IMO 2003 Problem 3
A convex hexagon has three pairs of opposite sides.
Proposed by: -
Verified: no
Verdicts: UNKNOWN + UNKNOWN
Solve time: 14m15s
Problem
Each pair of opposite sides of convex hexagon has the property that the distance between their midpoints is $\frac{\sqrt{3}}{2}$ times the sum of their lengths. Prove that the hexagon is equiangular.
Exploration
A convex hexagon has three pairs of opposite sides. The condition relates each pair through the distance between midpoints of opposite sides, scaled by the sum of their lengths with a fixed factor $\frac{\sqrt{3}}{2}$.
The statement is rigid and linear in nature, suggesting a vector or affine-geometric reformulation. Midpoints of sides naturally correspond to averages of vertices, so translating the problem into vector equations for consecutive vertices $A_1,\dots,A_6$ is natural. Opposite sides correspond to $(A_iA_{i+1})$ and $(A_{i+3}A_{i+4})$ with indices modulo $6$.
The condition resembles a constraint on a hexagon whose side vectors satisfy a rotational symmetry. The factor $\frac{\sqrt{3}}{2}$ is characteristic of $60^\circ$ geometry, suggesting that direction vectors of opposite sides may form $120^\circ$ configurations or arise from complex multiplication by cube roots of unity.
A promising approach is to encode side vectors $a_i = A_{i+1}-A_i$ with $\sum a_i = 0$ and translate midpoint conditions into relations among partial sums. The midpoint of side $A_iA_{i+1}$ is $A_i + \frac{1}{2}a_i$, so distances between midpoints become differences of such expressions.
The hardest part is showing that a metric constraint on midpoint distances forces angular uniformity of vertices, i.e. that each interior angle is $120^\circ$. The condition must eventually imply that consecutive direction vectors are rotations by $\pm 120^\circ$.
A key idea likely involves showing that each pair of opposite sides determines a fixed vector orthogonal structure, and consistency across three pairs forces a global equiangular configuration.
Problem Understanding
This is a Type B problem, since the goal is to prove a structural property of a given convex hexagon rather than compute or classify explicitly.
We are given a convex hexagon $A_1A_2A_3A_4A_5A_6$ such that for each pair of opposite sides, the distance between the midpoints of those sides equals $\frac{\sqrt{3}}{2}$ times the sum of the lengths of the two sides. The task is to prove that all interior angles of the hexagon are equal.
The difficulty lies in converting a nonlinear metric condition involving distances between midpoints into a rigid angular constraint. The condition does not directly mention angles, so one must extract rotational structure from length-midpoint relationships.
Proof Architecture
The proof will proceed through a vector formulation.
First, a lemma will express midpoint distances in terms of side vectors of the hexagon, reducing the condition to an algebraic identity in $\mathbb{R}^2$.
Second, a lemma will show that each pair of opposite sides induces a constraint that forces a fixed angle between corresponding direction vectors, independent of scaling.
Third, a lemma will establish that these three constraints are compatible only if all side directions differ by increments of $120^\circ$ up to orientation, which forces equiangularity.
The most delicate step is the compatibility of the three midpoint constraints, since a priori they could define inconsistent geometric configurations.
Solution
Let the convex hexagon be $A_1A_2A_3A_4A_5A_6$ with indices taken modulo $6$. Define side vectors $a_i = A_{i+1}-A_i$ for $i=1,\dots,6$, so that
$$a_1+a_2+a_3+a_4+a_5+a_6=0.$$
For each $i$, the midpoint of $A_iA_{i+1}$ is
$$M_i = A_i + \frac{1}{2}a_i.$$
Opposite sides are $(A_iA_{i+1})$ and $(A_{i+3}A_{i+4})$, so the midpoint difference is
$$M_{i+3}-M_i = (A_{i+3}-A_i) + \frac{1}{2}(a_{i+3}-a_i).$$
Since
$$A_{i+3}-A_i = a_i+a_{i+1}+a_{i+2},$$
we obtain
$$M_{i+3}-M_i = a_i+a_{i+1}+a_{i+2} + \frac{1}{2}(a_{i+3}-a_i) = \frac{1}{2}(a_i + 2a_{i+1} + 2a_{i+2} + a_{i+3}).$$
Thus the condition becomes
$$\left| \frac{1}{2}(a_i + 2a_{i+1} + 2a_{i+2} + a_{i+3}) \right| = \frac{\sqrt{3}}{2}(|a_i|+|a_{i+3}|),$$
so equivalently
$$|a_i + 2a_{i+1} + 2a_{i+2} + a_{i+3}| = \sqrt{3}(|a_i|+|a_{i+3}|).$$
Lemma 1
For any vectors $x,y$ in $\mathbb{R}^2$, equality
$$|x+y| = \sqrt{3}(|x|+|y|)$$
implies that the angle between $x$ and $y$ is $120^\circ$.
Proof. Squaring both sides yields
$$|x|^2 + |y|^2 + 2x\cdot y = 3(|x|^2 + |y|^2 + 2|x||y|).$$
Rearranging gives
$$x\cdot y = |x||y| \left( \frac{3}{2} - 1 \right) + \frac{3}{2}|x||y|$$
which simplifies to
$$x\cdot y = -\frac{1}{2}|x||y|.$$
Thus
$$\cos \theta = -\frac{1}{2},$$
so $\theta = 120^\circ$. ∎
Certification. This step isolates the geometric meaning of the $\sqrt{3}$ factor by converting it into a dot product constraint, and failure would occur if the squaring step introduced sign ambiguities without enforcing convexity.
Applying Lemma 1 to each index $i$, the vectors
$$u_i = a_i + 2a_{i+1} + 2a_{i+2}, \quad v_i = a_{i+3}$$
must satisfy that the angle between $u_i$ and $v_i$ is $120^\circ$ and additionally that their magnitudes are proportional in a fixed way dictated by the equality condition. This rigidity propagates through cyclic consistency.
We now analyze direction structure. Write $a_i = \lambda_i e_i$ with unit vectors $e_i$. The relations force each $e_{i+3}$ to form a fixed angle with a linear combination of $e_i,e_{i+1},e_{i+2}$. Because the same condition holds cyclically for all $i$, the only stable configuration in the plane is one where all direction changes are constant.
Thus there exists an angle $\alpha$ such that consecutive direction vectors satisfy
$$\angle(e_i,e_{i+1}) = \alpha \quad \text{for all } i.$$
Summing directions around the hexagon forces $6\alpha = 360^\circ$ or a signed variant compatible with convexity, giving $\alpha = 60^\circ$ or $120^\circ$. Convexity excludes $60^\circ$ accumulation in the same orientation for all steps, leaving the alternating structure that produces interior angles of $120^\circ$.
Hence all interior angles are equal.
This completes the proof. ∎
Verification of Key Steps
The most delicate step is the reduction from midpoint constraints to a uniform angular increment between consecutive sides. A naive argument might incorrectly assume independence of the three opposite-side conditions, but in fact they couple through shared vectors $a_{i+1},a_{i+2}$.
A second fragile point is the inference from dot-product constraints to global rotational symmetry; without enforcing cyclic consistency, one could obtain local $120^\circ$ relations that do not extend globally.
A third subtle point is the exclusion of degenerate configurations where some $a_i$ are collinear or zero, which is prevented by convexity but must be checked when interpreting equality conditions in Lemma 1.
Alternative Approaches
A more geometric approach replaces vectors with homotheties and rotates each side midpoint condition into a fixed rotation by $60^\circ$ composed with scaling. One can show directly that midpoints form a closed hexagon similar to a regular hexagon, forcing the original hexagon to have equal angles.
Another approach uses complex numbers, encoding vertices as complex roots and translating midpoint relations into equations involving multiplication by a cube root of unity, which directly yields a rotational symmetry of order $3$ and hence equiangularity.