ELASTOCOAST revetments are highly porous structures made of crushed stones which are durably and elastically bonded by Polyurethane (PU). To improve the understanding of the physical processes involved in the wave-structure-foundation interaction and to develop prediction formulae for both hydraulic performance and wave loading more than 75 large-scale model tests using both regular and irregular waves were performed. Three ELASTOCOAST revetment alternatives with the same slope (1:3) and the same revetment thickness (0.20 m) but with different thicknesses of the underlying filter layer (0.00 m, 0.10 m and 0.20 m for Model Alternatives A, B and C, respectively) were tested. More than 85 measuring devices synchronously connected to two video cameras were used. Prediction formulae are developed for wave reflection, wave run-up and run-down as a function of the surf similarity parameter which illustrate the advantage of ELASTOCOAST revetments as compared to conventional revetments. For instance, more than 25% less wave runup may result on comparison to smooth impermeable revetments. Using a surf similarity-based wave load classification as well as a systematic parametrization in both time and space, prediction formulae are also developed for both impact loads on and just beneath the revetment. These include the peak pressure pmax, its location in relation to still water level zpmax, the spatial pressure distribution and the time related parameters (rise time and total load duration). Prediction formulae for the wave-induced pore pressure in the sand core beneath the revetment are also provided, including the maximum pressure at the upper boundary of the sand layer and its development in deeper layers. Formulae are also proposed for the flexural displacement ä of the ELASTOCOAST revetment, showing that for impact load much smaller displacements would result than for non-impact load and that ä linearly increases with peak pressure pmax for a given revetment thickness. Finally, a stability analysis of Model Alternative A is performed on basis of the results of the measurements and the simultaneously recorded videos. The results illustrate why Model Alternative failed due to local transient soil liquefaction while Model Alternative B tested synchronously under the same wave conditions did not fail.