<p>David Lewis’s classical counterfactual theory of causation fails to capture what we call the “reliability constraint on causation,” according to which <i>c</i> causes <i>e</i> only if <i>e</i> would occur in pertinently similar circumstances where <i>c</i> occurs. In this article, we put forward a modified counterfactual account of causation that attempts to accommodate the reliability constraint. Our account adopts a strengthened possible-worlds semantics while retaining the strong centering thesis. Moreover, by contrasting our approach with several alternative counterfactual approaches to reliability, we aim to argue that our account is more plausible than the competing accounts.&#xa0;</p>