This paper examines a methodological question arising in orbital mechanics and the foundations of physical modeling: does successful prediction of a relativistic observable uniquely determine the underlying dynamics? Using Mercury perihelion precession as a representative case, several dynamically non-equivalent orbital systems are constructed that reproduce essentially identical relativistic-scale perihelion advance while differing in local temporal structure, phase-space evolution, and orbital
