Belief revision traditionally deals | from a first person perspective | with the question of what an agent should believe given an initial state and a revision input. This question is approached in two main ways: (i) formulating general properties a belief revision operator should satisfy and (ii) constructing specific revision operators. Reasoning about what another agent does in fact believe during a sequence of revisions is equally important as agents are not alone in the world and have to interact successfully with each other. This third person perspective, which we look at in this thesis, has received much less attention so far. In order to allow for a focused investigation, we assume the observed agent to employ a particular framework for iterated non-prioritised revision, i.e., a framework that allows for dealing with sequences of revision inputs that are not necessarily accepted by the agent. One important component of the agent's epistemic state is its core belief | a formula determining which revision inputs are accepted and which are not, a belief the agent commits to at all times.