How well can nonadiabatic dynamics forecast an experiment?