In a very abstract way, the account we have given in the previous chapters of causal Bayesian nets is all that is needed to use them effectively for the applications at hand. But this is true in the same sense that all that there is to classical electrodynamics is contained in Maxwell's four equations: a true statement that ill explains why books on the topic tend to be considerably longer than one half page.

In the same spirit the next three chapters are meant to help the reader use effectively the Bayesian-net technology. Aware of Yogi Berra's wisdom, we intend with these chapters to bridge as much as possible the gap between theory and practice. We do so by teasing out implied results, by providing ‘how-to’ suggestions, and by offering somewhat more advanced techniques for handling the practical problems that invariably arise.

In this part of the book we also present in detail the construction of a Bayesian net in a r ealistic case. The net thus constructed will then be used in the later parts of the book to derive the optimal portfolio allocations and to study the stability of the results.