Let me give an example at first: All of the calculations that are carried out in quantum chemistry rely on approximation methods to the Schrödinger equation. While these methods sometimes give quite good approximative results, they are computationally very hard.
I wonder whether one could find a theory, which agrees with the approximative results of the Schrödinger equation, but is computationally less hard.
This question may be asked in a more general context:
If you have a working, well known theory $T$ which predicts certain results $R$, can there be a theory $T^*$, which makes exactly the same predictions $R$ in a less computationally hard way?