[en] Determining how living beings react to tasks that reflect realistic situations of risk has given rise to a vast literature. However, I argue that the methodologies traditionally used to test humans and nonhumans relative to risk often fail to achieve their goal. When risk is modelled in laboratory, potential decision cost (or potential loss) typically denotes an absence of optimal gain. In contrast, when risk occurs in real-life situations, potential loss denotes the reduction in an individual’s limited resources – whether energetic, social, financial, etc. This conceptual difference about the nature of risk may have important implications for the understanding of the parameters that control risk-taking behaviour.