This paper investigates the effects of an unprecedented education policy called ‘double reduction’ (DR) on students’ performance, with evidence from China. One of the significant changes brought about by this policy is the transition of examination scores from a numerical system to a grading system. We found a junior middle school in a small town in Guangzhou that met the requirements for a natural experiment, and we used such unique data to examine three hypotheses. After the implementation of DR, the ‘rat race’ reflected on the students’ grade is reduced. To examine heterogeneity, we apply quantile regression showing that the high-ability students exert less efforts on the ‘rat race’ than low-ability students. Meanwhile, we also provide possible explanations in terms of this phenomenon. Overall, this study not only does it assess the grading transformation that DR policy induced but also fills the gap in the field of DR from an economic perspective.