If you shift the universe five meters to the left, there is no observational change. If you rotate the entire universe, the inertial frame is also rotated, and there is no observable change. If you freeze time in the universe for one billion years, then resume it, there is no observable change. If you boost the universe's velocity to five m/s, there is no observable change. You see where I'm going with this? If you increase the distance between all particles by 5 meters, is there an observable change? I know this has something to do with Lorentz/Poincare invariance, Galilean groups, etc. But I have no idea what those words mean. Imagine you have 3 gravitational particles of equal mass, with various velocities. If you scale their distances and velocities by five, will the evolution of the system change? If position, distance, and time are all arbitrary notions, what is mathematically the observationally best way to describe a classical system? Angles? Shapes?