If one plate has been subducted, you don’t have pairs of magnetic anomaly boundaries that formed at the same time. But you can reconstruct rotations between anomaly boundaries on the same plate, and use that to extrapolate how much crust was produced in the same time period on the now-subducted plate.