I search is:due -is:learn -prop:due=0 then flag and suspend everything. To get something like RRO, search, unsuspend is:suspended flag:flag_number prop:r>.9 slowly going through high R cards.
If the goal is to get the inverse Relative overdueness. Then you can assemble the cards into a filtered deck with Relative overdueness selection. When the cards are needed, you should simply reduce the limit.