Many scientists in China share Nai-Xing Wang's dissatisfaction with the dominant role of journal impact factors in the country's scientific evaluation system (Nature 476, 253; 2011). But I contend that even an imperfect law is better than no law.

Replacing this rigid evaluation system with a more flexible one could send Chinese academia into chaos. Leaders of universities and research institutions could then establish their own evaluation systems, designing them to favour their particular interests. For example, a professor who is connected to a scientific journal might be tempted to rank papers published in that journal more highly when evaluating the performance of his or her university.

Chinese researchers should benefit from the strict implementation of impact-factor evaluation criteria. But the rewards for meeting these targets aren't always forthcoming. A good relationship with the few leading executives who control China's academia is also important, as it is for gaining access to the best scientific projects and for promotions.

The key task is therefore to eradicate this autocratic control. Researchers would then be able to concentrate solely on their work.