The automatic derivation of informative heuristic functions has been a key development in modern domainindependent planning. Heuristic functions provide the search for plans with a sense of direction that allows large problems to be solved quite effectively. Heuristic search planners, on the other hand, are not transparent: it is not clear why and when they will work, it is not simple to explain, as humans do, why certain actions are selected and others are discarded, and most important of all, it is not simple to improve them in spite of their known limitations. The C3 planner approaches the problem of inference in planning from a different perspective. Rather than relying on the extraction and use of heuristic functions or reductions into SAT or CSPs, C3 prunes’ bad actions’ by appealing to the notion of consistent causal chains (Lipovetzky, Ramirez, and Geffner 2008).