From the ten breakwaters we selected, on the amount of reliable information available, one breakwater in Richards Bay, South Africa. The preliminary check, made in the first phase, showed the design of this breakwater was rather conservative. From surveys done on the breakwater although, it appeared that the armor-layer was damaged in storms with a significant wave height below the design wave height. In the chapters 1 and 3 a description is given of the breakwater in Richards Bay. The traditional design process of armor units on a breakwater consists of the dimensioning of the units with the Hudson-formula together with model tests. The Hudson formula is described in chapter 2, this formula has several short-comings such as: no influence of wave period, spectrum shape and permeability of the structure. Therefore model tests are required to complete the design. This traditional approach in the design process, based on creating a sufficient margin between the load and resistance, takes no notice of uncertainties in the results of wave climate, model testing and construction. Probability methods can be used to account for the distribution of the parameters in the design-formula. The theoretical background of these probabilistic methods is described in chapter 4. Chapter 5 includes a calculation of the probability of exceedance of a certain damage level to the armor layer. This calculation is done on level II, with data provided by the breakwater in Richards Bay. The parameters giving the highest contribution to the probability of exceedance in this calculation are the significant wave height and the damage coefficient. Although the calculation has several short-comings, it shows how this method can be used in the design process. With traditional methods as well as with probabilistic methods, errors may occur when a relevant failure mode is not taken into account. Chapter 6 observes the consequences of not considering breakage in the design process of the armor units used on Richards Bay's breakwater. Although there is little knowledge on the relation between wave action on the breakwater and breakage of armor units, in chapter 8 arelation will be derived from prototype data. Together with this relation a method will be derived to calculate the probability of failure of the breakwater concerning breakage and hydraulic failure. ( chapter 9 )