| Author |
Destroyer & Frigate Balance |
Jim Starluck
Marshal
Templar Knights
Joined: October 22, 2001
Posts: 2232
From: Cincinnati, OH
|  Posted: 2008-10-29 16:16
Quote:
|
On 2008-10-29 15:38, Fornax wrote:
I'm sure there's a way to approximate a leading targeting solution without a lot of calculations. Pretty much, anything would be better than what it's getting. I'd think that a +x gu's calculation based on the size of the target would be sufficient...Other options such as measuring the distance between the missle/target spot could adjust/reduce the lead targetting.
|
|
I mentioned something like this to Draf. He said it would also cause missiles to miss when the target is sitting still.
_________________
If at first you don't succeed, get a bigger space battleship and try again.
|
XcepticZP
1st Rear Admiral
Joined: October 23, 2006
Posts: 22
| Posted: 2008-10-30 07:11
Quote:
|
On 2008-10-29 16:16, Jim Starluck wrote:
Quote:
|
On 2008-10-29 15:38, Fornax wrote:
I'm sure there's a way to approximate a leading targeting solution without a lot of calculations. Pretty much, anything would be better than what it's getting. I'd think that a +x gu's calculation based on the size of the target would be sufficient...Other options such as measuring the distance between the missle/target spot could adjust/reduce the lead targetting.
|
|
I mentioned something like this to Draf. He said it would also cause missiles to miss when the target is sitting still.
|
|
Obviously someone is trying to hard to dismiss these ideas. It is absolutely trivial to implement. Shooting any projectile, currently, uses a pretty efficient system. This system is based on speed so this argument about the "sitting still" target is really just an insult to anyones intelligence. And if you decide to counter this by saying well "if we implement this like you say, then missiles have to use a lot of computation to adjust every second", well then I'll say, they already do. So that's a moot point.
Second. This game uses over 600mb of ram, and isn't even all that "big" compared to games these days. If they can bombard us with 600mb of unneeded size, then they can bombard our CPU with this simple, but needed, game calculation. This is under the assumption that a fix for missiles would require more processing power, which I seriously doubt, as the system already seems to be in place. It just needs a little tweaking.
_________________
~~~~ZOKI XCEPTIC~~~~
|
Sixkiller
Marshal
Courageous Elite Commandos
Joined: May 11, 2005
Posts: 1786
From: Netherlands
| Posted: 2008-10-30 11:13
Quote:
|
On 2008-10-30 07:11, XcepticZP wrote:
Quote:
|
On 2008-10-29 16:16, Jim Starluck wrote:
Quote:
|
On 2008-10-29 15:38, Fornax wrote:
I'm sure there's a way to approximate a leading targeting solution without a lot of calculations. Pretty much, anything would be better than what it's getting. I'd think that a +x gu's calculation based on the size of the target would be sufficient...Other options such as measuring the distance between the missle/target spot could adjust/reduce the lead targetting.
|
|
I mentioned something like this to Draf. He said it would also cause missiles to miss when the target is sitting still.
|
|
Obviously someone is trying to hard to dismiss these ideas. It is absolutely trivial to implement. Shooting any projectile, currently, uses a pretty efficient system. This system is based on speed so this argument about the "sitting still" target is really just an insult to anyones intelligence. And if you decide to counter this by saying well "if we implement this like you say, then missiles have to use a lot of computation to adjust every second", well then I'll say, they already do. So that's a moot point.
Second. This game uses over 600mb of ram, and isn't even all that "big" compared to games these days. If they can bombard us with 600mb of unneeded size, then they can bombard our CPU with this simple, but needed, game calculation. This is under the assumption that a fix for missiles would require more processing power, which I seriously doubt, as the system already seems to be in place. It just needs a little tweaking.
|
|
I think they assumed x was a fixed number, and you are assuming x is a variable. If it was a variable it would indeed work i guess, like x=speed or x=whatever*speed
Perhaps an idea? would fix the sitting still
_________________
|
Gejaheline
Fleet Admiral
Galactic Navy
Joined: March 19, 2005
Posts: 1127
From: UGTO MUNIN HQ, Mars
| Posted: 2008-10-30 12:40
It's not the client's RAM or CPU that's the problem; it's the fact that the server has to keep track of and perform all of these calculations.
Additionally, the calculation for leading a target, having tried to program something similar and failed, isn't particularly trivial, and it's compounded both by computers generally taking a lot of power to perform trigonometry (depending somewhat on how it's programmed) and the fact that the calculation needs to be performed every single tick.
It seems simple: All you have to do is establish how far the target will have moved between now and the impact time, and aim at that point.
In order to do that, you need to know how fast the target is going, what angle it's moving at, how fast you are going, the angle you are moving at, and the distance between you and the target.
Calculating the time-to-impact is relatively simple in the real world: You take a measure of how far you are from the target, wait for one second, measure it again, and calculate the rate of change. In a digital universe, you can't really do that since the calculation is occuring during a single frame and nothing is actually moving. You need to calculate the time to impact with the old time = distance/speed formula, applied to your speed and distance variables. Problem is, though, that if the target is moving away from you at the same speed as you're moving, a simple calculation like this will still give a time-to-target in spite of the common-sense knowledge that you're never going to hit the target. Additionally, your missile isn't necessarily directly facing the target, so you need to perform two computationally-intensive trig calculations to determine what the actual closing speed is, and use that as your velocity (And while you're at it, deal with the divide-by-zero/negative problems that fast fleeing targets will produce). Once you've got your accurate time-to-target, you can predict how far your target will move in that time. This is fairly simple, since it's just speed*time. The actual aimpoint will require another trig calculation so that the aimpoint can be correctly plotted in front of the target ship.
And then you'll need another trig calculation to tell the missile what direction it should be pointing, but that's already in the game so I'll let it off with that one.
And even once all that's done, your missile is still going to be slightly inaccurate, because the calculations asssume the missile has perfect manouevrability. In a perfect calculation, the aimpoint will remain in one place in space as the missile closes in, but the above procedure will produce an aimpoint that moves if the missile isn't pointing perfectly at it.
Then you'll have to repeat the calculation every tick, for every missile on the server. Plus a variation is used for firing cannons, although they take up a lot less power because they only check once per projectile.
So in short, missiles take up a lot of processing power already, but if you really wanted you could make them take up about three times more by making them guide properly.
It's been about a million years since I last attempted to seriously calculate this kind of thing, so maths boffins do feel free to correct/expand on this post.
[ This Message was edited by: Gejaheline on 2008-10-30 12:43 ]
_________________
[Darkspace Moderator] [Galactic Navy Fleet Officer]
|
XcepticZP
1st Rear Admiral
Joined: October 23, 2006
Posts: 22
| Posted: 2008-10-30 22:05
Ok, Geja. I don't know what you're trying to prove but its kind of silly to say "well I tried to code it and failed, therefore it's too hard". Yes it's too hard, but that's just based on your effectiveness as a programmer.
Second. I have no idea how the server/client calculations are partitioned, but my first guess(seeing as this isn't a "popular" game, therefore cheating isn't a big problem at all) is to try send as much computation as possible to the client. If this isn't the case with DS, then we clearly know the reason why missiles lag the server heavily or whatever.
Third. You say: "missiles take up a lot of processing power already", yes they do. And that is precisely my point, you just didn't understand it. My point is that this is already done every x slices of time. So why not simple aim slightly ahead of the ship, distance infront of the ship ofcourse being proportional to the speed the ship is going and the difference between the ship's heading and the missiles heading. This heading bit might need trig stuff, but the rest is vectors. You'd be surprised how much just this slight change can do for the effectiveness of missiles. Because both these functions are implemented and done efficiently by client/server. It's just that they are not being used _together_ in missiles.
Oh, and most of these calculations won't be using any trigonometry. They'll be using vectors. Perhaps you should read up on them next time you attempt to solve this kind of problem. It might actually make it easier for you, so you don't "fail". 
**Came back and added this.
So you're saying that missiles should just what, explode, when they detect they can't catch up to a target? That's just silly.
One more thing. All those funky calculations you were talking about "time*speed" and "time to target" and all that. Well, again, you need to replace all that with vectors. I hope you've been taught that at your school. Don't mean to insult you with that comment, but it's obvious that you're misunderstanding something, and I think it's because you're not well informed about vectors and what their purpose is.
[ This Message was edited by: XcepticZP on 2008-10-30 22:09 ]
_________________
~~~~ZOKI XCEPTIC~~~~
|
Gejaheline
Fleet Admiral
Galactic Navy
Joined: March 19, 2005
Posts: 1127
From: UGTO MUNIN HQ, Mars
| Posted: 2008-11-03 18:34
Quote:
|
On 2008-10-30 22:05, Xceptic ZP wrote:
|
|
Quote:
|
On 2008-10-30 12:40, Gejaheline wrote:
...so maths boffins do feel free to correct/expand on this post.
|
|
Damn, forgot to say "and if you're going to be an insulting smartarse, at least show me how you'd do it." I noted that I failed to write a gunnery control program in order to reinforce the fact that I don't know much about this area of programming, as I said multiple times. If I started talking about nuclear sequence alignment algorithms, I'm sure your eyes would begin to glaze over and you've have very little idea of what the hell I'm on about, but I'm not going to call you stupid for not knowing something outside your area of knowledge. So unless you're going to fill in some of the gaps in my knowledge, without implying that, wow, you're misunderstanding this totally elementary concept and thus your IQ must be cretinous.
Assuming you're talking about Euclidean vectors containing a length and angle, I don't recall any methods for projecting them onto a Cartesian map (which is what I'm presuming DS uses) without converting them into X and Y displacements, which tends to require trigonometry. And if you're using vectors containing the X and Y coordinates of the origin and terminus, you don't have the heading and distance values.
So, presumably I'm missing some magical property of vectors that allows you to calculate both positions and directions within a coordinate system that sidesteps all of the trigonometry that I've encountered in practically every video game and program I've had the pleasure of examining the code of.
Request to a mod that this topic might be better served in a thread of its own. Also, need to find a dev and find out how fire control is handled.
[Edit] Okay, after consulting with an eminent mathematician, there is indeed a way to render most of it down into quadratic/quartic equations (I suppose it would make for an easy way to calculate missile trajectories through more than three dimensions with a bit of adjustment, at least). Might have to rewrite the game somewhat, and I'm not sure if you'd be able to have a heading indicator on the interface, but you can indeed do it without trig. If you optimise the trig equations (for example using the Taylor Series to approximate), they're basically the same with regards to CPU cost, though.
[ This Message was edited by: Gejaheline on 2008-11-03 21:57 ]
_________________
[Darkspace Moderator] [Galactic Navy Fleet Officer]
|
Russian Roulette with Muskets
Grand Admiral
Joined: September 04, 2002
Posts: 393
| Posted: 2008-11-04 15:25
did i mention that frigate with heavy narrow forward arc Lasers are still the coolest option?
_________________
- In firepower we trust.
- I'm not buying this!
-we ran out of firepower.
|
Jim Starluck
Marshal
Templar Knights
Joined: October 22, 2001
Posts: 2232
From: Cincinnati, OH
| Posted: 2008-11-04 17:24
Lasers require them to get into point-blank range to be effective against Cruisers and Dreads--the range at which Cruisers and Combat Dreads can hammer them into scrap and Assault Dreads can pop them like balloons. They also have high power requirements which require Destroyers to fly slower which hampers their ability to get in and out quickly, which they have to in order to survive.
_________________
If at first you don't succeed, get a bigger space battleship and try again.
|
Russian Roulette with Muskets
Grand Admiral
Joined: September 04, 2002
Posts: 393
| Posted: 2008-11-05 09:28
Quote:
|
On 2008-11-04 17:24, Jim Starluck wrote:
Lasers require them to get into point-blank range to be effective against Cruisers and Dreads--the range at which Cruisers and Combat Dreads can hammer them into scrap and Assault Dreads can pop them like balloons. They also have high power requirements which require Destroyers to fly slower which hampers their ability to get in and out quickly, which they have to in order to survive.
|
|
Youre not supposed to fly int othe weapon arcs of said dreads.
Cruisers dont have the fire power to bust frigs that fast either.
The MO would be: gun and fckn run! (and do it in teams)
Its not supposed to be hard counter ala "will always beat the $@%!^?** outa dreads".
+ the energy management needs will make sure that idiots are weeded out of the ranks of fearless beam frig captains and make sure that a prologned battle is impossible, requiring proper tactics instead of "stay realy close to the rear and hammer it till the ship xplodes"
_________________
- In firepower we trust.
- I'm not buying this!
-we ran out of firepower.
|
doda *EP5 no longer exception...*
Grand Admiral
Joined: December 11, 2005
Posts: 1012
From: happy land
| Posted: 2008-11-05 18:34
In my opinion the best solution to this is adding a bridge between small ships and large ships.
Like a kamakazi transport. Very damaging SD but no jd, no armour, and little speed. Instead the tranny will depend on large ships like stations and dreads to drag it around.
Large ships will try to stay away from it cause they will not be able to out run it from it if they get close.
But small ships are relatively cheap, fast, and expendible so they can stay out of those tranny's SD range while slowly killing them.
We also need the exact opposite, large ships which are strong against small ships but weak against large ships. Ie like an interdictor station, a station with lots of assault beam weaps, tractors, and a dictor but little armour protection.
_________________
Please resize your Admin - signature
VCA since June 5th 06
|
Russian Roulette with Muskets
Grand Admiral
Joined: September 04, 2002
Posts: 393
| Posted: 2008-11-06 10:01
Quote:
|
On 2008-11-05 18:34, doda *EP5* (No Longer Exception...) wrote:
In my opinion the best solution to this is adding a bridge between small ships and large ships.
Like a kamakazi transport. Very damaging SD but no jd, no armour, and little speed. Instead the tranny will depend on large ships like stations and dreads to drag it around.
Large ships will try to stay away from it cause they will not be able to out run it from it if they get close.
But small ships are relatively cheap, fast, and expendible so they can stay out of those tranny's SD range while slowly killing them.
We also need the exact opposite, large ships which are strong against small ships but weak against large ships. Ie like an interdictor station, a station with lots of assault beam weaps, tractors, and a dictor but little armour protection.
|
|
this is as stupid as it reads.
Creating a counter for against dread that not only requires said dreads to deploy it but also is, in essence, AN EFFING MISSILIE but with FAIL attached.
And the counter to small ships swarming you is small ships of your own, not more big ships dedicated to killing small ships. Thats just stupid.
Do not post when under the influence of drugs.
_________________
- In firepower we trust.
- I'm not buying this!
-we ran out of firepower.
|
Gejaheline
Fleet Admiral
Galactic Navy
Joined: March 19, 2005
Posts: 1127
From: UGTO MUNIN HQ, Mars
| Posted: 2008-11-06 10:48
Probably not the best of ideas.
The idea isn't to have hard counters, ships that automatically go "hah, I kills j00" just by turning up.
An anti-dread destroyer would yes, be more effective against dreadnaughts, but it should have a difficult job of it and will probably require many of them to actually kill a dreadnaught.
Conversely, dreadnaughts are designed to be big and slow, and to kill other big and slow things. A big slow thing designed to kill small and fast things wouldn't get very far unless you had some kind of stupidly-long-range beam weapon that could shoot as far as a small ship's cannons and missiles.
Plus if you had both dread-ganking destroyers and destroyer-ganking dreadnaughts, they would become the only things anyone would ever use.
_________________
[Darkspace Moderator] [Galactic Navy Fleet Officer]
|
|
English
Deutsch
Nederlands
)
