On Subnautica's flaws in celestial mechanics

Next I show an:

Example of a better sky simulation

It is better but not yet realistic, as we explain below. It remains simple to program and retains some of the features of the original sky, like regular eclipses of the sun.

In this example the planet spins, and we are at lattitude +50°. The relation between the planet and the sun is like on Earth, except that we do not simulate the change of the sun's position in the celestial sphere as days pass in a year (due to the orbital motion of 4546B around its sun). The epoch has been chosen to be somewhere in the summer, so that the night lasts 3 times less than daylight, and so that the sun passes through the zenith (just for fun).

We also aligned the magnetic pole with the spin pole, which not realistic but is simpler for navigational purposes in the game.

Planet 4546B is tidally locked to its gray moon, whose orbit around 4546B is contained in 4546B's equatorial plane (this is not the case for Earth's moon). Hence the gray moon does not appear to move in the sky (its apparent position depends of where you sit on the planet). However, it still has to experience phases.

The red moon has a circular orbit around the planet, so its apparent motion is a superposition of its orbital motion and of the planet's spinning. Depending on parameters, tracking it could be quite fun. It would likely (but this is debatable) be a serious source of sea tides. Its orbital plan has been chosen to make an angle of 75° with the equatorial plane of 4546B. It hides the sun periodically (with an absolute regularity), with a period that is not in resonnance with the duration of a day: the ratio is about 1.62 (less than the original 2.75, so a day will typically contain only one solar eclipse, but with that choice the red moon has an interesting apparent trajectory in the sky). It hides the gray moon on some occasions, that come irregularly. I did not decide what is its actual size, only its apparent size. I did not try to determine the Roche limit. The angular size of the moon should change noticeably because the distance between the player and the moon varies (the moon, being a moon, is smaller than the planet, and is not far enough for this distance to be nearly constant), but this change is not simulated here.

Another inaccuracy is that the gravitational interplay between the two moons would perturb such a nice set of orbits (unless some stable resonnance phenomenon occurs), as could also the sun's influence. There is a possibility (to be studied) that such a system could be unstable and thus not exist long enough for things to settle and life to prosper there.

So it solves many orbital/geometrical mistakes of Subnautica's sky but not all the dynamical and astrophysical ones.

There are of course many alternatives, depending on the degree of realism that you look for and on the particularities of Subnautica's sky you are ready to give up on.