As a result of increased computing power and software development, the past twenty years has seen a growing trend amongst both academics and industry, to develop models for predicting total engine friction. The potential impact of such work centres around maintaining lubricant development and engine design costs to a minimum. Meanwhile the strongly led legislative demands on lubricant formulators, means that any indication of the performance of a new formulation under specific conditions, is a major advantage prior to commissioning expensive industry standard tests such as the Sequence VIB. This work reviews previous research on valve train friction modelling, with a focus on the approaches taken by the various authors to characterising the lubricant. A typical valve train model is described to provide a framework to evaluate methods of improving sensitivity to lubricant behaviour. Preliminary empirical work is described as a potential mechanism through which greater model refinement may be achieved.