Abstract The prediction of properties of coking blends from the properties of the component materials is complicated by interactions that may occur between the components being blended. A method is presented by which such interactions can be identified and accounted for in blends. This method assumes that the magnitude of the effect of an interaction between two coals in a blend of many coals is proportional to the product of the proportion of the two coals in the blend. Consequences of this assumption for predicting blend behaviour are explored. For example, this method predicts that, for a binary blend, for properties that measure the extent to which coals fuse during their pyrolysis, the temperature at which the magnitude of the effect of the interaction is greatest is independent of the proportions of the coals in the blend. If this method is successful, it may simplify blend design.