This abstract develops the argument that Policy and Design provide the (only) in-vivo laboratory for the science of complex systems. Scienitists do not have the mandate or the money to conduct experiments on large complex socio-techical systems. They must align themselves to policy makers. This leads to design. Prediction has to be seen in a new way in this context. There is scope and need for new scientific methodlogy including new statistical methods.

I would like to develop the argument into a long paper, adding many details.