Historians have long (0) HELD an uneasy relationship with counterfactual reasoning. To speculate about what might have happened had a single decision gone otherwise is, on one , a sober way of isolating the causal weight of that decision; on another, it risks reducing the vast contingency of events to a game of parlour arithmetic. The of the exercise depends almost entirely on the restraint with which it is conducted. A historian who entertains a single, modest counterfactual — what if the telegram had arrived an hour later? — may genuine insight into the fragility of a crisis. One who the practice into the nineteenth century finds the inferential chain growing with each substitution, until the supposed alternative world no relation to anything a responsible historian could claim to understand. Prudence therefore that counterfactuals be in scope and short in reach, used to illuminate particular junctures rather than to rewrite whole epochs.