For the cases where the design of concrete armour elements for breakwaters is based on deterministic calculations, the load-time response in the contact zone between two colliding elements should be known. Due to crushing, the load-time response may deviate from the linear elastic case considerably. In an extensive laboratory study, load-time histories were determined for concrete to concrete impact. Dynamic experiments were carried out as pendulum tests, in which a concrete striker was swung against a concrete target fixed on top of a prestressed concrete pile. The pile served essentially as measuring device. Variables in the experiments were the contact surface geometry (viz. two different target specimens: flat or corrugated; and three different striker geometries: spherical, conical and truncated conical); the striker mass; the striker velocity at impact; and the concrete quality. In addition to the dynamic experiments a number of static tests was carried out using the same contact surface geometries. In a second series of experiments the effect of size on the impact behaviour was investigated, and the characteristic dimension of the specimens was increased by a factor 2. The pendulum weight was increased by a factor 5 and 10, and amounted up to 2600 kg. An elasto-plastic model was developed, which can be used for calculating the contact zone effect in concrete to concrete collisions with given impulse and geometry. The model consists of three stages: elastic loading (defined by a contact parameter K e,1), a plastic plateau (defined by the size of the contact surface and the critical stress of the material) and a restitution phase (defined by a contact parameter K e,2, which is coupled to K e,1). The loading and restitution phases are modelled using Hertz's law. It is found that all parameters needed in the model can be derived from relatively simple static experiments. The geometry, viz. shape and size dependency of the parameters is discussed in the paper.