We study the problem of discriminating between non‐orthogonal quantum states with least probability of error. We demonstrate that this problem can be simplified if we solve for the error itself rather than solving directly for the optimal measurement. This method enables us to derive solutions directly and thus make definite statements about the uniqueness of an optimal strategy. This approach immediately leads us to a state‐discrimination analogue of Davies Theorem.
In the course of this, a complete solution for distinguishing equally likely pure qubit states is presented.