Absolutely. It has to do more than just explain current and past observations it needs to be used to predict future ones better than competing theories. If it can't be used to make predictions that can be tested then it hardly qualifies as a theory, more a hypothesis.