The use of deterministic stability analyses for geotechnical works is still common practice. Such analyses consider a single set of input parameters and therefore a single result is taken as definitive and compared to limits established by codes. However, a deterministic stability assessment taken as satisfactory may be associated with a probability of failure considered as high. This is why one approach should not suppress the other, they should rather be complementary. A similar comment can be made for the probability of failure, while a small value may not be accepted within a densely occupied urban scenario and a high value may be considered satisfactory within an uninhabited area. This is why the risk should also be evaluated rather than solely the probability of failure. Discussions concerning the risk of failure instead of mere deterministic approaches have significant importance, bearing in mind either insurance needs or the development of projects that are both more reliable and cost-effective.