How to Improve in Mathematics: A Practical, Step-by-Step Method

Almost every student who says "I'm bad at maths" is describing a symptom, not a cause. In practice, a low mathematics grade is nearly always the visible end of a chain: something specific was never fully understood two or three years ago, everything built on top of it has felt shaky ever since, and the gap has widened quietly because mathematics is more strictly cumulative than almost any other school subject.

That is genuinely good news, because it means improvement is a solvable engineering problem rather than a question of talent. What follows is the method we use with students across IB, IGCSE, GCSE, CBSE, ICSE and the American and Australian curricula. It is deliberately unglamorous, and it works at any level from Grade 6 arithmetic to Higher Level calculus.

Why More Effort Alone Often Fails

The most common response to a disappointing maths result is to do more of what already isn't working: reread the chapter, rewatch a video, redo the problems from the worked examples. Students who do this often study for hours and see their grade barely move, which is where the belief "I'm just not a maths person" takes root.

There are two reasons this happens.

The first is misdiagnosis. A student who loses marks on a quadratics question may not have a quadratics problem at all — they may have a fractions problem, or a sign-handling problem, or a problem reading what the question asked for. Studying quadratics harder does nothing for any of those.

The second is passive practice. Reading a solved solution produces a strong feeling of understanding, because following someone else's reasoning is genuinely easy. But exams do not test whether you can follow a solution; they test whether you can generate one from a blank page. Recognition and recall are different skills, and only one of them gets marks.

Step 1: Diagnose Before You Study

Before changing anything about how you study, you need to know what is actually broken. The fastest way to do this is to take your last two or three pieces of assessed work and classify every single lost mark into one of four buckets:

1. I had no idea how to start. A genuine content gap — the method was never learned, or was learned and forgotten.
2. I knew the method but executed it wrong. Usually algebra, signs, fractions, indices or arithmetic. A prerequisite problem, not a topic problem.
3. I answered a different question from the one asked. A comprehension and question-reading problem.
4. I ran out of time or panicked. A fluency and exam-strategy problem.

Do this honestly and the distribution is almost always lopsided. A student who assumed they "needed to redo the whole syllabus" frequently finds that 60–70% of their lost marks sit in bucket 2, which is a far smaller and far more fixable target than an entire course.

Step 2: Rebuild the Prerequisite Chain

Mathematics topics sit in dependency chains. If bucket 2 above was large, the repair work is upstream of whatever topic you are currently being taught. A small number of prerequisites cause the overwhelming majority of execution errors across every curriculum:

• Operations with negative numbers — particularly subtracting a negative, and squaring a negative quantity. Note that −3² = −9 while (−3)² = 9; the brackets are doing real work.
• Fractions — adding with unlike denominators, dividing by a fraction, and simplifying algebraic fractions. This single area silently breaks calculus for a large number of otherwise capable students.
• Index laws — x^a × x^b = x^a+b, (x^a)^b = x^ab, and the fact that x⁻¹ means 1/x, not a negative number.
• Expanding and factorising — reliable expansion of (a + b)(c + d), and recognising a difference of two squares on sight.
• Rearranging an equation — doing the same thing to both sides, with full confidence about what is allowed and what is not.

Two focused weeks on the two or three items from that list that your error analysis flagged will usually do more for a grade than two months of working ahead on new content. Our algebra and subject pages set out how these foundations map onto the topics taught at each stage.

Step 3: Practise the Way Exams Test You

Once the foundations are stable, the mechanism of practice matters more than the volume. Three changes account for most of the difference:

Work from a blank page

Close the book, close the notes, and attempt the problem cold. If you get stuck for more than a few minutes, look at only the next single step of the solution — then cover it again and continue on your own. Consuming a whole worked solution teaches you very little; recovering from a hint teaches you a great deal.

Interleave topics rather than blocking them

Twenty consecutive quadratic equations feel productive, but after the third question you are pattern-matching rather than deciding. Real exams give you unlabelled problems in random order, so practice sets should mix topics. Choosing the right method is itself a skill that only develops when the choice is genuinely open.

Practise under time, at least sometimes

Speed in mathematics is a by-product of fluency, not of rushing. But if you have only ever worked untimed, the exam introduces a variable you have never trained. One timed set per week is usually enough to remove that surprise.

Step 4: Close the Loop on Every Mistake

This is the habit that most reliably separates students who improve from students who plateau. For every question you get wrong, write down two things: the specific step where it went wrong, and the general rule that would have prevented it.

"Made a silly mistake" is not usable. "I distributed the minus sign across only the first term in the bracket" is usable, because it names a recurring behaviour you can watch for. Keep these in one running list — a single notebook page or document — and reread it before every test. Most students discover they have about six recurring errors rather than a hundred different ones, and that list shrinks fast once it is written down.

Then re-attempt the same problem three to four days later, from scratch. Solving a corrected problem immediately proves almost nothing; solving it again after the method has faded is genuine evidence of learning.

A Realistic Timeline

Improvement in mathematics is slow-then-fast, which is exactly the pattern most likely to make students quit early. A reasonable expectation for a student working consistently for around five hours a week:

• Weeks 1–2: diagnosis and prerequisite repair. Grades usually do not move yet, and may even feel worse, because you are working on basics rather than on current classwork.
• Weeks 3–6: execution errors drop noticeably. Questions you could previously start but not finish now get finished. This is where marks begin to appear.
• Weeks 7–12: new content starts landing faster, because the foundations under it are solid. Confidence follows the results, not the other way round.

Anyone promising a transformation in a fortnight is describing marketing, not mathematics. Anyone who tells you it takes years of talent is equally wrong.

Common Mistakes When Trying to Improve

• Starting from wherever the class is. If the prerequisites are missing, current classwork is the wrong entry point.
• Collecting resources instead of using one. Three textbooks, four channels and a question bank usually produce less progress than one good source worked all the way through.
• Only doing problems you can already do. This is comfortable, feels like revision, and builds nothing new.
• Treating the answer as the goal. The method is what gets marks in most curricula, and it is what transfers to the next question.
• Skipping the write-up. Doing mathematics in your head, then copying only the answer, hides exactly the steps where marks are lost.

None of this requires natural aptitude. It requires an accurate diagnosis, repair work in the right place, and practice that resembles the thing you are eventually assessed on. If you would like that diagnosis done properly, a free assessment maps your current gaps against your curriculum and sets out what to fix first — whether you are working within CBSE, IGCSE, IB or another system.