The approach described above follows the work of Diaz-Balteiro andRomero (2004), where the idea of capturing the worst weighted performance for each alternative is accomplished by computing the minimum(rather than maximum) value among the weighted indicator performances from a given set of positive (rather than negative) indicators.Although both strategies may seem equivalent, a closer look revealsthat under that setting a higher weight can mask rather than highlighta poor performance and therefore their index can fail to identify themajor vulnerability of each alternative, unless equal weights wereused. Thus, the normalization formula used here is crucial in order toobtain an index that effectively represents the intended concept for anarbitrary set of indicator weights.