In a number of industries, competitive constraints come mostly from outside the market. That is notably the case in most pharmaceutical markets where patents protect the position of an operator. In these markets there is uncertainty for the incumbent of where or when the threat to the market position may come. That begs the question of how this uncertainty should be addressed in a competition law assessment. This article seeks to answer that question by focusing on pay-for-delay cases in pharma markets. Many of the developments could be equally applicable to the other markets. A common thread in pay-for-delay cases is how to objectively prove a restriction of competition based on evidence that reflects the uncertainty at the time of the agreement: should ex post facto evidence or the perception of the parties at the time of the conduct be used? Relying on the pay-for-delay cases Lundbeck and Servier, this article looks at how the concept of potential competition has been applied, then at the value of the transfer that changes the incentive to compete for the generic company, and finally at a test for a restriction by effect analysis.