Consider modifying the Partition procedure by (uniform) randomly picking three elements (not necessarily distinct) from array A and partitioning about their median (the middle value of the three elements is used as pivot). What is the probability of obtaining a good split