Rail operations are housed inside a complex and extremely dynamic system where work is distributed in time and space. The train driver has traditionally relied on their own decisions, plans, and actions to navigate the rail environment, but the use of modern driver systems that force how these activities are regulated has altered this dynamic. This paper reports the findings of a study that set out to investigate the skills of modern (enhanced display-based) and traditional (real world) train driving. Data were collected from a variety of UK domain experts (n = 45) using an innovative methodology that converged multiple techniques for knowledge elicitation and analysis. The findings are represented in a model of dynamic train control and discussed according to the specific features and nature of tracking skill in the rail domain. The utility of the model is demonstrated through work of its application to the design of a train simulator and research tool for systematic study of rail human factor issues.