The concept of how probabilistic reactor dynamics applies to a realistic problem, an accidental transient of the primary side of a fast reactor, is demonstrated. A full description of the reactor model, including physical variables, evolution laws, and failure rates with their dependence on physical variables, is given. Failure probabilities and failure and success time distributions are evaluated. Vectorized and nonvectorized versions of a Monte Carlo algorithm as well as biased and nonbiased versions of this algorithm are compared.