The present paper reviews the problem of investigating incidents in systems with dependent states. Actions of an agent in such systems may lead to changes in the system, which could be made by an agent not directly, but indirectly through other agents. A method for modelling such systems is presented in this paper. The method allows defining a “manipulated system”, in which other agents can be used to obtain the changes. A method for analysis of such systems by deriving them to the non-manipulated type is also presented.