Abstract
Seventy-eight knee joints with varus malalignment were examined pre-operatively using three-point measurement. Operation was performed with or without 5° overcorrection of the varus deformity using random selection. the overcorrection group showed significantly better results than the normal-correction group.