One of the things I like about traditional RPGs, and older editions of D&D in particular, is that they have what David Brooks calls "epistemological modesty". That is, the designers seem to have understood the futility, indeed the foolhardiness, of attempting to come up with a coherent and robust system from the outset that will work for all players; their philosophy rather seems to have been to provide a starting point, a toolbox, with which DMs and players can make what they will, and which have almost been conceived with the expectation of evolution and change. The lack of cogency in these games, and the very necessity of house ruling them, is seen as a flaw by some, but it's all a matter of perspective; I suspect Karl Popper, William Morris and Michael Oakeshott would all have approved of OD&D's approach.
In fact the history of RPGs could in many ways be seen as a struggle between epistemological modesty and reductionism. On the one hand you have those who believe in decentralisation and freedom of interpretation. On the other are those who think that the complexity inherent in a game played by literally tens of thousands of different discrete groups at any one time could be simplified to the point where a satisfactory unified mechanical structure could be maintained. The latter seems like tilting at windmills to me, but it's been the philosophy of game design since at least the mid-90s.