Sometimes I cut up sections of old proofs that haven’t been successful and lay them on top of other failed proofs to see if different combinations might work better.
It certainly beats printing endless versions of the same print with only minor alterations, which then don’t work anyway…