Multilevel Flow Modelling can be used to identify causes or consequences of anomalies in process systems. The models can be used to identify numerous possible propagations of causes or effects but cannot distinguish between likely and unlikely causes or effects. We present a method for identifying likely and unlikely effect propagations in a given process window from Monte Carlo Simulations. We show that the joint probability of effects can be used to determine the probability of individual propagation paths. The analysis allows to identify subsets of the process window where certain effect propagations are more likely. The method enables prompt identification of likely propagations of effects from process anomalies.