Spaghetti Criticisms of System Dynamics

[Ian Griffiths <100326.1264@Compu]

In response to Primanata Tjandras comments in SD0270:

It seems to me that good models are accepted in part because they are spare.
The structure is seen to be simple but the insights into dynamic behaviour are
considerable. And model behaviour is close enough in its predictions to
reality.

It is often tempting to add some spaghetti in order to get something like the
behaviour the modeler wants. This is analogous to the method of adding epicycles
in Ptolemaic astronomy. That actually worked well for a very long time. But
(from our perspective) it was complex and failed to generate any new insight.
Success was judged by outcome: did the eclipse happen on the right day? (or
whatever). Similiarly econometricians are pretty exposed to the assertion that,
much of the time, they are really just playing with more epicycles or
spaghetti. Their focus is really on predicting outcomes; not on providing
insights. In some important ways, system dynamics is about reversing the
relative importance of these goals.

So the spaghetti criticism is one that we need to take seriously. It is
insidious -- and its growth must be resisted..

Ian Griffiths
100326.1264@compuserve.com
tel: (44) 01383 872937