Partial-hand amputees are often able to use their wrist when performing daily activities, but this wrist movement can interfere with electromyogram (EMG) pattern recognition of functional hand-grasps while controlling myoelectric prostheses. These grasp patterns also commonly require activation of similar muscle sets, resulting in poor discrimination and more frequent misclassification. In our recent work, we developed a classifier training paradigm and control system that improves real-Time control of a virtual prosthesis capable of selecting between 4 grasps in multiple wrist positions. However, it is unclear if there were adverse effects associated with operating the virtual prosthesis in certain wrist positions or with attempting to select specific grasps. The primary purpose of this study is to determine whether the required wrist position or grasp affected task timeout rates, and to determine the number of grasp selection attempts for both a baseline pattern recognition controller and our proposed controller. We show that the specific wrist position of a given task does not significantly affect performance for either the baseline controller (p>0.575) or the proposed controller (p>0.459). However, while the grasp required for a task significantly affects a user's ability to complete the task when using the baseline controller (p0.429). Thus, subjects using the proposed controllers were more easily able to complete tasks involving grasps difficult to select with the baseline controller.