Why does the recently proposed simple empirical formula for the lowest excitation energies work so well?