Current guidelines support the well-established clinical practice that patients who present with atrial fibrillation (AF) of less than 48 hours duration should be considered for cardioversion, even in the absence of pre-existing anticoagulation. However, with increasing evidence that short runs of AF confer significant risk of stroke, on what evidence is this 48-hour rule based and is it time to adopt a new approach? We review existing evidence and suggest a novel approach to risk stratification in this setting. Overall, the risk of thromboembolism associated with acute cardioversion of patients with AF that is estimated to be of <48 hours duration is low. However, this risk varies widely depending on patient characteristics. From existing evidence, we show that using the CHADS-VASc score may allow better selection of appropriate patients in order to prevent exposing specific patient groups to an unacceptably high risk of a potentially devastating complication.