Today, designers are forced to reduce performance and increase power requirements in order to reserve larger margins that are required due to the greater variability introduced by smaller feature sizes and manufacturing variations of mod IC designs. The better-than-worst-case design can both address the variability problem and achieve higher performance/energy efficiency than the worst-case design. This paper surveys the progress to date, provides a snapshot of the most representative methods in this field, and discusses the future research directions of the better-than-worst-case design.