How scientifically accurate are orbital mechanics in KSP?

[Ford_Prefect]

So, we just finished some important chapters in astronomy. After 8 months of not playing KSP, I got the urge to complete a mission just with math (no mission planner tool), just to see if I can.

I wondered how accurate orbital mechanics are in KSP. Can I apply all the usual laws (Kepler, two bodies etc)? I remember reading something back in the days that Kerbin's year is too short for it's path, and some other stuff.

So... can I head right into KSP with all the math from school or do I have to take some "special" physics into account?

If there is a thread about this already, please link it to me! Thanks.

[Smidge204]

It's my understanding that all orbital mechanics are two-body systems, and the Sphere of Influence system is used to avoid more complex interactions.

I would imagine Kepler's laws still apply.

You will want to account for how some objects are on "rails" instead of real-time-calculated orbits. For example a craft in a degrading orbit just grazing the atmosphere will stop degrading if you switch to another craft, because the physics simulation is stopped. Things that are actually in an atmosphere and lose focus (and are too far away from something that does have focus) are liable to just vanish since it's assumed they burn up/crash while unattended.

=Smidge=

[Eric S]

The lack of N-body physics is the biggest issue, with "on rails" calculations being the second largest issue, as Smidge204 says, as far as orbital mechanics. Other minor issues would be the lack of tidal locking and barycenters.

I'm not sure how you can say that Kerbin's year is too short for it's path, unless someone botched the precomputed orbit for Kerbin. The planets are on rails, which is to say that they're following precomputed orbits, but I suspect that the precomputed orbits were computed with the relevant gravitational factors in mind.

Once you look beyond orbital mechanics, other things pop up, like the aerodynamics model which is at best a gross approximation.

[boomerdog2000]

Disregarding decay due to the atmosphere that Smidge pointed out and n-body physics which you probably aren't using anyway I believe the math works out. I was actually using some to figure out the most efficient orbit to put my mapping satellite.

[arq]

Two-body physics with the large-mass approximation (your craft is considered massless with respect to the celestial bodies) are modeled in KSP.

[Geschosskopf]

KSP is about getting a ship from Point A to Point B. For this application, the orbital mechanics are quite accurate. Sure, it's not n-body but the difference that makes is insignificant in most cases. And the "on rails" thing is fine from a playability standpoint, so you actually have stable orbits you don't have to tweak periodically due to tiny n-body effects adding up over the long term. Plus, of course, you don't have to devote computer horsepower to keeping track of all that, either. So all that's fine and dandy.

The main place where KSP's system noticeably differs from reality is that it has no true Lagrange points because those rely on n-body forces being significant. However, the Lagrange points along the same orbit as 1 of the bodies might as well exist--KSP fakes them quite well. It's the ones where both bodies and the ship are all on the same line that can't be done in KSP, and those happen to be the ones that are most interesting.

[Ford_Prefect]

Thank you guys, that's all I needed.

[Kasuha]

Patched conics approximation and spheres of influence are actually used in real space programs for initial estimates of trajectory and required delta-v.

How accurate they are? When crossing SOI boundaries at speeds near escape velocity of the smaller body, the accuracy is very high (omitting KSP boundary crossing bugs). When crossing the boundary at low speeds, KSP starts to behave very differently from what would happen in real universe.

[Monkeh]

Other minor issues would be the lack of tidal locking...

Laythe is tidally locked to Jool.

[Kerbart]

Laythe is tidally locked to Jool.

No, it is not, because in the orbital model that KSP uses there's no such thing as tidal locking.

Laythe's orbit (and Mun's as well) is set up to appear to be tidally locked. Tidal locking suggests some kind of stable, self correcting system and that's not the case. I could claim that my car can run at 100mph for 24 hours straight without the need to refuel simply by pinning the speed dial in that position, but that doesn't mean I'm really doing that. It just appears that way. Same with Laythe's tidal locking.

[Superfluous J]

No, it is not, because in the orbital model that KSP uses there's no such thing as tidal locking.

That's about as relevant as the idea that you didn't actually land on Laythe because Laythe is just a construct inside your computer. They gave Laythe a rotation and an orbit that obviously intentionally mimics tidal locking. Saying it's not tidally locked is like saying it doesn't actually have oceans because really the oceans are a second planet that just happens to be at the same place as the "land" Laythe. Technically correct but completely irrelevant.

[Monkeh]

I don't know who to listen to and now my head hurts...

*retreats to a corner, collapses onto his haunches and proceeds to moan to himself while rocking back and forth on his heels and hitting his head on the wall*

[Geschosskopf]

Laythe's orbit (and Mun's as well) is set up to appear to be tidally locked. Tidal locking suggests some kind of stable, self correcting system and that's not the case. I could claim that my car can run at 100mph for 24 hours straight without the need to refuel simply by pinning the speed dial in that position, but that doesn't mean I'm really doing that. It just appears that way. Same with Laythe's tidal locking.

But the cool thing is, this doesn't make any difference. In gaming, perception is all that really matters. If what you see approximates reality closely enough, then it doesn't matter what's going on under the hood to create that perception. So if you can come up with a way to fake the appearance of reality that 1) provides the desired level of realism in gameplay and 2) requires WAY less computing power than real-life calculations, you have an ideal solution.

So who cares if Laythe only APPEARS to be tidally locked because the devs built it on rails so its daily rotation period equals its orbital period. For players concerned with landing ships at specific places in local daylight hours, the result is the same as if all the real-world gravity things were going on under the hood, only without needing countless CPU cycles that could better be spent on the rockets that the player actually cares about.

[Kasuha]

Tidal locking of planets is a phenomenon which only occurs over millions or billions of years. For instance Earth is still not tidally locked with Sun or Moon even after over 4 billion years. There's no difference between planets being actually tidally locked and being set up to appear so.

[Monkeh]

The moon is tidally locked to Earth isn't it?

[wasmic]

The moon is tidally locked to Earth isn't it?

Yes, it is, but Earth is not tidally locked to the Moon. If you look at Pluto and Charon, they are tidally locked to each other, mening that no matter if you are on Pluto or Charon, the other body will always look like it's at the same spot in the sky.

Writing that made me think: isn't Earth, technically, in a lunasynchronous ortbit?

[blizzy78]

Writing that made me think: isn't Earth, technically, in a lunasynchronous ortbit?

From the Moon, it might look like that, but of course it isn't, because it's not orbiting the Moon.

[Monkeh]

Ah-ha, I see. Must learn to read proper innit loike safe brah yo etc.

[Kasuha]

From the Moon, it might look like that, but of course it isn't, because it's not orbiting the Moon.

But of course Earth is orbiting the Moon. Or more accurately, Earth and Moon are orbiting their common barycenter so neither is really orbiting the other.

[Eric S]

Laythe is tidally locked to Jool.

Sorry, I wasn't very clear on that. I know that there are moons that are tidally locked. Heck, Duna is tidally locked to Ike which is tidally locked to Duna.

I was referring the forces that create tidal locking, which makes it harder to create something like the ISS where the same part of the station always faces earth. So no, not a very significant effect for most players, but someone coming from a real celestial mechanics background might notice the lack of it.