suppose that n+1 wrongs does not ever fix n wrongs. consider any set of n+1 wrongs W and any additional n+2th wrong w*. Order W by wrongness and pick max w\in W. It is well known there is a threshold of wrongness at which a wrong graduates from a "little" to a "big" wrong. So consider two cases: