Consider a set of elements being organized as a binary search tree $($BST$).$ The elements in

the set may be of different popularity. That is$,$ a few elements are searched for much more

frequently than others. In order to improve the average search time, one can consider

moving an element in the set that is being searched for to the top of the BST $($so that if

the same element is searched for again soon, the element is likely close to the top of the

BST and is thus located with fewer node checkings$).$