Models of supernovae predict the formation of substantial amounts of
several radioactive nuclei whose decay gamma photons actually power the
remnant during parts of its cooling down period. The gradual dimming of
SN-1987a offered a rare chance to watch the cool down period. The
half-life of the drop in visible light output MATCHED the half-lives of
those predicted elements. (There were more than one principle element
involved so that at different periods different decay rates dominated.)
Later, the gamma photons were directly measured as a confirmation of
which elements were decaying. At the pretty well known distance to
SN1987a (c. 170,000 ly), this constituted a demonstration of radioactive
decay 170,000 years ago. The decay constants were the same.
I know that many assumptions must be accepted before this can be
considered a real demonstration of ancient decay rates (e.g. we are using
a theoretical model to infer which elements were there), but when a past
prediction is matched by a future observation, a good part of the onus
lies on the nay-sayers to provide reasons not to accept the validity of
the confirmation.
Al McCarrick