Understanding how antioxidants work in lipid-based systems is the first necessary step to optimize current antioxidant strategies and develop new ones. To do so, we need a model that can predict the activity of various antioxidant molecules in different systems. Hitherto, the polar paradox was the only available model. It predicts that polar antioxidants would generally tend to be more active in bulk oils than their nonpolar counterparts, while the reverse phenomenon is thought to occur in lipid dispersions such as oil-in-water emulsions, micelles, membranes, etc. This chapter questions the relevancy of the polar paradox hypothesis and suggests other models, taking into account challenging results to the polar paradox that have been accumulated in the past decade. General principles are given here to try to answer the difficult question of knowing what makes a good antioxidant in bulk oils, lipid dispersions, and cultured cells.