Severe damage to dwellings and other types of structures can be caused by the impact of a moving, or falling, object in the extreme events of rockfalls, landslides, severe windstorms, and hailstorms. Such damage which is usually localised in nature is controlled by the magnitude of the impact force occurring at the point of contact and is referred herein as the contact force. The magnitude of the contact force can be many times higher than the design impact force (which is essentially the equivalent quasi-static force derived from momentum-energy principles) as per stipulations by codes of practices. Currently, the compressive stiffness properties of the colliding materials controlling the conditions of contact has not been factored into code models.Non-linear visco-elastic behaviour of the contact spring forming part of the two-degree-offreedom (2DOF) analytical lumped-mass model has been proposed in previous investigations for predicting the magnitude of the contact force that is generated by the impact of hail or windborne debris based on idealising objects into spheres. To ensure realistic modelling outcomes, gas-gun experimentations involving machining, or moulding, impactor specimens into spheres would be required in order that contact stiffness parameters can be obtained by calibrating against results from high-velocity impact. This article presents the development, and experimental validation, of an extension of this deterministic modelling methodology for predicting the magnitude of the contact force generated by much larger impactors such as boulders. Impact testing involving real large boulders would be very costly. An important innovation presented in this article is compression testing (by a common test rig) of cylindrical specimens cored from representative boulder, and concrete, samples to determine their compressive stiffness properties. This eliminates the need of any impact experimentations and yet achieves the desired modelling outcomes of accurately predicting contact force generated by the impact of an idealised boulder on a concrete surface. Details of the extensive experimental validation of the proposed deterministic model are reported. Size effects of the impactor on contact force behaviour have also been investigated. A design chart is introduced for predicting the peak amplitude of the contact force for given boulder size and velocity of impact.