This would be useful for inclusion in trial protocols, or to help choose the number of looks during planning. The probability of early stopping at each look would be tied a selected design and already calibrated thresholds so would be an add-on function like calc_decision_rules().