A method is provided for testing a reliability of a plurality of driven component variants (22, 24, 30), such as those to be used as power train elements in one or more wind turbines. The method includes conducting physical success run testing on a test bench (42) of a first subset of test specimens (50) provided for the component variants (22, 24, 30), and conducting virtual success run testing of a second subset of the test specimens (50). The virtual success run testing does not use test bench time or physical component samples, which reduces the costs typically needed to provide reliability results for specified operating time durations or other parameters at desired confidence levels for operators of driven components such as wind turbine operators, who want to avoid unscheduled downtime based on component failures.