Example 4: No Man’s Land
Not always is it possible, or sensible, or desirable, indeed necessary, to take account of all the possible/plausible values of a parameter in order to determine the robustness of a decision against variations in the value of this parameter. But, it is imperative to make sure that the robustness analysis in question yields meaningful results. This means that it is one’s duty to ensure that the values of the parameter that are incorporated in the analysis adequately represent the parameter space under consideration.
As might be expected, when it comes to voodoo methodologies, this consideration is not even on the agenda. And to illustrate this point, consider the following picture depicting the No Man’s Land effect created by the robustness analysis prescribed by info-gap decision theory.
Here the large rectangle represents the parameter space and the small yellow rectangle represents the range of values of the parameter of interest that affect the results generated by the robustness analysis. The black area represents the No Man’s Land of the analysis. It is that part of the parameter space that is left out of the robustness analysis and which therefore has no impact whatsoever on the robustness of the decision under consideration.