That's kind of a strange approach. If scores rose significantly, then potentially a former 99 percentile could become much lower. That doesn't really make sense as the student was 99 percentile among his peers at the time of testing.