Ten years of real questions are almost the same, and you can do other years when you have enough time. The problem is not accuracy.

How should everyone do the real questions?

In the improvement stage, we first classify the real questions according to the types of questions, and start with the types of questions to do the real questions. This has two purposes. First, you can know how big the gap between your current level and the exam is. Secondly, when we do different kinds of real questions, we can also know where our knowledge is relatively weak, which is convenient for us to check and fill in the gaps and do special review.

In the sprint model test stage, we also need to take the real question as the fundamental starting point, and everyone needs to continue to do the real question. But at this time, we don't have to classify the real questions, we can do the whole set of real questions directly. At this time, some students may say, I did the real problem in the improvement stage, why do I still do the real problem now? Everyone must understand that the real questions you do and the whole set of real questions are two concepts. We don't need too much thinking span when we do classified real questions. But when we do a whole set of real questions, we need to think across, which is quite harsh on everyone during the exam.

Of course, there are also certain simulation questions to be interspersed. After all, everyone has already done the real problem in the improvement stage. The suggestion for everyone here is to do two sets of real questions and one set of simulation questions.