The reactivity effects of material rearrangements, simulating conditions in a postulated liquid-metal fast breeder reactor accident, were measured in three different critical assemblies. SNEAK-12A, a single-zone core, fueled with enriched uranium metal plates; SNEAK-12B, which had a central test zone fueled with Pu0202 rod bundles surrounded by a buffer and a driver zone; and SNEAK-12C, which had nearly the same integral compositions as SNEAK-12B but was loaded totally with plates. The reactivity effects were calculated using current Kernforschungszentrum Karlsruhe methods and data and, in part, also using the corresponding modules of the SIMMER-11 accident analysis system. Also, for some configurations, a comparison of measured and calculated fission rate distributions was performed., The evaluation yielded similar results for the three assemblies. For most cases investigated, satisfactory agreement between theory and experiment was reached when two-dimensional transport eigenvalue calculations or exact transport perturbation methods were used. As long as larger deviations occurred, transport results generally were on the conservative side. First-order transport perturbation theory only worked well in a limited number of cases. Diffusion calculations often led to large discrepancies, particularly when the experiments involved fuel dilution.