In a 1981 article, Allan Gibbard and William Harper explained causal decision theory as maximization of the expected utility U of an action A of an action “calculated from probabilities of counterfactuals”:
U(A)=\sum\limits_{j} P(A > O_j) D(O_j),
where D(Oj) is the desirability of outcome Oj and P(A > Oj) is the counterfactual probability that, if A were done, then Oj would hold.
David Lewis proved that the probability of a conditional P(A > Oj) does not always equal the conditional probability P(Oj | A). If that were the case, causal decision theory would be equivalent to evidential decision theory, which uses conditional probabilities.
Wikipedia page on causal decision theory says:
Can somebody explain this strange statement?
An important aspect of a decision theory is how it defines counterfactuals. Anna Salamon wrote a good sequence on this topic.