Ullrich and colleagues construct pedestrian modelling as an explicitly scalar problem. Their iconic operation is the coupling of a network model, which predicts aggregated pedestrian flows at the urban scale, with an agent-based model capable of representing local movement, restrictions and individual behaviour. The theoretical contribution lies in refusing the choice between macro-pattern and micro-action: each becomes meaningful through translation into the other. Methodologically, real-world measurements are used to test the combined workflow, turning validation into a bridge between abstract spatial representation and observed urban behaviour. The model becomes a planning instrument through which alternative design scenarios can be examined before implementation. Its wider significance lies in computational urbanism and evidence-based design, where the critical issue is no longer simply whether a model predicts accurately, but whether different scales of evidence can be connected without losing the spatial conditions that generated them.