вот этот кусок у абрамского меня заинтересовал недавно
The key feature of games, by comparison with the many extant models of computation (labelled transition systems, event structures, etc. etc.) is that they provide an explicit representation of the environment, and hence model interaction in an intrinsic fashion. (By contrast, interaction is modelled in, say, labelled transition systems using some additional structure, typically a “synchronization algebra” on the labels.)