Research by academics and competition agencies on evaluating competition policy has grown rapidly during the last two decades. This paper surveys the literature in order to (i) assess the fitness for purpose of the main quantitative methodologies employed, and (ii) identify the main undeveloped areas and unanswered questions for future research. It suggests that policy evaluation is necessarily an imprecise science and that all existing methodologies have strengths and limitations. The areas where the need is most pressing for further work include: understanding why Article 102 cases are only infrequently evaluated; the need to bring conscious discussion of the counterfactual firmly into the foreground; a wider definition of policy to include success in deterrence and detection. At the heart of the discussion is the impact of selection bias on most aspects of evaluation. These topics are the focus of ongoing work in the CCP.