The paper deals with effectiveness of particle size separation. The investigations of limestone screening were run on a laboratory scale using a rolling screen. Batch tests were run for various batch outputs as well as different angles of deviation of motovibrators. Quantitative and qualitative effectiveness indices were calculated as well as the imperfection and probable error, which depend on the sieve mesh. A model of determining the screening effectiveness as a function of four independent variables: the amplitude of vertical vibrations of sieve decks (described by means of the angle of deviation of motovibrators), screening time, batch output and the mesh sizes of respective sieves was proposed. The results presented in the paper and the applied approach to stochastic modeling can be used under industrial conditions for rolling screens with variable dynamic parameters at different screening parameters.