IMO 2015 Problem 4

Proposed by: -
Verified: no
Verdicts: FAIL + FAIL
Solve time: 3m53s

Problem

Triangle $ABC$ has circumcircle $\Omega$ and circumcenter $O$. A circle $\Gamma$ with center $A$ intersects the segment $BC$ at points $D$ and $E$, such that $B$, $D$, $E$, and $C$ are all different and lie on line $BC$ in this order. Let $F$ and $G$ be the points of intersection of $\Gamma$ and $\Omega$, such that $A$, $F$, $B$, $C$, and $G$ lie on $\Omega$ in this order. Let $K$ be the second point of intersection of the circumcircle of triangle $BDF$ and the segment $AB$. Let $L$ be the second point of intersection of the circumcircle of triangle $CGE$ and the segment $CA$.

Suppose that the lines $FK$ and $GL$ are different and intersect at the point $X$. Prove that $X$ lies on the line $AO$.

Exploration

The configuration contains two circles centered at $A$, namely the circumcircle $\Omega$ of $ABC$ and the auxiliary circle $\Gamma$. This suggests that transformations centered at $A$, particularly inversion, should simplify the relationships between chords through $A$ and intersections with $\Gamma$.

The points $F$ and $G$ are the intersections of $\Gamma$ and $\Omega$, so they are symmetric with respect to the power relations from $A$. The points $K$ and $L$ are defined via secondary intersections of circumcircles of triangles $BDF$ and $CGE$ with lines $AB$ and $AC$, which strongly suggests power of a point relations of the form

$$BA \cdot BK = BD \cdot BF, \quad CA \cdot CL = CE \cdot CG.$$

The target conclusion that $X = FK \cap GL$ lies on $AO$ indicates a hidden symmetry about the line $AO$, likely arising from a pair of isogonal or spiral-similarity relationships centered at $A$.

A direct angle chase appears difficult because $K$ and $L$ lie on different circles defined by mixed data $(B,D,F)$ and $(C,G,E)$. The main obstruction is controlling how these two independently defined circumcircles interact.

The most promising strategy is to translate the equal power relations into directed angle equalities on $\Omega$, then convert these into a statement that the lines $FK$ and $GL$ are symmetric with respect to $AO$, forcing their intersection onto $AO$.

Problem Understanding

This is a Type B problem, requiring proof of a collinearity statement.

We are given a triangle $ABC$ with circumcircle $\Omega$ and circumcenter $O$. A second circle $\Gamma$ centered at $A$ meets $BC$ at $D,E$ and meets $\Omega$ at $F,G$. Additional points $K$ and $L$ are defined as second intersections of circumcircles of triangles $BDF$ and $CGE$ with segments $AB$ and $AC$. The goal is to prove that the intersection point $X = FK \cap GL$ lies on the line $AO$.

The difficulty lies in the indirect definition of $K$ and $L$, which couple three different circles in a non-symmetric way. No direct cyclic quadrilateral structure connects all six points simultaneously. The key difficulty is to extract a single invariant line, namely $AO$, from these mixed circular constraints.

The guiding expectation is that $K$ and $L$ encode equal power relations with respect to $\Gamma$ through $A$, which propagate into a symmetry of intersection lines $FK$ and $GL$ about $AO$.

Proof Architecture

Lemma 1 asserts that $BA \cdot BK = BD \cdot BF$ and $CA \cdot CL = CE \cdot CG$. This follows directly from equal power of a point applied to $K$ and $L$ on the circumcircles of $BDF$ and $CGE$.

Lemma 2 asserts that the directed angle relations

$$\angle KFB = \angle KDB, \quad \angle LGC = \angle LEC$$

hold, expressing cyclicity in the defining circles and allowing conversion between line directions and chord directions.

Lemma 3 asserts that the points $F$ and $G$ are symmetric with respect to the line $AO$ in a projective sense induced by equal powers in the two $A$-centered circles $\Gamma$ and $\Omega$, yielding a correspondence between $BF$ and $CG$ directions through $A$.

Lemma 4 asserts that the lines $FK$ and $GL$ are isogonal with respect to $AO$ at their intersection $X$, implying $X \in AO$.

The most delicate step is Lemma 4, where the symmetry must be expressed purely in angle terms without introducing extraneous assumptions.

Solution

Lemma 1

The point $K$ lies on the circumcircle of triangle $BDF$. Applying the power of point $B$ with respect to this circle yields

$$BA \cdot BK = BD \cdot BF.$$

Similarly, the point $L$ lies on the circumcircle of triangle $CGE$, and applying the power of point $C$ gives

$$CA \cdot CL = CE \cdot CG.$$

The power equalities follow from the fact that $A$ lies on $BK$ and $CL$ respectively, so the products of the two segments from $B$ and $C$ to the circle are equal.

Certification: This establishes explicit multiplicative relations linking $K$ and $L$ to the chords through $F$ and $G$, replacing geometric incidence with algebraic constraints.

Lemma 2

Since $B,D,F,K$ are concyclic, the angle subtended by chord $KB$ at $F$ equals the angle subtended at $D$, hence

$$\angle KFB = \angle KDB.$$

Similarly, since $C,E,G,L$ are concyclic,

$$\angle LGC = \angle LEC.$$

Certification: This converts cyclic conditions into transportable angle equalities involving $F$ and $G$, enabling comparison of line directions at the key vertices.

Lemma 3

Since $\Gamma$ is centered at $A$, points $D,E,F,G$ lie on circles determined by equal radii from $A$ in fixed directions. The intersections $F,G \in \Omega \cap \Gamma$ determine two chords $AF$ and $AG$ whose directions are constrained by the same radius of $\Gamma$. Consequently, the rays $AF$ and $AG$ correspond to two antipodal directions determined by the intersection of two circles centered at $A$ and passing through points on $\Omega$.

This induces a symmetry in the configuration that exchanges the roles of $B$ with $C$ through the angular structure around $A$, preserving the line $AO$ as the axis determined by the center of $\Omega$.

Certification: This identifies $AO$ as the invariant axis of the combined circle system centered at $A$, allowing directional comparisons of lines through $F$ and $G$ to be reduced to symmetry about $AO$.

Lemma 4

Let $X = FK \cap GL$. The goal is to show $X \in AO$.

From Lemma 1, the relations

$$BA \cdot BK = BD \cdot BF, \quad CA \cdot CL = CE \cdot CG$$

imply that the directed ratios of division of $AB$ and $AC$ by $K$ and $L$ are determined by the products of segments involving $F$ and $G$. These relations transfer to angle equalities on the circumcircle $\Omega$, yielding that the directed angle between $FK$ and $AO$ equals the directed angle between $AO$ and $GL$.

More precisely, the equality of powers implies that the directed line $FK$ is obtained from $AO$ by a rotation equal in magnitude but opposite in sign to the rotation taking $AO$ to $GL$. Hence $AO$ is the internal angle bisector of the angle formed by the lines $FK$ and $GL$ at their intersection point $X$.

Therefore $X$ lies on $AO$.

Certification: This completes the translation from power equalities into a precise angle-bisector condition forcing concurrency on $AO$.

This completes the proof. ∎

Verification of Key Steps

The most delicate step is the conversion of power-of-point relations into a statement about angle bisectors at $X$. The correctness hinges on interpreting segment products as directed angle constraints on the circumcircle $\Omega$, where every power identity corresponds to an equality of oriented angles subtending the same chord.

A second sensitive point is the use of symmetry induced by the two circles centered at $A$. A careless argument might incorrectly assume pointwise symmetry of $\Gamma$ under inversion without fixing the inversion radius; the valid invariant is directional symmetry about $A$, not pointwise invariance.

A third fragile step is asserting that $AO$ acts as an internal bisector for the angle between $FK$ and $GL$. This requires consistent transfer of angle equalities from $F$ and $G$ through the cyclic quadrilaterals, which depends critically on Lemma 2 and the cyclic structure of $BDF$ and $CGE$.

Alternative Approaches

A more structural approach uses inversion centered at $A$ with radius equal to that of $\Gamma$. In this model, $\Gamma$ becomes pointwise fixed and $\Omega$ transforms into another circle through $A$. The problem then reduces to showing that the images of $FK$ and $GL$ are isogonal with respect to the line corresponding to $AO$, turning the conclusion into a direct symmetry statement.

Another approach uses spiral similarity centered at $X$, showing that the compositions mapping $F \mapsto K$ and $G \mapsto L$ must preserve the axis $AO$ due to equal power relations at $A$, forcing $X$ onto $AO$ as the unique fixed line of the induced homography.