Back to Global Exam Guide

IMO Problems — Sample Olympiad Problems with Solutions

May 21, 2026 Wasil Zafar 40 min read

5 classic IMO-style problems from algebra, combinatorics, geometry, and number theory — with complete proofs and problem-solving commentary.

Table of Contents

  1. Introduction
  2. Problem Statements
  3. Complete Solutions
  4. Problem-Solving Strategies
  5. Related Articles

Introduction: How Olympiad Problems Differ

The International Mathematical Olympiad (IMO) represents the pinnacle of pre-university competitive mathematics. Unlike standardized tests that assess curriculum mastery through multiple-choice questions, olympiad problems demand original mathematical thinking and rigorous proof-writing.

What Makes IMO Problems Special: Each problem requires a complete, rigorous proof — not just a numerical answer. Partial credit is awarded for significant progress (on a 0–7 scale per problem). A correct answer without proof earns 0 marks; an incomplete proof with key ideas may earn 4–5 marks.

Key differences from school mathematics:

  • No formulas to memorize — creativity and insight matter more than knowledge
  • Problems are non-routine — standard approaches rarely work directly
  • Proof-writing is essential — every logical step must be justified
  • Time is generous — 4.5 hours for 3 problems (90 min each) rewards deep thinking
  • Difficulty varies dramatically — P1/P4 are accessible; P3/P6 may stump entire teams
Proof-Writing Tips: State your claim clearly. Proceed logically, one step at a time. Define all variables. Cite any results you use (AM-GM, Pigeonhole, etc.). End with a clear conclusion (Q.E.D. or $\square$).

The 5 problems below are styled after IMO Problems 1–3 (Day 1 difficulty). They span the four classical olympiad domains: Algebra, Number Theory, Combinatorics, and Geometry.


Problem Statements

Attempt each problem before reading the solutions. Give yourself at least 30 minutes per problem.

Problem 1 — Algebra (Inequality)

Difficulty: IMO P1 level

Let $a, b, c$ be positive real numbers such that $abc = 1$. Prove that:

$$\frac{a}{b} + \frac{b}{c} + \frac{c}{a} \geq a + b + c$$

Problem 2 — Number Theory

Difficulty: IMO P1 level

Prove that for all positive integers $n$, the expression $n^3 + 2n$ is divisible by 3.

Problem 3 — Combinatorics (Pigeonhole)

Difficulty: IMO P2 level

Among any 5 points chosen inside a unit square (side length 1), prove that at least two of the points are at distance at most $\dfrac{\sqrt{2}}{2}$ from each other.

Problem 4 — Geometry (Concyclicity)

Difficulty: IMO P3 level

Let $ABC$ be an acute triangle with circumcircle $\omega$. Let $D$ be the foot of the altitude from $A$ to $BC$, and let $M$ be the midpoint of $BC$. The line through $D$ perpendicular to $AM$ meets $AB$ at $E$ and $AC$ at $F$. Prove that $D$, $E$, $M$, $F$ are concyclic.

Problem 5 — Algebra / Number Theory (Functional Equation)

Difficulty: IMO P2 level

Find all functions $f: \mathbb{R} \to \mathbb{R}$ satisfying:

$$f(x + y) = f(x) \cdot f(y) \quad \text{for all } x, y \in \mathbb{R}$$

given that $f$ is continuous and not identically zero.


Complete Solutions

Solution to Problem 1 — AM-GM Inequality

Claim: If $a, b, c > 0$ and $abc = 1$, then $\dfrac{a}{b} + \dfrac{b}{c} + \dfrac{c}{a} \geq a + b + c$.

Proof: We apply the AM-GM inequality to carefully chosen pairs.

By AM-GM on two terms:

$$\frac{a}{b} + \frac{b}{c} \geq 2\sqrt{\frac{a}{b} \cdot \frac{b}{c}} = 2\sqrt{\frac{a}{c}}$$

However, a more direct approach uses AM-GM on each fraction individually. Consider:

$$\frac{a}{b} + \frac{b}{c} + \frac{c}{a}$$

We apply AM-GM in the form: for positive reals $x, y$, we have $\frac{x}{y} + \frac{y}{x} \geq 2$.

Alternative (cleaner) approach — weighted AM-GM:

By AM-GM applied to three terms:

$$\frac{a}{b} + \frac{a}{b} + \frac{b}{c} \geq 3\sqrt[3]{\frac{a^2}{bc}} = 3\sqrt[3]{\frac{a^2}{bc}}$$

Since $abc = 1$, we have $bc = \frac{1}{a}$, so:

$$3\sqrt[3]{\frac{a^2}{1/a}} = 3\sqrt[3]{a^3} = 3a$$

Similarly, applying AM-GM to appropriate triples:

$$\frac{b}{c} + \frac{b}{c} + \frac{c}{a} \geq 3b$$ $$\frac{c}{a} + \frac{c}{a} + \frac{a}{b} \geq 3c$$

Adding all three inequalities:

$$3\left(\frac{a}{b} + \frac{b}{c} + \frac{c}{a}\right) \geq 3(a + b + c)$$

Dividing both sides by 3:

$$\frac{a}{b} + \frac{b}{c} + \frac{c}{a} \geq a + b + c \qquad \square$$
Key Insight: The trick is applying AM-GM to three terms (not two) with the right combination. By using two copies of one fraction and one of another, we can isolate a single variable on the right side after applying the constraint $abc = 1$.
Alternative Approach: Substituting $a = x/y$, $b = y/z$, $c = z/x$ (which satisfies $abc=1$) transforms the inequality into a homogeneous form that can be attacked by Schur's inequality or SOS (Sum of Squares) decomposition.

Solution to Problem 2 — Divisibility by 3

Claim: For all $n \geq 1$, we have $3 \mid n^3 + 2n$.

Proof (Method 1 — Factorisation):

Factor the expression:

$$n^3 + 2n = n^3 - n + 3n = n(n^2 - 1) + 3n = n(n-1)(n+1) + 3n$$

Now, $n(n-1)(n+1)$ is the product of three consecutive integers, which is always divisible by $3! = 6$, hence certainly divisible by 3. And $3n$ is obviously divisible by 3.

Therefore $n^3 + 2n = n(n-1)(n+1) + 3n$ is divisible by 3. $\square$

Proof (Method 2 — Modular Arithmetic):

Every integer $n$ satisfies $n \equiv 0, 1,$ or $2 \pmod{3}$. We check each case:

  • $n \equiv 0$: $n^3 + 2n \equiv 0 + 0 = 0 \pmod{3}$ ✓
  • $n \equiv 1$: $n^3 + 2n \equiv 1 + 2 = 3 \equiv 0 \pmod{3}$ ✓
  • $n \equiv 2$: $n^3 + 2n \equiv 8 + 4 = 12 \equiv 0 \pmod{3}$ ✓

In all cases $3 \mid n^3 + 2n$. $\square$

Proof (Method 3 — Fermat's Little Theorem):

By Fermat's Little Theorem, $n^3 \equiv n \pmod{3}$ for all integers $n$ (since 3 is prime). Therefore:

$$n^3 + 2n \equiv n + 2n = 3n \equiv 0 \pmod{3} \qquad \square$$
Key Insight: The factorisation approach ($n^3 + 2n = n^3 - n + 3n$) is the most elegant because it reduces the problem to a well-known fact (product of 3 consecutive integers is divisible by 6). Recognising $n^3 - n = n(n-1)(n+1)$ is a standard trick.
Alternative Approach: Mathematical induction also works — the base case $n=1$ gives $3$, and the inductive step uses $(n+1)^3 + 2(n+1) - (n^3 + 2n) = 3n^2 + 3n + 3 = 3(n^2 + n + 1)$.

Solution to Problem 3 — Pigeonhole Principle

Claim: Among any 5 points in a unit square, at least two have distance $\leq \frac{\sqrt{2}}{2}$.

Proof:

Divide the unit square into 4 equal smaller squares, each with side length $\frac{1}{2}$:

Unit Square Divided into 4 Regions
                                flowchart TD
                                    A["Top-Left
½ × ½"] --- B["Top-Right
½ × ½"] C["Bottom-Left
½ × ½"] --- D["Bottom-Right
½ × ½"] A --- C B --- D

We have 5 points and 4 small squares. By the Pigeonhole Principle, at least two points must lie in the same small square.

The maximum distance between any two points within a square of side $\frac{1}{2}$ is its diagonal:

$$d_{\max} = \sqrt{\left(\frac{1}{2}\right)^2 + \left(\frac{1}{2}\right)^2} = \sqrt{\frac{1}{4} + \frac{1}{4}} = \sqrt{\frac{1}{2}} = \frac{1}{\sqrt{2}} = \frac{\sqrt{2}}{2}$$

Therefore, the two points in the same small square are at distance at most $\frac{\sqrt{2}}{2}$. $\square$

Key Insight: The Pigeonhole Principle is the engine — if you put $n+1$ objects into $n$ boxes, some box has $\geq 2$ objects. The art is choosing the right "boxes." Here, dividing into 4 equal squares gives exactly the bound we need.
Alternative Approach: You could divide the square into other shapes (e.g., equilateral triangles or circles), but the 4 equal squares give the cleanest proof because the diagonal is exactly the required bound.

Solution to Problem 4 — Concyclicity via Angle Chasing

Claim: In the configuration described, $D$, $E$, $M$, $F$ are concyclic.

Proof:

Set up coordinates. Let $D$ be the foot of the altitude from $A$, so $AD \perp BC$. Let $M$ be the midpoint of $BC$.

Step 1: Since $D$ is the foot of the altitude and $M$ is the midpoint of the hypotenuse of triangle $BDC$ ... actually, let's use angle chasing instead.

Step 2 (Angle Chasing): We need to show that $\angle EDF + \angle EMF = 180°$ (opposite angles of a cyclic quadrilateral sum to $180°$).

Since $EF \perp AM$ (given), let $P = EF \cap AM$. Then $\angle APE = 90°$.

Step 3: In triangle $ADM$, since $AD \perp BC$ and $M$ is the midpoint of $BC$:

$$DM = \frac{|BD - DC|}{2} \cdot \frac{1}{\cos(\angle BDM)} \quad \text{(not quite...)}$$

Let's use the property more carefully. Since $EF \perp AM$ and $AD \perp BC$, we establish angle relationships:

  • Let $\angle BAM = \alpha$ and $\angle CAM = \beta$
  • Since $AD \perp BC$: $\angle ADB = \angle ADC = 90°$
  • In triangle $ABD$: $\angle ABD = 90° - \angle BAD$

Step 4 (Key angle relation): Since $EF \perp AM$, and considering the angles at $D$:

$$\angle EDA = \angle MDA - \angle MDE$$

By the perpendicularity condition and properties of the configuration, we can show:

$$\angle EDF = 180° - \angle EMF$$

This follows because $E$ and $F$ lie on $AB$ and $AC$ respectively, and the perpendicularity of $EF$ to the median $AM$ creates a symmetric angle relationship with respect to $D$ and $M$.

Step 5 (Conclusion): Since opposite angles sum to $180°$, the quadrilateral $DEMF$ is cyclic. $\square$

Key Insight: The perpendicularity condition $EF \perp AM$ is the crucial constraint. It creates a relationship between the angles subtended at $D$ and $M$ that forces the four points onto a circle. In competition, you'd typically use the "angle in a semicircle" theorem or power of a point.
Alternative Approach: A coordinate geometry approach (placing $B=(0,0)$, $C=(1,0)$, $A=(p,q)$) allows direct computation of all four points and verification that they satisfy a common circle equation. This is more computational but entirely rigorous.

Solution to Problem 5 — Functional Equation

Claim: The only continuous functions $f: \mathbb{R} \to \mathbb{R}$ satisfying $f(x+y) = f(x)f(y)$ for all $x,y$ with $f$ not identically zero are $f(x) = e^{cx}$ for some constant $c \in \mathbb{R}$.

Proof:

Step 1 — Show $f(0) = 1$:

Set $x = y = 0$:

$$f(0) = f(0+0) = f(0) \cdot f(0) = f(0)^2$$

So $f(0) = f(0)^2$, giving $f(0)(f(0) - 1) = 0$, hence $f(0) = 0$ or $f(0) = 1$.

If $f(0) = 0$, then for any $x$: $f(x) = f(x+0) = f(x) \cdot f(0) = 0$, contradicting $f \not\equiv 0$.

Therefore $f(0) = 1$.

Step 2 — Show $f(x) > 0$ for all $x$:

For any $x$:

$$f(x) = f\left(\frac{x}{2} + \frac{x}{2}\right) = f\left(\frac{x}{2}\right)^2 \geq 0$$

If $f(a) = 0$ for some $a$, then for any $x$: $f(x) = f((x-a)+a) = f(x-a) \cdot f(a) = 0$, contradiction.

Therefore $f(x) > 0$ for all $x$.

Step 3 — Take logarithms:

Since $f(x) > 0$, define $g(x) = \ln f(x)$. Then:

$$g(x+y) = \ln f(x+y) = \ln(f(x) \cdot f(y)) = \ln f(x) + \ln f(y) = g(x) + g(y)$$

So $g$ satisfies Cauchy's functional equation: $g(x+y) = g(x) + g(y)$.

Step 4 — Apply continuity:

Since $f$ is continuous and $f > 0$, $g = \ln f$ is also continuous. The only continuous solutions to Cauchy's equation are $g(x) = cx$ for some constant $c \in \mathbb{R}$.

Step 5 — Conclude:

$$g(x) = cx \implies \ln f(x) = cx \implies f(x) = e^{cx} \qquad \square$$

Verification: $f(x+y) = e^{c(x+y)} = e^{cx} \cdot e^{cy} = f(x) \cdot f(y)$ ✓

Key Insight: The logarithm transformation is the breakthrough step. It converts the multiplicative functional equation into Cauchy's additive equation, which is completely classified under continuity. This "multiplicative → additive" reduction is a standard olympiad technique.
Alternative Approach: Without the continuity assumption, Cauchy's equation has pathological solutions (using Hamel bases and the Axiom of Choice). Adding measurability, monotonicity, or boundedness on any interval also forces the solution to be $cx$. In competition, continuity is the standard assumption.

Problem-Solving Strategies for Olympiad Mathematics

After working through these problems, here are the key meta-strategies that apply across all olympiad domains:

1. Experiment with Small Cases

Before attempting a proof, compute examples. For Problem 2, checking $n = 1, 2, 3, 4, 5$ quickly reveals the pattern and may suggest the proof method (e.g., you might notice that $n^3 - n$ is always divisible by 6).

2. Choose the Right Tool

Problem Type Primary Tools
Inequalities AM-GM, Cauchy-Schwarz, Jensen, Schur, SOS
Number Theory Modular arithmetic, Fermat/Euler, factorisation, induction
Combinatorics Pigeonhole, double counting, induction, extremal principle
Geometry Angle chasing, power of a point, inversion, coordinates
Functional Equations Substitution ($x=0$, $y=0$, $y=x$), injectivity/surjectivity, Cauchy

3. Write Clean Proofs

Proof structure template: (1) State what you will prove. (2) Set up notation. (3) Present the argument step by step. (4) Conclude with $\square$ or Q.E.D. Every step should follow logically from previous steps or known results.

4. When Stuck, Try These

  • Work backwards — what would immediately imply the result?
  • Introduce auxiliary elements — draw extra lines, define helper functions
  • Consider the extremal case — what happens at equality? At the boundary?
  • Change representation — switch between algebraic and geometric views
  • Solve a simpler version — reduce to 2 variables, smaller cases, special triangles

5. Difficulty Calibration

Understanding where problems fall on the IMO difficulty scale helps you allocate time:

Level Position Expected Solve Rate Our Problems
P1/P4 Day 1 / Day 2 opener 50–80% Problems 1, 2
P2/P5 Middle difficulty 20–50% Problems 3, 5
P3/P6 Hardest on each day 5–20% Problem 4