This paper presents a hybrid pronoun resolution system for German. It uses a simple rule-driven entity-mention formalism to incrementally process discourse entities. Antecedent selection is performed based on Markov Logic Networks (MLNs). The hybrid architecture yields a cheap problem formulation in the MLNs w.r.t. inference complexity but pertains their expressiveness. We compare the system to a rule-driven baseline and an extension which uses a memory-based learner. We find that the MLN hybrid outperforms its competitors by large margins.