I came up with basic floops specifically becuase causal diagrams were too large and great for mapping things, but poor for understanding potential control points. It sounds weird, but a lot of causal loop diagrams explode to the point where they may not be overwhelming. http://www.stat.columbia.edu/~cook/movabletype/archives/2009/12/how_to_win_in_a.html

The number of loops and interactions by definition limits their effectiveness due to the rise in response uncertainty. Floops are designed for simple feedback systems that in should have a high enough degree of certainty that a control action could be taken with an expected result. The idea being that a single floop should be abstracted or focused relative to the scope in which a control variable will have a somewhat expected response.

large systems diagrams pose the challenge of “so what next” in that the information while all encompassing may not actionable. The diagram captures state variables and relationships but may not offer immediate representation of control response.

just a thought. One doesn’t rule out the other and perhaps a large system diagram map is used for plotting all relationships and then a simple Floop is used to highlight a “tight” cuasal feedback component. the floop could be examined and tweeked and then looked at relative to the causal diagram. Of course most of the feedback/response mechanisms in such diagrams are either a-temporal, response amplified or attenuated or non-linear response functions.

I guess my fear causal diagrams, is that the completeness of the diagram masks the true complexity of the system. In a way the causal diagram may give a false sense of control. A floop may be a bit more humble.

Nick