In two experiments, children and adults rated pairs of faces from election races. Naïve adults judged a pair on competence; after playing a game, children chose who they would prefer to be captain of their boat. Children's (as well as adults') preferences accurately predicted actual election outcomes.