Necessity and sufficiency are well-established notions in logic and causality analysis, but have barely received attention in the formal methods community. In this paper, we present temporal logic characterizations of necessary and sufficient causes in terms of state sets in operational system models. We introduce degrees of necessity and sufficiency as quality measures for sufficient and necessary causes, respectively, along with a versatile weight-based approach to find “good causes”. The resulting optimization problems of finding optimal causes are shown to be solvable in polynomial time.