Very common question.
No I don't adjust for attrition for two reasons. First, the prediction rate is around 70% without adjusting. Second, when I ran a sample adjusting for attrition, the changes were extremely slight and rarely accounted for a change in predictions.
Bottom line: most teams suffer from a similar rate of attrition, you just pay more attention to one team than others
