All interaction – positive as well as negative – requires opportunity as well as motive. An important part of the former is geographical opportunity, which is usually proxied by some measure of the distance between the interacting parties. Certain well-publicized strands in the globalization literature have revived older notions that the importance of distance is decreasing along with the declining cost of interaction and have announced ‘the death of distance’ and even ‘the end of geography’. What we call the interaction perspective, on the other hand, emphasizes the relative costs of interaction, which are not directly affected by changes in absolute cost. We argue that the importance of relative cost, and thus of distance, should hold up in a globalized world. Our empirical analysis shows that interstate disputes and wars are influenced by distance, and continue to be so in a globalizing world. The analysis also demonstrates the robust and persistent effect of the neighborhood on the risk of conflict. Relatively globalized states are just as affected by a hostile neighborhood and the distance decay as less liberal states. We discuss whether a virtual world will finally relegate the distance-interaction relationship to the scrap-heap of history, but reject this notion. We also reject the idea that the persistent influence of distance on conflict is due merely to territorial disputes.