Predictions of performance are at the core of geotechnical engineering design. Predictions based solely on deterministic analyses suffer from unquantifiable uncertainties and the implied absoluteness of the prediction. On the other hand, probabilistic estimates suffer from being vague, which is unsettling to most geotechnical engineers. Combining deterministic and probabilistic analyses offers synergies that are best utilized only if the geotechnical engineer appreciates the relative role of each type of analysis. This paper describes an overview of the hierarchy of probability-based analyses in geotechnical engineering predictions. The aim is to provide geotechnical engineers with a framework that integrates analytical and probabilistic analyses. The available probabilistic analyses, their level of complexity, applicability and limitations are considered, in order to enable the geotechnical engineer to choose correctly the optimum analysis that best suits their specific project and circumstances.