In two experiments, one hundred and sixty‐two 6‐ to 8‐year‐olds were asked to reason counterfactually about events with different causal structures. All events involved overdetermined outcomes in which two different causal events led to the same outcome. In Experiment 1, children heard stories with either an ambiguous causal relation between events or causally unrelated events. Children in the causally unrelated version performed better than chance and better than those in the ambiguous condition. In Experiment 2, children heard stories in which antecedent events were causally connected or causally disconnected. Eight‐year‐olds performed above chance in both conditions, whereas 6‐year‐olds performed above chance only in the connected condition. This work provides the first evidence that children can reason counterfactually in causally overdetermined contexts by age 8.