Math can only describe reality correct, when reality is actually known, otherwise you are only making the math fit, with what you think is, now. math, like any simulation that uses it, can only ever hope to describe, that which we currently understand, and if it actually does that, then we like it and keep it, so how can it ever, actually really "prove" anything? we use it as a tool, and in the case that math and the resulting visual simulation, matches that which we observe, then the math is said to be correct, but what if what we observe indirectly and from our point of perspective, is not actually what is in reality, when seen from close and more "real-time", thus directly? what if the math describes what we observe correctly, but explains what we observe incorrectly, thereby justifying the invented math itself, and not reality which is, as there appears to be a disconnect.
i personally like to "think" orbits in our solar system are perfectly circular in their nature, with balanced out forces over time, one constant orbital velocity, and sometimes one constant rotational velocity, yet seen under an angle from the earth's surface. we also have this atmospheric lensing going on, creating the "appearance" of elliptical orbits. the distance through space from source, or indirect, location of last reflection in space towards the earth is variable, and thus the light reaching us here, ON our atmospheric lens, and the time it takes to reach us here, is a variable. what do you think is more logical, planets speeding up and down in space precisely according to our perspective (foci in relation to earth's perspective?) on them? or the elliptic shape of a (hemi) spherical) atmospheric lens, the light falls on and through, warping the final image, bending the light to perspective, before it is seen on a 2D photo by one on the surface of the earth?
math is used to describe motion, as observed from point of perspective, warped through a lens, be it the atmospheric lens, or your photo camera lens. we only get long exposure images put together, forming a "projected orbit", and not actual film from space, yet. if a camera held a single long exposure in it's memory, and then added a new frame, and cut or dropped the first frame, and then put the entire windowed long exposure frame set, out as a single movie frame, and repeated the process for every new frame from the "almost same" common data set used, we could get a long exposure direct motion clip from space perhaps. if we had set a ten minute exposure, we could have all this captured light put into a set of (shared buffer) frames, to form a "real time" timelapse movie, made up of longer exposed frames, as semi direct output. just start cranking out a movie, after the set exposure time has lapsed?
even he who believes, that only what he observes is, must realize how and what he actually observes, from where and when, captured and represented with what and how... why... while math can result in a number for distance, it can never prove, the distance actually exists. only experience of it, can do that.