IB Maths Matrices (AI HL): Common Questions Answered

Most marks vanish through arithmetic slips in row operations, sign errors when finding a determinant, or forgetting that matrix multiplication isn't commutative. A huge chunk also lose marks by solving with a GDC but not showing the method the mark scheme wants — matrices questions carry method marks, not just answer marks.

For most students it's using matrices to solve systems of linear equations, especially when a system has infinitely many solutions or none — you need to interpret a zero determinant correctly, not just calculate it. Eigenvalues and eigenvectors, if your teacher covers them as extension, trip up nearly everyone at first too.

For 2×2 matrices, yes — know the formula $A^{-1}=\frac{1}{\det A}\left(\begin{array}{cc}d& -b\\ -c& a\end{array}\right)$ cold, since Paper 1 (non-calculator-style reasoning, though AI HL papers all allow GDC) often wants working shown. For 3×3 and larger, your GDC does it — but you must show the matrix you entered and the inverse it returned.

Matrices appear only in Applications and Interpretation (AI) HL — not in AI SL, and not in Analysis and Approaches (AA) at either level. According to the IB, the current Mathematics: Applications and Interpretation guide (first exams 2025) places matrices in AI HL Topic 1, alongside eigenvalues and eigenvectors as core content.

The AI HL syllabus covers matrix addition, subtraction and multiplication, determinants and inverses (2×2 and 3×3), solving systems of linear equations using matrices, and eigenvalues/eigenvectors including diagonalisation. Matrices also link to the Markov chains sub-topic, where a transition matrix models probability moving between states over time.

Matrices can appear on any of the three AI HL papers, since all three allow a GDC. Paper 1 and Paper 2 tend to test direct calculation — solving systems, finding inverses, Markov chain states. Paper 3, the extended-response paper, often embeds matrices inside a longer modelling context worth 20+ marks.

A Markov chain models a system moving between states with fixed transition probabilities, and a transition matrix stores those probabilities. Multiplying the matrix by a state vector repeatedly (or raising it to a power) predicts long-run behaviour — including the steady-state, where the distribution stops changing.

Start with the arithmetic — addition, multiplication, determinants — done by hand until it's automatic, then move to GDC methods for larger matrices. Work through past paper questions by topic rather than by year, since matrices questions repeat similar structures every session: solving a system, then interpreting the answer in context.

You need to enter matrices into your calculator's matrix editor, compute determinants and inverses, multiply matrices of compatible dimensions, and raise a matrix to a power for Markov chain problems. Examiners expect you to state the matrix you used, not just the final numeric answer — a bare decimal with no matrix shown often loses a method mark.

Eigenvalues and eigenvectors are conceptually the trickiest matrix sub-topic in AI HL, mainly because students memorise the characteristic equation $\det (A-\lambda I)=0$ without understanding what an eigenvector actually represents. Once you see it as 'a direction the transformation doesn't rotate, only scales', the rest — diagonalisation, matrix powers — follows more naturally.

Struggling with matrices alone isn't a reason to avoid AI HL — it's one topic among many, worth roughly one Paper 3 question and scattered marks elsewhere. The bigger question is whether your child prefers applied, technology-driven maths (AI) over the more abstract, proof-based style of AA — that preference matters more than any single topic.

Matrices typically contribute one substantial question per paper session — often a multi-part Paper 3 question worth 15 to 25 marks when combined with Markov chains, plus smaller standalone questions on Paper 1 or 2. It's not the highest-weighted topic overall, but skipping it entirely risks losing an easy, formulaic set of marks.