Theories of individualized evidence have been offered to show why, inter alia, we are not justified in finding a defendant legally responsible on the basis of mere statistical evidence even if the probability of his guilt is very high. Yet, there is little discussion of properties that we would want in a robust theory of individualized evidence. In my dissertation, I have four primary goals. First, I propose four desiderata that a robust theory of individualized evidence ought to possess. Then, I show how many contemporary theories of individualized evidence do not possess all four of the desirable properties. I then develop, what I call, legally relevant alternatives (or, LRA for short) - a theory of individualized evidence that is rooted in the relevant alternatives account of knowledge in epistemology. Finally, I show how LRA does satisfy the aforementioned desiderata.