In a recent article, Hu et al. (2019) argued that the commonly used Taylor diagram in climate model evaluation only considered centred statistics, while the absolute error (AE) was not considered. Therefore, they proposed a new index, termed Distance between Indices of Simulation and Observation (DISO), that takes AE, correlation coefficient, and uncentred RMSE into account to comprehensively measure climate model performance. Note that the widely used uncentred RMSE is a function of the AE and correlation coefficient. Therefore, it is not necessary to combine AE and correlation coefficient with the uncentred RMSE again to define DISO. In addition, model performance does not improve monotonically with the decrease in the DISO index, which is also a flaw of the index. Indeed, the Taylor diagram was constructed based on the cosine law between three‐centred statistics that cannot measure the mean error of the climate model. However, this limitation can be mitigated by showing the mean error or percent bias on the diagram or constructing the Taylor diagram with uncentred statistics. The uncentred statistics still satisfy the cosine law and can also be illustrated in the Taylor diagram. In our viewpoint, there is nothing wrong with the Taylor diagram. On the contrary, the Taylor diagram can effectively visualize multiple statistics and compare the performance of different climate models.