This paper applies Bloor's extension of Douglas's grid/group theory to a novel case-namely, a debate which took place concerning the `retrodiction' or `backcasting' of system dynamics global simulation models. The paper has two aims: first, in terms of understanding this debate, it seeks to link the different treatments of anomalies within it both to the strategies for dealing with mathematical counter-examples described by Lakatos, and to the social experiences of the different protagonists; secondly, in terms of methodology, it attempts to illuminate both the possibilities and the difficulties of applying Bloor's scheme to cases outside pure mathematics.