Calculations show that the surprising ``orientation independence'' of the Co/Pd interface anisotropy found experimentally is an accidental consequence of the precise degree of strain exhibited by this system and is not generally expected. The conclusions are supported by the good agreement between experiment and our electronic structure calculations for the anisotropy magnitude, and by the close correlation with the results of a simple symmetry-based theory in which the effects of strain are clear.