How to Actually Get Better at Maths: Diagnose, Go Back, Rebuild

Why re-reading your notes is not studying

Re-reading notes and watching worked solutions is the most common way students prepare, and it is close to worthless. Both activities are passive. You are watching someone else do the thing, and the fluency you feel is theirs, not yours. Recognition is not the same as recall. When a worked example is on the screen, every step looks obvious, because the hard part, deciding which step comes next, has already been done for you.

Active retrieval is the opposite. You close the book and produce the method from an empty page. It is uncomfortable, it is slow, and it is the only version that builds anything. The discomfort is the signal that something is being constructed. A student who spends forty minutes copying out worked solutions has revised nothing. A student who spends fifteen minutes attempting three questions unaided, then marks them, has revised properly.

There is a quick test for which one you have been doing. Cover a solved question and write the first line yourself. If you cannot start without looking, you have been reading rather than practising. Almost every student who revised for hours and still failed the paper was doing passive review at high volume. The hours were real. The learning was not.

How do you find the gap that is actually stopping you?

Work backwards from a question you cannot do until you reach the step you cannot justify. Take one question, open the solution, and go down it a line at a time. For each line ask two things: could I have produced this, and can I say why it is allowed. The first line where the answer is no is a starting point, not the answer. Keep digging. The gap usually sits two or three layers below it.

A worked case. A student cannot solve an equation with an unknown in two denominators. The first step of the solution is cross-multiplication, which they can do. The next step expands brackets, and their expansion is wrong about half the time. Tested on expanding brackets alone, they are fine except when a negative sits outside. Tested on multiplying negative numbers, they are guessing. The stuck topic was never equations. It was a rule about signs met in Year 7 and never secured.

This is why drilling the current chapter harder makes things worse. Every extra hour spent on the chapter reinforces the experience of failing at it while the actual cause sits untouched three years back, and the student concludes that the problem is them. At Learning Loft Institute the free demo class is spent on this backwards search rather than on teaching a lesson, because the diagnosis determines everything that follows it.

• Pick one question you cannot do, not ten.
• Read the solution line by line and mark the first line you could not have written.
• Ask why that line is allowed. If you cannot say, it is a candidate.
• Test the candidate on its own with three simple questions.
• If you pass, go one layer deeper. If you fail, you have found it.

Derive the rule instead of memorising it

A memorised rule has one failure mode: you forget a piece of it and have nothing to fall back on. A derived rule can be rebuilt from scratch in about ninety seconds. Under exam pressure that difference decides whether a grade holds. It is also the reason concept teaching is worth the extra time it costs at the start.

Take the quadratic formula, which nearly everyone memorises and almost nobody can explain. Start from a x squared plus b x plus c equals zero. Divide every term by a, then move the constant to the right. Now complete the square: halve the coefficient of x to get b over 2a, and add its square to both sides. The left side is now x plus b over 2a, all squared. The right side tidies into b squared minus 4ac, over 4a squared.

Take the square root of both sides. That is where the plus-or-minus comes from, and where the root of b squared minus 4ac appears, with 2a underneath. Subtract b over 2a and you have the formula exactly as it is printed in the book. Do that derivation three times on paper and you will never again be the student who writes 2a under only half the numerator. You will also see why a negative discriminant gives no real solutions.

The same principle runs through everything. Cross-multiplication is not a rule handed down; it is what happens when you multiply both sides of an equation by both denominators. Dividing by a fraction is not flip-and-multiply; it is asking how many halves fit into three. Every method in the syllabus came from somewhere, and learning where costs twenty extra minutes once.

Twenty focused minutes daily beats three hours on Sunday

Maths is stored by repeated retrieval spread over time, not by total hours logged. Six sessions of twenty minutes across a week will leave more behind than one three-hour session, even though the Sunday session is longer on paper. Every time you pull a method out of memory after a gap, that memory gets harder to lose. Six days of pulling is six strengthening events. One long sitting is one.

The long session has a second problem. Within an hour you are tired, and tired practice teaches you to make careless mistakes fluently. Twenty minutes is short enough that you stay sharp, and short enough that you will actually begin, which matters more than any of the theory. The hardest part of a maths habit is opening the book, and a twenty-minute commitment is easy to open.

Deliberately revisiting old topics is the part students skip. It feels like wasted time because you already did that chapter. It is the opposite. A topic you have not touched for three weeks is precisely the one that has started to fade, so retrieving it now is worth more than four more repetitions of today's work. Build the old-topic warm-up in permanently and never negotiate it away.

• Five minutes: three questions from a topic you last studied two or three weeks ago, from memory.
• Ten minutes: today's topic, book closed, working written out in full.
• Five minutes: mark honestly and write every error into the log.

Keep an error log and sort mistakes into four types

Getting a question wrong tells you almost nothing on its own. Knowing why you got it wrong tells you what to do tomorrow. Keep one notebook for this. For every wrong answer, write the question reference, what you did, what was correct, and which of four categories the mistake belongs to. It takes about a minute per question and it replaces guesswork with a record.

The categories matter because the fixes are completely different. Concept errors need re-teaching from below. Method errors need wider, mixed practice so you learn to recognise which tool a question is calling for, which is a separate skill from using the tool. Arithmetic errors need slowing down and a checking habit, not more theory. Misreads need one small ritual: underline what the question actually asks for before you write anything.

After a fortnight the log tells you the truth about yourself, and it is rarely what you expected. Many students who are certain they are bad at maths find that most of their lost marks are arithmetic slips and misreads, which are mechanical and fixable and have nothing to do with ability. Others find every error clustered inside one topic. That cluster is the gap, now labelled and easy to hand to a tutor.

• Concept: you did not understand what the question was about. Rebuild the topic from below.
• Method: you understood it but chose the wrong procedure, or did the right steps in the wrong order.
• Arithmetic: the reasoning was sound and the numbers went wrong. Usually signs or fractions.
• Misread: you correctly solved a different question from the one printed.

Explain the method out loud to someone who does not know it

The real test of understanding is whether you can teach a method, unprompted, to somebody who does not already know it. Reading a solution and thinking that it makes sense is recognition. Standing at a whiteboard producing that solution while a confused listener keeps asking why is understanding. Most exam marks are lost in the space between those two states.

Do it with anyone: a parent, a younger sibling, a friend in the year below, or an empty chair. The listener does not need to know any maths. They need to ask why you did that at every step and refuse to accept because that is the rule. You will hear yourself go vague at one specific point. That vagueness is the next thing to study, and you have just diagnosed yourself for free.

Explaining is also the fastest way to turn a shaky topic into a secure one. It forces you to sequence the steps and justify each one, which is exactly what a written exam answer demands. Twenty minutes of explaining a topic aloud is worth more than an hour of reading it again. If nobody is available, talk to the wall and write on paper as you go. It feels ridiculous and it still works.

What does a realistic timeline actually look like?

Weeks, not days. If the gap opened two or three years ago it will not close over a weekend, and anyone promising otherwise is selling something. A realistic shape is about a week to locate the gap, two to four weeks to rebuild it, then steady improvement in current work as the foundation stops giving way underneath. Our students improve by an average of two grade bands within six months, which is a fair target and clearly not a fortnight's work.

Expect the marks to lag the effort. For the first two or three weeks the repair is happening below the level being examined, so test scores may not move at all. What changes first is smaller and more reliable. You start knowing where you are stuck instead of feeling generally lost, and questions that used to stay blank now get a first line. Track that, not the percentage, or you will misread real progress as failure.

The usual failure mode is quitting during the lag. A student abandons the method in week three because the last test was no better, goes back to re-reading notes the night before, and confirms to themselves that they cannot do maths. Hold the daily twenty minutes and the error log for six weeks and you will have real evidence either way. Most students never run the experiment long enough to find out.

Keep a record so that the evidence is not just a feeling. At Learning Loft Institute every class is recorded, weekly quizzes are marked within 24 hours and parents receive a written report each month, for exactly this reason: three weeks in, nobody remembers accurately what they could and could not do at the start. Your error log does the same job on your own.