In contrast to much of the literature in this field, rather than focusing on the nature and aetiology of riots, in this article I investigate why riots do not happen. During times of widespread disorder there are some locations that do not fall prey to rioting themselves despite apparently sharing many of the features of the places where significant disorder is found. Why? Focusing on the 2011 England riots—though the arguments developed here have international application – I use two case studies involving semi-structured interviews with key informants in two such locations to reflect on why riots do not happen. Initially drawing on Waddington’s ‘flashpoints’ model, I argue that it was primarily matters at an interactional level that appear to have been crucial in the absence of riots in these particular cases. In order to facilitate a more detailed analysis at this level in particular, I propose some elaboration of the flashpoints approach, using Reicher’s ‘social identity’ model, as the basis for understanding the nature of police–crowd and police–community interaction.