2016 UMO Problems/Problem 3

Revision as of 03:09, 22 January 2019 by Timneh (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Problem

Can each positive integer $1,2,3,\ldots$ be colored either red or blue, such that for all positive integers $a,b,c,d$ (not necessarily distinct), if $a + b + c = d$ then $a,b,c,d$ are not all the same color?


Solution

Solution We claim that no such coloring exists. For the sake of contradiction, assume that such a coloring exists. Without loss of generality, we may assume that 1 is colored red. In the string of steps below, if a number n is known to be colored red, we denote it as $n^R$, and if it is known to be colored blue, we denote it as $n^B$. Each of the implications below follow from the fact that if all but one of the distinct numbers in $a + b + c = d$ are the same color, then the final number must be a different color. $1^R + 1^R + 1^R = 3 \Rightarrow 3$ is blue. $3^B + 3^B + 3^B = 9 \Rightarrow 9$ is red. $1^R + 4 + 4 = 9^R \Rightarrow 4$ is blue. Now we have $1^R + 1^R + 9^R = 11 \Rightarrow 11$ is blue AND $3^B + 4^B + 4^B = 11 \Rightarrow 11$ is red. This is a contradiction because $11$ can only have one color. Therefore, no such coloring exists.


See Also

2016 UMO (ProblemsAnswer KeyResources)
Preceded by
Problem 2
Followed by
Problem 4
1 2 3 4 5 6
All UMO Problems and Solutions