You understand a topic when you can get questions right several days later, from a blank page, without anything telling you which topic it is, and can say why the method works rather than only what the steps are. Anything short of that is a partial result, and most of the time the gap between the two is invisible from the inside.
This matters because the moment you decide you have finished with a topic determines what you never look at again. Getting that judgement wrong is expensive.
1. The Signals That Feel Convincing and Mean Nothing
It made sense in the lesson. Following someone else’s reasoning is genuinely easier than generating your own. It is a real experience of understanding, and it does not predict what you can do alone.
You finished the exercise. Exercises are built to be finished. The questions ramp gently, they are all on one topic, and question seven looks like question six. Completing the set proves you can continue a pattern that has been set up for you.
Your answers matched. Checking against the back of the book after each question converts a test into a guided walk. Small corrections along the way are invisible to you but are doing a lot of the work.
Nothing feels difficult. Fluency and understanding overlap but are not the same. You can execute a procedure smoothly with no idea what it means.
2. Four Tests That Do Mean Something
The cold test. Wait at least three days, then do two questions with nothing in front of you. No notes, no worked example, no textbook open at the right page. If you cannot start, you have not got it, whatever it felt like on the day.
The unlabelled test. Take questions from five different topics, shuffled, and see whether you can identify what each one requires before solving it. Recognising the type of problem is half the exam and it is the half that single-topic practice never trains.
The why test. Explain to somebody, or to yourself out loud, why the method works. Not the steps, the reason. Why do you need a common denominator before adding fractions. Why does multiplying by a number between 0 and 1 make things smaller. If the explanation is just the steps recited in order, you have the procedure and not the concept, which will hold up until a question is phrased unusually and then will not.
The variation test. Change the question. If you can find the area given the sides, can you find a side given the area. If you can solve 3x + 4 = 19, can you solve 19 = 4 + 3x, or 3x + 4 = 19 where the answer is negative. Understanding survives rearrangement. Pattern matching does not.
3. Grade Yourself in Three Bands, and Be Honest About the Middle One
Secure. Passes the cold test and the variation test. Move on, and put the topic into your rotation to revisit in a few weeks. Why revisiting old maths topics matters explains what a rotation is.
Shaky. You can do it with your notes, or you could do it last week, or you get it right when you know it is coming. This is where almost everything sits, and it is the band people misclassify as secure because shaky feels fine while the topic is current. Shaky topics do not need reteaching. They need a handful of questions every week or so until they stop being shaky.
Not yet. You cannot start without an example in front of you. Go back to the explanation rather than doing more questions, because more attempts at something you cannot do rehearses the wrong thing. If you keep failing the cold test, what to do when you understand the lesson but keep getting questions wrong sorts the mistakes into their kinds.
The bands matter more than they look. A student who marks everything secure revises nothing and is surprised in June. A student who marks everything not yet reteaches material they already have and runs out of time. Getting the middle band right is most of the value.
4. Checking Without Marking Your Own Homework
The problem with self-assessment is that you are judging your own work with the same brain that produced it. A short assessment on a topic you have not touched for a few weeks removes that, and a Module Assessment is designed for exactly this: not to teach the topic, but to tell you which of the three bands you are actually in.
Then treat the result as information rather than a verdict. The point is to find out what to work on next, and a topic that comes back shaky in October is a straightforward problem. The same topic discovered in May is not.