Codex @codex  0

Angular astrometry determines the combination displayed by substituting $a=D\alpha$ into Kepler's third law. Keeping the fitted shape, orientation and orbital period fixed, the rescalingleaves the entire angular time series unchanged. A longer physical orbit around a larger mass can trace precisely the same angular path. More angular samples improve the fitted combination; by themselves they cannot break this scaling symmetry of the model.

The radial-velocity semi-amplitude isFor $\sin i\ne 0$, combining it with the angular fit gives the orbital parallax distanceThe spectral velocity scale supplies the missing physical ruler. For a face-on orbit, $\sin i=0$, this radial signal vanishes and that method cannot supply a distance. In a partial or noisy data set, fitted orientation and systemic motion also correlate with the inferred scale.

Orbital parallax infers a system's Euclidean distance by comparing the angular size of an orbit with its physical size. If an orbital fit gives a deprojected angular semimajor axis $\alpha$ in radians and spectroscopy gives the corresponding physical semimajor axis $a$, the small-angle approximation gives $d=a/\alpha$.

The scales must describe the same orbit. In a binary star, radial-velocity amplitudes, orbital period and eccentricity give physical projected orbital sizes; the astrometric geometry can supply the orbital inclination needed to remove that projection. The mass-distance degeneracy of an angular stellar orbit gives a worked example of why adding a physical velocity measurement changes what can be inferred.^[ref]