Let $\omega$ be the incircle of triangle $ABC$. For any positive real number $\lambda$, let $\omega_{\lambda}$ be the circle concentric with $\omega$ that has radius $\lambda$ times that of $\omega$. Let $X$ be the intersection between a trisector of $\angle B$ closer to $BC$ and a trisector of $\angle C$ closer to $BC$. Similarly define $Y$ and $Z$. Let $\epsilon = \frac{1}{2024}$. Show that the circumcircle of triangle $XYZ$ lies inside $\omega_{1-\epsilon}$. Note. Weaker results with smaller $\epsilon$ may be awarded points depending on the value of the constant $\epsilon <\frac{1}{2024}$.
Problem
Source: 2024 Korea winter program practice test P8
Tags: geometry, circumcircle
31.01.2024 20:26
It looks like the numerical value $\epsilon=1/2024$ matters and that scares me. Also, may you please post the other problems too?
31.01.2024 20:35
Well, from empirical observations, it might even be true with $\varepsilon \approx 0.45$, but it is very much conceivable that there is a solution which produces a value around $\frac{1}{2024}$ (and that there are other solutions which produce even smaller values). Note that it is not even obvious that this is true for any fixed value of $\varepsilon$.
02.02.2024 12:47
Disclaimer: The proposers of this problem spent many hours on this problem and concluded that any choice of $\epsilon\le 1-\frac{2}{\sqrt{6}}$ will have similar difficulty. The only reason $\epsilon=\frac{1}{2024}$ was given was to intimidate the students and as an extent of the P2 troll(or not troll).
05.02.2024 06:36
MNJ2357 wrote: Disclaimer: The proposers of this problem spent many hours on this problem and concluded that any choice of $\epsilon\le 1-\frac{2}{\sqrt{6}}$ will have similar difficulty. The only reason $\epsilon=\frac{1}{2024}$ was given was to intimidate the students and as an extent of the P2 troll(or not troll). can you give an outline of the solution, plz?
08.02.2024 19:14
bumpity bump
14.02.2024 17:43
bumpity bump bump
12.03.2024 12:16
does anyone have a solution
12.03.2024 16:59
bumpity bump bump bUmP
25.03.2024 18:25
bumpity bump bump bUmP BuMp
07.05.2024 01:31
Seems very interesting, anyone?
07.05.2024 09:12
Here's a proof of the bound given in #4. Note that $\triangle XYZ$ is the first Morley triangle, which is equilateral with sidelengths $8R \sin\left(\frac{A}{3}\right)\sin\left(\frac{B}{3}\right) \sin\left(\frac{C}{3}\right)$. By here we can conclude that the incenter lies inside $\triangle XYZ$. It remains to show that $(1 - \varepsilon) \cdot r \ge 8R \sin\left(\frac{A}{3}\right)\sin\left(\frac{B}{3}\right) \sin\left(\frac{C}{3}\right)$. Substituting the relation for $r = 4R\sin\left(\frac{A}{2}\right)\sin\left(\frac{B}{2}\right) \sin\left(\frac{C}{2}\right) $, we want to show that \[ (1 - \varepsilon) \cdot \sin\left(\frac{A}{2}\right)\sin\left(\frac{B}{2}\right) \sin\left(\frac{C}{2}\right) \ge 2 \sin\left(\frac{A}{3}\right)\sin\left(\frac{B}{3}\right) \sin\left(\frac{C}{3}\right) \] Now consider $K = \prod_{\text{cyc}} \frac{\sin(A/2)}{\sin(A/3)}$. Note that $f(x) = \frac{\sin(x/2)}{\sin(x/3)}$ is strictly decreasing on $(0, \pi)$. If any of $A$ is larger than $\frac{\pi}{2}$, then \[ K \ge f(\pi) \cdot \min_{x + y = \frac{\pi}{2}} f(x)f(y) \ge \frac{2}{\sqrt{3}} \cdot \frac{3}{\sqrt{2}} = \sqrt{6} \]where the value is minimal at $\{x, y\} = \{0, \frac{\pi}{2}\}$ (this is easy to see graphically but there's no immediate nice proof of it). Else, $K \ge f\left(\frac{\pi}{2}\right)^3 = 2\sqrt{2}$. As such, we get that $\varepsilon \le 1 - \frac{2}{K} \le 1 - \frac{2}{\sqrt{6}}$. (The weaker bound $K \ge \frac{4}{\sqrt{3}}$ suffices for the original problem).