An evidential simulation or Bayesian simulation is a simulation whose transition probabilities are given by the output of a predictive world model conditioned on the current observed state. The next state is obtained by sampling from the model's predictions and then updating the state as if the sampled event or outcome had come to pass, simulating time progression via hallucination.
Whether a dynamical process qualifies as an evidential simulation depends on its semantic properties: if the state can reasonably be called an "observer state" (a record understood as evidence for an underlying state that is only partially observed) and the time evolution operator a "predictive world model" (interpreter of the observer state as evidence for the hidden state that outputs a distribution over next observations conditioned on the update). In a pure or solipsistic evidential simulation, transition probabilities are determined solely by the world model; if extrinsic factors also play a role, it is mixed.