When imagining how an outcome might have turned out differently, people’s “if only” thoughts consistently focus on some kinds of antecedent events rather than others. The current research investigates a tendency to select an antecedent based on similarity or representativeness. We predict a focus on antecedents representatives of the actual outcome or, in other words, unrepresentive of the imagined counterfactual outcome. In Experiments 1 and 2 subjects were put in the role of a manager who had invested in advertising in several media and then experienced a low sales outcome. As predicted, when imagining how sales might have been higher subjects’ “if only” thoughts consistently focused on the advertising media in which the lowest investment had been made. In Experiments 3 and 4 subjects were put in the role of an athlete reflecting upon her finish in a recent triathalong. In Experiment 3, we varied the subject’s imagined “target” outcome (i.e., a better or worse performance) through instructions. Consistently, “if only” throughts focused on the antecedent event that was unrepresentative of the target outcome. Experiment 4 varied the race strategy described as normal for the athlete. Subjects focused on the antecedent that was unrepresentative even when this was presented as part of the athlete’s normal strategy. We discuss how this tendency in counterfactual though differs from previously identified tendencies.