The determination of material strength parameters from structural test measurements yields in general an inverse problem. The measured test values are e.g. loads and displacements applied on the specimen. The unknown material parameters are e.g. local stress values at particular strain states. Naturally, material models should only contain material parameters which can be identified by such back calculation. However, if material models want to be practically useful, the back calculation should be a simple procedure, which is derived from adapted analytical and numerical models, and where the material parameters are gained in a most direct way. In a complete approach the introduction of a new constitutive model is accompanied with the specification of an according test setup and a validated evaluation procedure, with which the engineer can determine the occurring material parameters. The current article discusses this issue on hand of two simple, but nonstandard material strength data for metals: First, the flow curve under pure shear, which is determined by using a plane torsion test rig, and second, the compressive strength under high strain rates, which is determined by using a cam drive test rig. Both, analytical considerations and numerical Finite Element simulations are used for back calculation.