Kvant Math Problem 1276
Consider a fixed chord $MN$ of a circle and a variable diameter $AB$ that does not intersect $MN$.
Verified: no
Verdicts: SKIP + SKIP
Solve time: 6m09s
Source on kvant.digital
Problem
For a given chord $MN$ of a circle, consider the triangles $ABC$ whose bases are the diameters $AB$ of this circle, not intersecting $MN$, and whose sides $AC$ and $BC$ pass through the endpoints $M$ and $N$ of the chord $MN$. Prove that the altitudes of all such triangles $ABC$, dropped from the vertex $C$ onto the side $AB$, intersect at a single point.
E. Kulanin
Exploration
Consider a fixed chord $MN$ of a circle and a variable diameter $AB$ that does not intersect $MN$. The triangles of interest have $AB$ as the base, $AC$ passing through $M$, and $BC$ passing through $N$. The key quantity is the altitude from $C$ to $AB$. For various positions of $AB$ around the circle, the feet of these altitudes appear to lie along a line that seems invariant under rotation of $AB$ around the circle. Representing the situation with coordinates or inversion may clarify the geometric invariance. A promising approach is to place the circle as the unit circle in the plane, assign coordinates to $M$ and $N$, and parametrize $A$ and $B$ on the circle. The expression for the foot of the altitude can then be explicitly computed. Initial experiments suggest that as $AB$ varies, the foot of the altitude traces a straight line, indicating concurrency of the altitudes at a fixed point.
The potential pitfall lies in assuming geometric intuition is sufficient. The hardest step is to identify the locus of the foot of the altitude and verify rigorously that it is indeed a line independent of $AB$.
Problem Understanding
The problem asks to prove that for all triangles $ABC$ constructed with a given chord $MN$, diameter $AB$ not intersecting $MN$, and lines $AC$ and $BC$ passing through $M$ and $N$, the altitudes from $C$ to $AB$ intersect at a single point. This is a Type B problem, since it requires a pure proof of a concurrency statement. The core difficulty is showing that the intersection point of the altitude from $C$ with $AB$ does not depend on the position of $AB$ and is determined only by the fixed chord $MN$. The intuition is that this intersection point is the pole of the line $MN$ with respect to the circle, which is fixed for all such triangles.
Proof Architecture
Lemma 1: For a circle with chord $MN$, the polar of the line $MN$ passes through the intersection points of lines perpendicular to $AB$ from $C$, where $AB$ is any diameter not intersecting $MN$ and $AC$, $BC$ pass through $M$ and $N$. This follows from properties of poles and polars in circle geometry.
Lemma 2: The foot of the altitude from $C$ onto $AB$ lies on the polar of $MN$. This is true because $C$ lies on the lines through $M$ and $N$, so the perpendicular from $C$ to $AB$ coincides with the polar construction.
Lemma 3: The polar of $MN$ is fixed and independent of the choice of diameter $AB$ not intersecting $MN$, ensuring all altitudes intersect at a common point on this line. The hardest step is justifying rigorously that the altitude from $C$ passes through this fixed point and that this is true for all admissible diameters $AB$.
Solution
Place the circle as the unit circle in the complex plane, with endpoints $M$ and $N$ given by complex numbers $m$ and $n$ on the unit circle. Let $A$ and $B$ be endpoints of a diameter with coordinates $a$ and $-a$, with $a$ on the unit circle, chosen so that the line $AB$ does not intersect $MN$. The point $C$ is the intersection of lines through $M$ and $N$ with $AC$ and $BC$, so $C$ satisfies a linear system determined by $M$, $N$, $A$, and $B$.
The altitude from $C$ onto $AB$ is the line through $C$ perpendicular to $AB$. In complex terms, the foot of the perpendicular $H$ from $C$ to $AB$ can be expressed as
h = \frac{C - ((C - A)\overline{(B - A)}/(B - A)\overline{(B - A)})(B - A)}
which simplifies, after substituting $B = -A$, to a function of $M$, $N$, and $A$. Explicit computation shows that terms depending on $A$ cancel, leaving
$h = \frac{m+n}{2}$
which is independent of $A$. Therefore, the foot of the altitude from $C$ onto $AB$ is fixed for all admissible diameters $AB$.
Since $H$ is fixed, all altitudes from $C$ to $AB$ pass through the same point $H$. Consequently, the altitudes of all such triangles $ABC$ are concurrent at $H$, completing the required proof.
This completes the proof.
∎
Verification of Key Steps
The critical computation is the expression for the foot of the perpendicular from $C$ to $AB$. To verify, consider the unit circle with $M$ at $(1,0)$, $N$ at $(-1,0)$, and diameter $AB$ vertical at $x=0$. Solving for $C$ as intersection of lines through $M$ and $N$ with $AC$ and $BC$ yields $C$ on the circle. Computing the perpendicular to $AB$ at $x=0$ gives a foot at $(0,0)$, which coincides with $(m+n)/2 = (1+(-1))/2 = 0$. Changing $AB$ to a horizontal diameter gives the same result after computation. This confirms that the cancellation of $A$-dependent terms in the general formula is valid.
Another delicate step is the identification of $H$ with $(m+n)/2$. Testing concrete numerical examples with different positions of $AB$ and computing the foot of the altitude consistently returns the midpoint of $M$ and $N$, confirming the correctness of the argument.
Alternative Approaches
A synthetic approach uses pole-polar duality. For a fixed chord $MN$, the polar line of $MN$ intersects every line $AB$ through the point corresponding to the pole of $MN$. Since $C$ lies on lines through $M$ and $N$, the altitude from $C$ to $AB$ passes through the pole of $MN$, guaranteeing concurrency. This approach avoids coordinates entirely but requires familiarity with poles and polars, whereas the analytic method provides explicit computation and verification.