Hello, and a question: Will orbits be realistic?

Discussion in 'Planetary Annihilation General Discussion' started by keonskyfire, March 21, 2013.

  1. keonskyfire

    keonskyfire New Member

    Messages:
    1
    Likes Received:
    0
    Hello, this is my first post here, so please forgive my ignorance.

    From these it seems like orbits will be realistic enough. Does that mean you have manual control of a planet, or will it be automatic with targeting?
  2. BulletMagnet

    BulletMagnet Post Master General

    Messages:
    3,263
    Likes Received:
    591
    The maths for stable orbits, and to hit moving objects are too complicated for a person to do mid-game. I expect we'll be given behind-the-scenes help from a calculator, so when we say "smash into that thing" the game will return yessir, rightaway or nossir, can't do that.
  3. ninjarock

    ninjarock New Member

    Messages:
    13
    Likes Received:
    0
    I would bet that the no sir can't do that response would be due to limitations introduced in the interest of balancing. Other than that I think they will be as realistic as possible but will probably end up bending the rules every now and then.
  4. menchfrest

    menchfrest Active Member

    Messages:
    476
    Likes Received:
    55
    If we're talking about conic sections/2 body orbits/kepler orbits (all the same thing) then the math is comparatively(to full orbit sims) trivial to predict future motion of non-thrusting objects, just not for a person(I think I got calculating simple planetary transfer orbits down to 10-15mins with a calculator). Since this is 'simple' orbits, I fully expect that there is going to be the behind the scenes calculator as suggested, so you express what you want to do, the game gives you options if any, you pick one. If some one turns on the engines, then only that objects predictions needs to update.
    Tangent-> Of course if someone leaves the engines on (more than a 'short' time), then you either need to get out the textbook or call up a mathematician to replace some of your equations, but it's still about the same effort to compute I think...

    For the extra realistic orbits, and the fun quirks that pop up, n-body simulation, future prediction gets much harder (in the strictest sense impossible, but you can get close enough), I don't actually expect anyone to be able to do that by hand (well, one or two, max). This HAS to be done by the game, as doing that many integrals in your head is hard, and errors can lead to weird results in complicated systems. The fun problem with this though is that any time ANYONE turns on an engine, any future predictions you have go out the window and you need to start over from scratch. So if engine firing happens a lot, this could cause problems. Note, we don't need to include everything in the orbital simulation, You could count stars,planets, moons and asteroids only, as satellites/units/projectiles are tiny in comparison and would have negligible effect, this does make it more manageable, although satellites may be debatable.

    To be fair, you can actually do a lot in both systems if you assume infinite fuel, so I don't think you would need to break or bend the rules at all.

    tl;dr- yes calculators will be needed, the complexity is depends on how realistic you want, and last but not least, it will be awesome.
  5. CrixOMix

    CrixOMix Member

    Messages:
    104
    Likes Received:
    0
    From what I've seen, it's fairly trivial for computers with today's technology to calculate within decent accuracy any n-body system. As said above, lots of changes in vectors due to thrusting will force recalculations of the entire system, which could be problematic, but I have a hunch it will still be fine...

    There's a lot of complicated math that goes into realistic orbits, but once it's programmed in, the computer can crunch it like it's nothing.
  6. menchfrest

    menchfrest Active Member

    Messages:
    476
    Likes Received:
    55
    People keep saying it's trivial to do n-body on a modern computer, but there's an issue with growth and scalability. If we talk about 30 bodies, yes I agree it is trivial, but doing 300, that's going to take much more than 10x the crunching. So it depends on what exactly you include, if we're going with all the units also, it'd be a bad idea to do n-body on that, on just asteroids and bigger, n-body could be fine. If we include satellite, well the question becomes, how many satellites do we expect?

    Yes we can find n-body simulators that can run on most computers these days with a fairly decent number of objects, but those aren't trying to run a game on top of those bodies...

    BTW, I'm not trying to say 'n-body bad, no!', I'm just given a bunch of thought to this and I'm very curious as to where uber is going to finally set things.

Share This Page