Why is it erroneous to impute “likelihood\ to Info-Gaps robustness?
Answer-78: As attested by its founder, Info-Gap decision theory was designed as a non-probabilistic, likelihood-free — to which I should add, membership-free — theory. The immediate implication of this is that any talk of “likelihood” in the Info-Gap context runs counter to its conceptual foundation, indeed its justification. In other words, the attribution of “likelihood” to any core element of Info-Gap decision theory, be it the estimate, the robustness result, or whatever, not only introduces a radical change in its foundation, it effectively results in a new theory that is utterly different from the original one. To be specific, the mere fact that to model the uncertainty under consideration, Info-Gap defines the regions of uncertainty U(α,û),α≥ 0, does not in any way shape or form make us the wiser as to what values of u are more/less likely than others. To repeat, defining U(α,û),α≥ 0, as such, gives no indication whatsoever what values of u are more/less likely than others. In
