Let $A_1B_1C_1D_1E_1$ be a regular pentagon.For $ 2 \le n \le 11$, let $A_nB_nC_nD_nE_n$ be the pentagon whose vertices are the midpoint of the sides $A_{n-1}B_{n-1}C_{n-1}D_{n-1}E_{n-1}$. All the $5$ vertices of each of the $11$ pentagons are arbitrarily coloured red or blue. Prove that four points among these $55$ points have the same colour and form the vertices of a cyclic quadrilateral.
Problem
Source:
Tags: combinatorics
20.01.2019 16:56
PHP FTW. I got 5 is enough instead of 11
20.01.2019 16:59
TheDarkPrince wrote: PHP FTW. I got 6 is enough instead of 11 No, it may be there that the 2 edges sharing the same coloured vertices are of different colour. So, there must me 3 parellel edges, so it should be$\geq11$
20.01.2019 17:00
TheDarkPrince wrote: PHP FTW. I got 6 is enough instead of 11 Your solution?
20.01.2019 17:03
TheDarkPrince wrote: PHP FTW. I got 6 is enough instead of 11 I guess I know your solution, but I think that it should be 11 instead of 6. I may be wrong.
20.01.2019 17:09
According to my solution, 9 are enough
20.01.2019 17:10
Math-wiz wrote: According to my solution... Can you show us?
20.01.2019 17:11
Math-wiz wrote: According to my solution, 9 are enough No, way, still share the solution.
20.01.2019 17:27
EASY PHP (I guess....) Notice that some two adjacent vertices of every regular pentagon must be of the same color. Also by joining the mid points of a regular pentagon we get another regular pentagon whose sides are parallel to each other. Now, note that for any pentagon we have 5 pairs of adjacent vertices and 2 colors. Therefore, there are 10 possible colorings of the adjacent vertices. And since we have 11 pentagons ,the corresponding adjacent vertices of some two pentagons must be of the same color (by PHP). These four points form a monochromatic issosceles trapezium which is cyclic. Hence we are done!!!
20.01.2019 18:43
Proof for $n=5$: (high chance its false). Notice that $A_iA_jB_iB_j$ is isosceles trapezium. If $A_nB_nC_nD_n$ are same color, we are done. WLOG, $A_1,B_1,C_1$ are red. Therefore for all $n\geq 2$, at least two of $A_n,B_n,C_n$ are blue. WLOG $A_2,B_2$ are blue. Now for $n=3,4,5$ by PHP at least one of $A_n,C_n$ and $B_n,C_n$ are blue for more than one $n$. Therefore we are done if $11$ is replaced by $5$ also.
20.01.2019 18:53
I brute forced it and it got killed in n=3.
20.01.2019 18:57
A counterexample with n=9, Sorry for my bad handwriting, but I am too lazy to latex this.
Attachments:

20.01.2019 19:03
...........
20.01.2019 19:04
^^ What's $D_4$? If it's $R_1$, $B_1D_1D_4B_4$ is cyclic and if it's $R_2$, $A_2D_2D_4A_4$.
20.01.2019 19:08
If anyone is trying to find a counterexample of any sorts, keep in mind that 2 consecutive pentagons(like a1b1c1d1e1 and a2b2c2d2e2) can't both have 3 vertices red or 3 vertices blue- it has to be alternating-this is easy to prove by some length casework.
20.01.2019 21:00
aadhitya wrote: EASY PHP (I guess....) Notice that some two adjacent vertices of every regular pentagon must be of the same color. Also by joining the mid points of a regular pentagon we get another regular pentagon whose sides are parallel to each other. Now, note that for any pentagon we have 5 pairs of adjacent vertices and 2 colors. Therefore, there are 10 possible colorings of the adjacent vertices. And since we have 11 pentagons ,the corresponding adjacent vertices of some two pentagons must be of the same color (by PHP). These four points form a monochromatic issosceles trapezium which is cyclic. Hence we are done!!! Why should two adjacent vertices of a regular Pentagon be of the same colour?
20.01.2019 21:06
If we can prove that then it follows that if AkBkCkDkEk has 2 red and 3 blues(WLOG) Then Ak+1Bk+1Ck+1Dk+1Ek+1 should have 3 red and two blue vertices (Which can be proved by taking cases) And the number of red and blue vertices alternates for Pentagon k and Pentagon k+1 WlOG assume A1B1C1D1E1 has 2 red and 3 blue Then out of the 55 vertices 28 should be red and 27 blue...But we wanted 22 red and 33 blue or 22 blue or 33 red... Q.e.d
21.01.2019 11:16
I think $n=3$ suffices.
Can anyone point out the mistake? Because this is (probably) wrong.
21.01.2019 11:22
@Supercali, Bravo!! Your solution works just fine and $n=3$ suffices.
21.01.2019 11:39
Supercali wrote: I think $n=3$ suffices.
Can anyone point out the mistake? Because this is (probably) wrong. Yup l also think $n=3$ works. I case bashed all the possibilities.
21.01.2019 13:42
Wow the original bound is way too weak. In fact, if we had 4 colours, the result would hold true for $n=21$ (because then each pentagon will have at least one monochromatic side/diagonal). @below: Edited, sorry
21.01.2019 17:16
@above, Are you sure? Because now you have $11$ colours and $4\times5=20$ labels. However, $3$ colours and 8 pentagons work as $2\times8>3\times5$.
14.02.2019 07:46
TheDarkPrince wrote:
Does this work btw?
14.02.2019 08:13
TheDarkPrince wrote: Proof for $n=5$: Notice that $A_iA_jB_iB_j$ is isosceles trapezium. If $A_nB_nC_nD_n$ are same color, we are done. WLOG, $\color{red} A_1,B_1,C_1$ are red.Therefore for all $n\geq 2$, at least two of $A_n,B_n,C_n$ are blue. WLOG $A_2,B_2$ are blue. Now for $n=3,4,5$ by PHP at least one of $A_n,C_n$ and $B_n,C_n$ are blue for more than one $n$. Therefore we are done if $11$ is replaced by $5$ also. If I understand correctly, $ A_1,B_1,C_1$ are three consecutive vertices, but then how are you WLOG assuming this? Also you should clarify on how you are naming the vertices of the “inner” pentagons. @below. Your proof seems correct then
14.02.2019 09:39
ayan.nmath wrote: If I understand correctly, $ A_1,B_1,C_1$ are three consecutive vertices, but then how are you WLOG assuming this? Also you should clarify on how you are naming the vertices of the “inner” pentagons. The naming is like: $A_{n+1}$ is the midpoint of $C_nD_n$, $E_{n+1}$ is midpoint of $B_nC_n$, and so on. Well, it doesn't matter if they are consecutive or not. Any three points have the same proof (as my proof never uses the fact of consecutive).
16.06.2019 14:17
Well, it's pretty late now but here's the solution which fetched me 2/17 @below 31 (same as your post number)
Attachments:
INMO P2.pdf (146kb)
16.06.2019 14:20
AhirGauss wrote: Well, it's pretty late now but here's the solution which fetched me 2/17 What was ur total?
30.11.2019 07:03
12.01.2020 13:43
monsterDJ wrote: I brute forced it and it got killed in n=3.
05.02.2020 13:41
richardrichard wrote: monsterDJ wrote: I brute forced it and it got killed in n=3.
Which material do you use to prepare for combinatorial colouring theory? Please help as I am having problems with colouring problems