A method is presented for the fuel element designer to relate the reliability required of the fuel element to its design. Random and systematic uncertainties are used to determine the fraction of fuel rods that can exceed some limit and to determine the probability that the fraction exceeding the limit is less than that allowed. The method is used with analytical models of fuel and cladding behavior. The method requires that the designer not only know the values for the variables in his analytical model, he must also know the uncertainties in these variables. When using this technique, the fuel element designer can see which of the various uncertainties are contributing the most to the uncertainty in the margin. Those uncertainties that contribute the most are those that merit additional expenditure for research and development or additional quality control effort.