We approach the problem of determining a general method for augmenting haptic simulators to amplify the perceptually salient aspects of the interaction that induce effective skill transfer. Using such a method, we seek to simplify the design of haptic simulators that can improve training effectiveness without requiring expensive improvements in the capability of the rendering hardware. We present a decomposition approach to the automated design of perceptually augmented simulations, and we describe a user-study of the training effectiveness of a search-task simulator designed using our approach vs. an un-augmented simulator. The results indicate that our decomposition approach allows existing psychophysical findings to be leveraged in the design of haptic simulators that effectively impart skill by targeting perceptually significant aspects of the interaction.