ABSTRACT The ability to accurately predict the evolution of tidally stripped haloes is important for understanding galaxy formation and testing the properties of dark matter. Most studies of substructure evolution make predictions based on empirical models of tidal mass-loss that are calibrated using numerical simulations. This approach can be accurate in the cases considered, but lacks generality and does not provide a physical understanding of the processes involved. Recently, we demonstrated that truncating NFW distribution functions sharply in energy results in density profiles that resemble those of tidally stripped systems, offering a path to constructing physically motivated models of tidal mass-loss. In this work, we review calculations of mass-loss based on energy truncation alone, and then consider what secondary effects may modulate mass-loss beyond this. We find that a combination of dependence on additional orbital parameters and variations in individual particle energies over an orbit results in a less abrupt truncation in energy space as a subhalo loses mass. Combining the energy truncation approach with a simple prediction for the mass-loss rate, we construct a full model of mass-loss that can accurately predict the evolution of a subhalo in terms of a single parameter ηeff. This parameter can be fully determined from the initial orbital and halo properties, and does not require calibration with numerical simulations.