This paper evaluates the performance of currently available analytical procedures to assess building response to tunnelling-induced ground displacements. The focus is on methods that account for the interaction between the soil and the structure during tunnelling. These methods relate the soil to the building stiffness and are often called Relative Stiffness Methods (RSMs). Results from centrifuge model tests are used to evaluate the ability of these RSMs to predict building deformations. This evaluation benefits from detailed building models including facade openings, intermediate walls and strip footings. The range of RSM predictions was large, and the accuracy of each RSM was quantified. It was found that no RSM accurately predicts flexural building deformations. Recommendations that consider the building-to-tunnel position to achieve accurate predictions are indicated. This contribution provides a better understanding of the performance of currently available criteria to assess the risk of urban tunnelling.