How to use fractional cascading to improve query complexity from O(lognlogn) to O(logn). I could not find any nice blogs on this technique. Only research papers and stuff. I know there are not a lot of problems on this technique since O(lognlogn) is pretty good too :p. For example : Merge sort tree as query complexity of O(logn*logn), which can be lowered to O(logn).