Let's say, you have missed thinking about some particular use case in a problem and a very large test case fails. how do you go about identifying the use case ?