A system of multiple degrees of freedom composed out of three masses and three springs has been presented in 2008 for analyzing rockfall impacts on protective structures covered by a cushion layer. The model has then been used for a blind prediction of a large-scale test carried out in Sapporo, Japan, in November 2009. The test results showed substantial deviations from the blind predictions, which led to a deeper evaluation of the model input parameters showing a significant influence of the modeling properties for the cushion layer on the overall results. The cushion properties include also assumptions for the loading geometry and the definition of the parameters can be challenging. This paper introduces the test setup and the selected parameters in the proposed model for the blind prediction. After comparison with the test results, adjustments in the input parameters in order to match the test results have been evaluated. Conclusions for the application of the model as well as for further model improvements are drawn.