The idea that passivity of alloys such as Fe-Cr is controlled by the richness of interconnection in a disordered system was originally proposed by Sieradzki and Newman. That approach was based on percolation on the metal lattice: Cr-O-Cr bridges form during selective dissolution of Fe. McCafferty has developed another approach using graph theory, in which connectivity within the metal does not play a role: instead, he analyzes the connectivity within a binary oxide of the alloying elements, and cites experimentally determined enrichments of one or other alloying element. McCafferty’s model is really a reactivation model, not a passivation model: it describes the requirement for the film to be stable against dissolution. His graph theory is not equivalent to percolation theory and does not identify the first point where the Cr “sublattice” becomes infinitely connected – rather, it identifies a point or region where the interconnection becomes statistically dense. A similar idea can be framed in the language of percolation.