02:00 -
[5] - Quote
Kallas Hallytyr wrote:@Stefan: I think one manoeuvre that pilots would try would be to flip the ship to being perpendicular to their line of travel while letting go of the thrust (to prevent it being a curve) and then hitting thrust under an afterburner to get as radical a course change as possible. Not sure if that's clear enough for you... Do you have any idea how long an afterburning ADS takes to reach full speed? Even if it loses height while at it. If I know that I can tell you whether that maneuver is viable. |
Stefan Stahl
Seituoda Taskforce Command
820
|
Posted - 2014.11.02 12:01:00 -
[6] - Quote
I twiddled around with matlab a bit to see how maneuverable a DS has to be to be able to dodge a missile. As the basic scenario I chose a SL in the center of the coordinate system and a static DS 150 east of it at the same height (to facilitate visualization in 2D). The swarmer launches a missile that tracks at a given maximum turn rate. The DS has the task of maneuvering such that it maximizes angular velocity towards the missile at any moment. Notice that this is not optimal behavior. Ideally you'd want a combination of dodging until the SL passes you and then fleeing to outlast it's fuel reserve. This test is merely looking into how well a DS can outmaneuver swarms, not on optimal flight strategy.
I plotted three scenarios:
A)
- The missile has a maximum turn rate of 70 -¦/s (as of right now)
- The DS has a maximum acceleration of 10 m/s (as of right now without AB)
B)
- The missile has a maximum turn rate of 70 -¦/s (as of right now)
- The DS has a maximum acceleration of 25 m/s (I think that's how the 150% bonus of the AB works)
C)
- The missile has a maximum turn rate of 48 -¦/s (fantasy number)
- The DS has a maximum acceleration of 25 m/s
In C I chose the maximum turn rate like that because that's the first time I don't get an impact on the DS. For reference, I count as an impact when the missile is within 10 m of the DS center of mass, which I think is reasonably close to how it works in the game.
Anyway, here are the plots. A-B-C are left to right. We're looking top-down on the battlefield. In each plot the missile starts in the center and the DS starts on the right. Impact is indicated if it happens, in C you can see there is no impact.
I can share that A and B are pretty much what it looks like from a DS perspective with and without the afterburner. I managed to test this for a while thanks to a couple of stubborn MLT SL users yesterday evening (no redberries where harmed during this particular experiment  ).
I did this mostly for the fun of it. There's some neat math involved that'll be helpful to me in other projects. I'm sharing though to show how you can design the gameplay using a quick mockup simulation. This reduces turn-around times significantly. I can show you the approximate effect of 100 combinations of values within an hour. On the live server you'd have to test each combination for two weeks to get conclusive data.
Anyway, if my assumptions turn out to be true (1. we're discussing maximum SL turn rate and 2. that's actually how afterburners work) I've shown that even if right now a successful dodge is unlikely (confirmed ingame), we aren't far away from a point where this may become possible. However I can't say anything about whether attempting to dodge three consecutive swarms is ever going to work or what the effect of this is going to be on the SL's effectiveness against HAVs and LAVs. |
Stefan Stahl
Seituoda Taskforce Command
822
|
Posted - 2014.11.02 14:24:00 -
[7] - Quote
CCP Rattati wrote:Matlab, my all time favorite software You know things get serious when during a discussion someone pulls out his laptop and boots matlab. That's an engineer's version of raising your fists. 
Anyway, regarding ABs and out-maneuvering missiles:
Trouble is that the AB increases acceleration, not top speed. So right now ABs are what you'd use to out-maneuver missiles.
We could change that and get a very different game where a pilot has two options:
A) Have ABs grant a straight line speed bonus so you can outrun swarms in a straight line
Right now a 20% speed bonus would be sufficient. It'd take a DS 6 seconds to get to missile speed and be safe forever.
If it is deemed necessary to have ABs grant an acceleration bonus it could be worthwhile to think about it also conveying a malus to moment of inertia. That way you accelerate fast in a straight line but can't really turn all that well.
B) Fit engine upgrades (the old ones with the torque bonus) and nanofibre structures (the old ones with the mass reduction) to outmaneuver swarms
We'd have to go comparatively low on the missile turnspeed unless we want such a DS to feel like constantly having a current AB on. At 70 -¦/s a swarm does a full circle in about 5 seconds. That means you'd have to dodge every missile volley twice. At 40 -¦/s a missile would be capable of doing exactly one circle in it's 9 second lifespan.
[Edit]
Also, you go and have Sunday. Sundays are not the time where you do work related stuff. |
|
|
| |
|