An Interesting Orbit Pattern of an Astroid that Almost Hit Earth Several TImes (gif)

How are true orbital mechanics unfeasible? In reality it's actually pretty simple: Have every object apply a force on every other body in the system, based on its mass and distance, and use elementary physics to change the directional vector of the body. Position the bodies (or balance the constants of gravity and whatnot) so that orbits become feasible over long durations, and you've got yourself a system. Then just make some math where you can automatically generate the right mass and speed for the planet (use the existing system editor as a base, with sliders for mass and speed that affect the other) and you've got the system. Given you balance the equations and constants right, foreign bodies (for instance, Jupiter vs Earth) would have negligible gravitational attraction on each other, unless you, say, stick a bunch of Halleys on Earth and rocket it into Jupiter.

I still don't want to worry about a moon knocking a planet into the sun 30 mins in because I didn't expect it.

That would be terrible for competitive games as it is, let alone people just wanting to play for a while with out having the solar system destroy it's self at the first opportunity.

N Body is achievable in the engine, they chose to go 2 body because of some issue with how the engine tracks units of time and distance causing weird orbits in edge cases with N body. Apparently its basically toggle-able. This was all in one of dev videos.

This type of encounter is actually fairly common and does require a 3 body system to obtain. It is usually called a "rosette".

Another reason that space is terrifying....spirograph of oblivion.

I'm not certain how uber have implemented planetary motion, but based on the visuals I would bet that it's not calculated frame by frame, which would be required for the system you've outlined. There would be no simple equation that would describe a planet's position over time, and the error it would introduce could potentially fling planets around if the simulation isn't updated frequently enough.
Based on the initial conditions, you would not be able to definitively say what state the system will be in after some time, and that means (as igncom1 mentioned) you wouldn't be able to tell if your planet will eventually careen off into deep space or some other celestial body.

The current system isn't even 2 body, it's 1 body. I wasn't aware that they'd tried N body physics before, do you have a link to where that comment was made? I'd like to know more about how they handle planetary motion.

I had expected to see a shape of a penis and testicles.

There is no such thing as 1 body sim. The current system is a modified 2 body (i.e the star/planet and the satellite, not actually sure what part is modified). 2 body mechanics is Kepler orbits when one mass is much greater than the other. GalacticCow, the problem with what you're suggesting, in addition to what aevs said, is that there is no way to know if a system is stable over any duration. Real life only SEEMS like a stable system, planet orbits change over time, just not at human time scales. Another issue with N-body is that it scales poorly compared to 2-body, so if we wanted to do, say 100 or 1000 asteroids in a system, it starts becoming more of a problem to do the physics (in addition to the rest of the game). It's not really a problem you can just tweak numbers to fix, it's a problem inherent in the type of math.

They did try N-body and showed it off in a stream when they were first releasing the system editor. The problem with N-body is the lack of stability of created systems and difficulty predicting things. Because the game is not fun when the physics starts throwing random planet smashes at you with no warning, or moons stop being moons, again, with no warning.

The current system can handle 3D orbits, but it's more of an UI issue to make them usable. Also, Beta.

Other general thread comments and facts for consumption: IAU has no official binary planet definition, the ESA uses the term for Earth Moon because of the mass of both objects, not their orbits (which I think is silly, but hey, I'm not them, I prefer the center of mass definition, but Jupiter is weird then). All pairs orbit a single point (center of mass), but usually it's inside of one of those objects. Also, defining things like the size of gas/plasma objects gets hard because they have no distinct end(see the atmosphere on Earth). Also the difference between an asteroid and moon in real life is does it orbit a planet or a star, shape has no relevance (look at the moons of Mars and look at Ceres).

Should have specified what I meant when I said 1 body problem; only the motion of 1 body is modelled based on the gravitational field of the other body, so yeah, a Kepler orbit. By 2 body mechanics, I meant a system where it isn't assumed that the orbiting body is of negligible mass.

Never heard of it referred to that way, I've only heard it referred to as 2-body problem with people meaning the kepler solutions (Though I admit terrible memory for such things and a lack of extensive involvement in the subject). I'd have to dig through the live streams to see what Uber actually implemented in regards to that.