Work on opinion and sentiment tends to focus on explicit expressions of opinions. However, opinions may be expressed implicitly via inference over explicit sentiments and events that positively/negatively affect entities (goodFor/badFor events). We develop an annotation scheme for goodFor/badFor events and for the sentiment of the writer toward their agents and objects. The results of an inter-annotator agreement study of the corpus we annotate are pretty good. Based on the corpus, we investigate how such inferences may be exploited to improve sentiment analysis, given goodFor/badFor event information. We apply Loopy Belief Propagation to propagate sentiments among entities. The graph-based model improves over explicit sentiment classification by 10 points in precision and, in an evaluation of the model itself, we find it has an 89% chance of propagating sentiments correctly.