A novel randomized approach to fixed-order controller design is proposed for discrete-time SISO plants. It is based on the Monte Carlo sampling Schur stable polynomials using so-called Fam--Meditch parametrization and projecting them onto the affine set of closed-loop characteristic polynomials, which is defined by the controller parameters. If the sampling-projecting procedure fails to find a stabilizing controller, certain candidate controllers are then locally optimized by means of an iterative method of nonsmooth optimization.