I don't think there can be a general answer to this question, because it depends entirely on the kind of pre-test and the nature of the mistakes being made. On the one hand, you might not want a silly slip e.g. an arithmetic error to be a reason not to study something higher than arithmetic. At the other extreme, you might not want failing to see a clever application of a technique to be a reason to restudy material that was going to introduce the technique. If your aim is to make sure that the student doesn't skip something such that they end up weaker than the students who don't skip it, the data you need is "what marks do students who have just studied the material get on this test?" and then you want to be towards the upper end of that distribution.

Email: my username, followed by 2, at google's mail