Difference between revisions of "2013 APMO Problems/Problem 5"
(→Solution 6) |
|||
(3 intermediate revisions by the same user not shown) | |||
Line 39: | Line 39: | ||
Use barycentrics on <math>CBD</math>, in that order. It is easy to derive that <math>P=(-a^2:b^2:c^2)</math> and <math>A=(-a^2:2b^2:2c^2)</math>. Clearly, line <math>PD</math> has equation <math>\frac{x}{y}=\frac{-a^2}{b^2}</math>, and the tangent from <math>C</math> has equation <math>c^2y+b^2z=0</math>, so we get that <math>Q=(-a^2:b^2:c^2)</math>. Line <math>AD</math> has equation <math>\frac{x}{y}=\frac{-a^2}{2b^2}</math>, so we also get that <math>R=(-a^2:2b^2:-2c^2)</math>. Line <math>AQ</math> has equation <math>4b^2c^2x+3a^2c^2y-a^2b^2z=0</math>, and line <math>BR</math> has equation <math>\frac{x}{z}=\frac{a^2}{2c^2}</math>, so we quickly derive that <math>E</math> has coordinates <math>(3a^2:-2b^2:6c^2)</math>, and it is easy to verify that this lies on the circumcircle <math>{\sum_{\text{cyc}}a^2yz=0}</math>, so we're done. | Use barycentrics on <math>CBD</math>, in that order. It is easy to derive that <math>P=(-a^2:b^2:c^2)</math> and <math>A=(-a^2:2b^2:2c^2)</math>. Clearly, line <math>PD</math> has equation <math>\frac{x}{y}=\frac{-a^2}{b^2}</math>, and the tangent from <math>C</math> has equation <math>c^2y+b^2z=0</math>, so we get that <math>Q=(-a^2:b^2:c^2)</math>. Line <math>AD</math> has equation <math>\frac{x}{y}=\frac{-a^2}{2b^2}</math>, so we also get that <math>R=(-a^2:2b^2:-2c^2)</math>. Line <math>AQ</math> has equation <math>4b^2c^2x+3a^2c^2y-a^2b^2z=0</math>, and line <math>BR</math> has equation <math>\frac{x}{z}=\frac{a^2}{2c^2}</math>, so we quickly derive that <math>E</math> has coordinates <math>(3a^2:-2b^2:6c^2)</math>, and it is easy to verify that this lies on the circumcircle <math>{\sum_{\text{cyc}}a^2yz=0}</math>, so we're done. | ||
+ | |||
+ | === Solution 5 === | ||
+ | |||
+ | First off, let's state the constructions: | ||
+ | |||
+ | Let <math>AE \cap BD \cong K </math>; let the tangent from <math>R </math> to <math>(ABCD) </math> not containing <math>C </math> be <math>l </math>, and let <math>l \cap (ABCD) \cong F </math>. | ||
+ | |||
+ | Now we make a couple of slick observations; | ||
+ | |||
+ | Since <math>ACED</math> is harmonic, <math>AE </math> is the <math>A-</math> symmedian in <math>\triangle ACD </math>, and since <math>ABCD </math> is harmonic, <math>DB </math> is the <math>B-</math> symmedian in <math>\triangle ACD </math>. Then <math>K </math> is the symmedian point of <math>\triangle ACD </math>. | ||
+ | |||
+ | Now by the definition of <math>F </math>, <math>ACDF </math> is harmonic. Hence <math>CF </math> is a symmedian of <math>\triangle ACD </math> as well! Hence <math>K </math> lies on <math>CF </math>. | ||
+ | |||
+ | Now since <math>R </math> is the pole of <math>\overline{CF} </math>, <math>R </math> lies on the polar of point <math>K </math> (w.r.t. <math>(ABCD) </math>, obviously). But by Brocard's theorem on quadrilateral <math>ABED </math>, <math>BE \cap AD </math> lies on the polar of <math>K </math>. | ||
+ | |||
+ | Hence <math>R </math> lies on <math>\overleftrightarrow{BE} </math> and we are done. | ||
+ | |||
+ | === Solution 6 === | ||
+ | |||
+ | Note that this problem is purely projective! So we can take a projective transformation fixing <math>(ABCD) </math> and taking <math>\overline{AC} \cap \overline{BD} </math> to the center of <math>(ABCD) </math>. This implies <math>ABCD </math> is a rectangle. But it is also harmonic; hence it is a square. | ||
+ | |||
+ | Now by some simple angle chasing, we get <math>\angle DEB = 90^{\circ} </math>, <math>\angle DEQ = 135^{\circ} </math> and <math>\angle REQ = 45^{\circ} </math>. | ||
+ | |||
+ | <math>\implies </math> <math>\angle DER = 90^{\circ} </math>, which proves that <math>B </math>, <math>E </math>, <math>R </math> are collinear, finishing our proof. | ||
http://www.artofproblemsolving.com/community/c6h532682p3046946 | http://www.artofproblemsolving.com/community/c6h532682p3046946 |
Latest revision as of 07:46, 25 November 2024
Contents
Problem
Let be a quadrilateral inscribed in a circle , and let be a point on the extension of such that and are tangent to . The tangent at intersects at and the line at . Let be the second point of intersection between and . Prove that , , are collinear.
Solution
Solution 1
Let . Note that the tangents at and concur on at , so is harmonic, hence the tangents at and concur on at , say.
Now apply Pascal's Theorem to hexagon to find that , and are collinear. Now note that and both lie on the tangent at , hence also lies on the tangent at . It follows that . So and are in fact the same point. Since lies on by definition, it follows that , are indeed collinear, and thus the problem is solved.
Solution 2
We use complex numbers. Let be the unit circle, and let the lowercase letter of a point be its complex coordinate.
Since lies on the intersection of the tangents to at and , we have . In addition, lies on chord , so . This implies that , or .
lies on the tangent at , and lies on , so .
lies on chord and on the tangent at . Therefore we have and . Solving for yields are collinear, so we have , or We must prove that are collinear, or that or Cross-multiplying, we have which is true.
Solution 3
Set , , and , where . Note that since is harmonic, we have collinear and with But is harmonic; therefore .
Solution 4
Use barycentrics on , in that order. It is easy to derive that and . Clearly, line has equation , and the tangent from has equation , so we get that . Line has equation , so we also get that . Line has equation , and line has equation , so we quickly derive that has coordinates , and it is easy to verify that this lies on the circumcircle , so we're done.
Solution 5
First off, let's state the constructions:
Let ; let the tangent from to not containing be , and let .
Now we make a couple of slick observations;
Since is harmonic, is the symmedian in , and since is harmonic, is the symmedian in . Then is the symmedian point of .
Now by the definition of , is harmonic. Hence is a symmedian of as well! Hence lies on .
Now since is the pole of , lies on the polar of point (w.r.t. , obviously). But by Brocard's theorem on quadrilateral , lies on the polar of .
Hence lies on and we are done.
Solution 6
Note that this problem is purely projective! So we can take a projective transformation fixing and taking to the center of . This implies is a rectangle. But it is also harmonic; hence it is a square.
Now by some simple angle chasing, we get , and .
, which proves that , , are collinear, finishing our proof.
http://www.artofproblemsolving.com/community/c6h532682p3046946