There is an emerging strand in the agricultural economics literature which examines the calibration of risk programming models using the principles of Positive Mathematical Programming (PMP). In a recent contribution to this journal, Liu et al. (2020) compare three different PMP approaches and attempt to find the ‘most practical’ method for calibrating risk programming models to be used in policy analysis. In this article, we argue that the comparison design by Liu et al. (2020) is problematic, as it is based on inappropriate metrics and it ignores recent advancements in PMP. This word of caution intends to provide constructive criticism and aims at contributing to the use of risk programming models in policy analysis.