The original paper investigates the concept of an ``ideal'' overall stoichiometry of a given set of steps, starting materials, and target product, which is not irrelevant to a theory of multistep reactions. Such a theory should be able to determine the best yield obtainable from a given mechanism.
Ross states: ``In the study of chemical kinetics ... the stoichiometries ... are almost always known.'' My paper does not concern chemical kinetics, it concerns chemical reactions. Reactions can be studied, and their pathways or mechanisms talked about rationally and usefully, without knowing the overall stoichiometry. This occurs constantly in the study of complex, multistep reactions most often carried out by chemical engineers. If the final products or the reduction in starting materials cannot be measured within the discretionary effort put forth by the experimenter, then the reaction stoichiometry remains unknown.
Stoichiometry is surely a powerful constraint on possible mechanisms. But it is only one among many pieces of constraining evidence. In fact, it is less powerful than is often realized, since a set of reactions could generate the observed stoichiometry by weighted summation, rather than by simple summation, contrary to the suggestion in Ross's last paragraph.
Even if Ross's assertion about prior knowledge of stoichiometry were conceded, it would nonetheless be interesting to compare an observed stoichiometry with the theoretical optimum allowed by the mechanism, especially since yield can vary with reaction conditions. It helps to have defined beforehand the concept of optimal stoichiometry, as well as how to compute it.
In brief, my point is this: sets of reactions exist, independently of whether we know their overall stoichiometry or not. Within the space of stoichiometries spanned by a set, one can identify an interesting special case: the optimal stoichiometry(ies) resulting in the best yield of a target product. The original paper develops the details.