How to keep a mistake book for A-levels (and actually stick to it)
It is the most recommended revision technique that almost nobody keeps for more than a fortnight. The problem is not discipline. It is that most advice tells you to write your mistakes down and stops there.
A mistake book — an error log, or 错题本, which is standard practice across Chinese schooling — is a record of every question you got wrong, kept so you can come back to it. Teachers recommend it constantly. The research behind it is solid: you learn far more from retrieving something you got wrong than from re-reading something you got right.
And most students abandon it by week three. Not because they stopped believing in it. Because copying out a question, your wrong answer, the right answer and an explanation takes about ten minutes per mistake, and a bad past paper generates fifteen of them.
So this is written to fix the two things that actually break it: knowing what to write, and keeping the writing short enough to survive contact with a Tuesday evening.
The part almost every guide misses
Not all mistakes are the same kind, and they need opposite fixes. A mistake book that treats them identically produces a list of things you got wrong and no idea what to do about any of them.
There are three kinds, and you can tell them apart in about five seconds.
1. You did not know it
A fact, a formula, a definition you did not have. You could not have answered, because the thing was not in your head.
The unit of magnetic flux density. You wrote "newtons per amp" and it is teslas.
The fix is memorising it. There is no clever version of this. It goes on a card and you drill it.
2. You misunderstood it
You had the facts and applied them to the wrong situation. You used a formula outside the conditions it holds under, or assumed something about the problem that was not true.
Using v = u + at on a question where the acceleration is not constant.
The fix is going back to the idea — the boundaries and conditions of it, not the formula. Drilling this one is a waste of an evening, because you will reproduce the same misunderstanding faster.
3. You knew it, and the marks went elsewhere
This is the biggest category at A-level and the one students most often misdiagnose as stupidity. You understood the physics. The marks went missing somewhere between your head and the page.
- You did not use the marker's words. You wrote "force equals rate of change of momentum" and the scheme wanted resultant force, and proportional to rather than equals.
- You answered a question that was not asked. The command word was "explain" and you described. A stated condition went unused.
- You did not show what the marks were for. The method, the units, the significant figures, the line you did in your head.
- You slipped. A sign, an arithmetic step, a value copied wrong.
The fix is a specific habit, and it is different for each of those four. It is never "be more careful" — that names no habit anyone can act on, and telling yourself to concentrate harder has never once added a mark.
Why the sorting matters more than the writing. From the inside, not knowing something and knowing it but writing it un-examinably feel identical. Both just feel like being wrong. But one means go back to the textbook and the other means you already understand it and need the examiner's vocabulary — and sending yourself back to the textbook when the problem was wording wastes a week you did not have.
What to actually write down
Four lines. Not the question, not your full answer, not a worked solution. Those are already in the paper and the mark scheme, and copying them out is the tedium that kills the habit.
| Line | Example |
|---|---|
| Where | AQA Physics Paper 1 2019, Q4b |
| Which kind | Knew it — wrong words |
| What went missing | Wrote "force", scheme wanted "resultant force". 1 mark. |
| What I do differently | On any Newton's laws question, check every noun against the scheme's noun. |
Twenty seconds a mistake instead of ten minutes. That difference is the whole reason one version survives to May and the other does not.
The fourth line is the one people skip and the only one that changes anything. If you cannot write a specific action, you have not finished diagnosing the mistake. "Revise moments" is not an action. "Write the pivot point down before taking any moment" is.
When to look at it again
A mistake book you never reopen is a diary. The reviewing is the part that works, and the schedule matters because forgetting is not linear — you want to meet something again just as it is about to slip.
The standard schedule, and it is worth following exactly:
- Review a mistake 1 day after you make it.
- Get it right, and the next review is 3 days later.
- Then 7, then 14, then 30.
- Get it wrong at any point and it goes back to the beginning.
- Clear all five and it is genuinely yours. Stop reviewing it.
This is a Leitner system, and the doubling is the point: things you keep getting right stop taking your time, and the schedule concentrates itself on the things you keep getting wrong.
The trap that ruins most people's reviews. Do not re-read the same question and check whether you now know the answer. You will — because you remember writing the correction down, not because you understand it. That feels like progress and measures nothing.
Cover your notes and answer it cold, in full, on paper. Better still, find a different question on the same idea. If you can only do the version you have seen, you have not learnt the idea; you have learnt the question.
The minimum version that works
One notebook. Four lines per mistake. A date in the margin so you know when it is due back.
That is it, and it costs nothing. If you keep it honestly for a month you will find something uncomfortable and useful: the same three or four mistakes account for most of your lost marks, and at least one of them is a habit rather than a gap in your knowledge.
You do not need software for any of this. The only thing software genuinely removes is the tedium — the writing out, and the remembering of when each thing is due back.
Or have it kept for you
That is what we built mistakelog to do. You upload your own past papers and mark schemes, answer questions in the app, and it marks you against the real scheme — then tells you which of the three kinds of mistake you just made. Every one is logged automatically, and comes back on exactly the schedule above, as a new question rather than the one you already know the answer to.
It is free to start, and the method above works perfectly well in a notebook if you would rather.