Kvant Math Problem 196

Work in the unit circle centered at $O$.

kvantmathematicsolympiad

Verified: no
Verdicts: UNKNOWN + UNKNOWN
Solve time: 2m06s
Source on kvant.digital

Problem

In a circle of radius 1, several chords are drawn. Prove that if each diameter intersects at most $k$ chords, then the sum of the lengths of the chords is less than $\pi k$.

A. T. Kolotov

Exploration

Work in the unit circle centered at $O$. A chord is determined by its endpoints on the circle, and its length depends only on the central angle it subtends. If a chord subtends angle $2\varphi$, then its length is

$$\ell = 2\sin \varphi.$$

A diameter through $O$ is determined by a direction angle $\theta$. The condition that this diameter intersects a fixed chord depends only on how $\theta$ is positioned relative to the chord’s endpoints. If we vary $\theta$ continuously, the set of directions producing an intersection should form an arc on the circle of directions.

The key idea is to convert geometric incidence “a diameter meets a chord” into a statement about angular measure. Then the global constraint “each diameter meets at most $k$ chords” becomes a bound on how many times directions can be covered by these angular intervals. The main task is to relate chord length to the angular size of this interval and compare them quantitatively.

The potential failure point is the exact relation between chord length and the measure of directions of intersecting diameters. This must be computed explicitly and checked carefully.

Problem Understanding

This is a Type B problem, requiring a proof of an inequality.

We are given chords in the unit circle such that any diameter intersects at most $k$ of them, and we must prove that the sum of chord lengths is strictly less than $\pi k$.

The central difficulty is to convert a local intersection constraint over all diameters into a global bound on total chord length, using angular parametrization of directions.

Proof Architecture

For each chord, define the set of directions $\theta \in [0,\pi)$ such that the diameter at angle $\theta$ intersects the chord. This set will be an interval whose length is computed in terms of the chord.

A lemma identifies that if a chord has length $L$, then the measure of directions of intersecting diameters equals $2\arcsin(L/2)$.

A second lemma proves the inequality $2\arcsin(L/2) \ge L$ for $L \in [0,2]$.

A counting argument integrates the overlap condition: since every direction corresponds to a diameter intersecting at most $k$ chords, the total measure of all direction-intervals is at most $\pi k$.

The most delicate step is the precise computation of the direction interval for a chord and its dependence on $L$.

Solution

Fix the unit circle centered at $O$. Identify a diameter with its direction angle $\theta \in [0,\pi)$, where $\theta$ and $\theta+\pi$ define the same unoriented line through $O$.

Let $AB$ be a chord. Let its endpoints subtend a central angle $2\varphi \in (0,\pi]$. Then its length is

$$|AB| = 2\sin \varphi.$$

Consider the set of directions $\theta$ such that the diameter through $O$ at angle $\theta$ intersects the segment $AB$. This happens exactly when the line through $O$ is not separating $A$ and $B$ with respect to perpendicular projection. Equivalently, the diameter misses the chord precisely when both endpoints lie strictly on the same open half-plane bounded by the diameter line.

The boundary cases occur when the diameter passes through $A$ or $B$. As $\theta$ varies, these boundary positions correspond to the two lines through $O$ tangent to the angular sector determined by $OA$ and $OB$. The set of intersecting directions is therefore an interval on $[0,\pi)$ whose endpoints correspond to directions orthogonal to $OA$ and $OB$ shifted appropriately. A direct angular computation shows that the length of this interval equals the central angle subtended by the chord:

$$\mu(AB) = 2\varphi.$$

Using the chord-length relation $|AB| = 2\sin\varphi$, we express

$$\mu(AB) = 2\arcsin\left(\frac{|AB|}{2}\right).$$

Now consider the family of all chords. For each direction $\theta \in [0,\pi)$, let $n(\theta)$ be the number of chords intersected by the diameter in direction $\theta$. By assumption, for every $\theta$,

$$n(\theta) \le k.$$

Integrating over all directions,

$$\int_0^\pi n(\theta),d\theta \le \pi k.$$

On the other hand, each chord contributes exactly the measure of directions for which it is intersected, hence

$$\int_0^\pi n(\theta),d\theta = \sum \mu(AB) = \sum 2\arcsin\left(\frac{|AB|}{2}\right).$$

We compare each term with the chord length. For $x \in [0,1]$, the function $\arcsin x$ satisfies $\arcsin x \ge x$. This follows from convexity of $\arcsin x$ on $[0,1]$ and equality at $x=0$. Substituting $x = |AB|/2$, we obtain

$$2\arcsin\left(\frac{|AB|}{2}\right) \ge |AB|.$$

Summing over all chords,

$$\sum 2\arcsin\left(\frac{|AB|}{2}\right) \ge \sum |AB|.$$

Combining inequalities,

$$\sum |AB| \le \int_0^\pi n(\theta),d\theta \le \pi k.$$

Since equality in $\arcsin x \ge x$ is strict for $x \in (0,1]$, and at least one chord has positive length, the final inequality is strict:

$$\sum |AB| < \pi k.$$

Verification of Key Steps

The first delicate point is the identification of the direction set measure for a chord. This depends only on the geometry of a central angle $2\varphi$. The two boundary directions correspond to when the diameter through $O$ passes through an endpoint of the chord. These two directions differ by angle $2\varphi$, since rotating the diameter by that amount moves one endpoint to the other in angular position. This confirms $\mu(AB)=2\varphi$.

The second point is the inequality $\arcsin x \ge x$ on $[0,1]$. Differentiating gives

$$\frac{d}{dx}(\arcsin x - x) = \frac{1}{\sqrt{1-x^2}} - 1 \ge 0,$$

since $\sqrt{1-x^2} \le 1$. Hence $\arcsin x - x$ is increasing with value $0$ at $x=0$, so it is nonnegative.

The third point is the exchange of sum and integral. Each chord contributes to $n(\theta)$ exactly on the set of directions where it is intersected, so summing indicator functions over chords reproduces $n(\theta)$ pointwise, validating the identity under integration.

Alternative Approaches

A different approach uses integral geometry directly via Crofton’s formula, interpreting chord length as an integral over projections of the segment onto varying directions. One can also argue using probabilistic interpretation: choosing a random diameter direction, expected number of intersections is bounded by $k$, while each chord contributes a probability proportional to its angular visibility, leading again to a linear bound in terms of $\pi k$. The angular-interval method is preferable here because it avoids external machinery and keeps the argument fully elementary while still capturing the same geometric content.