A maximum likelihood framework for probabilistic assessment of post cyclic deformations is described. A database, consisting of cyclic laboratory test results including maximum shear strains in conjunction with relative density, number of stress (strain) cycles and "index" test results is used for the development of probabilistically-based post cyclic deformation correlations. For this purpose series of stress controlled cyclic triaxial tests were performed on laboratory constituted clean sand specimens consolidated isotropically to a confining pressure of 100 kPa. The proposed stochastic method allows full and consistent representation of uncertainties, including model imperfection and statistical uncertainty. Volumetric and shear strain boundary curves are developed for the assessment of post cyclic deformation problem. The resulting correlations represent a significant improvement over prior efforts, producing predictive relationships with enhanced accuracy and greatly reduced overall model uncertainty.