# Show that this inequality doesn't hold

Given $(a,b,c) \in \mathbb R^3_+$ show that atleast one of the real numbers $a(1-b)$, $b(1-c)$ and $c(1-a)$ is less than or equal to 1\4.

I tried to show it by contradiction i.e Suppose that $$a(1-b) > \frac{1}{4}$$ $$b(1-c) > \frac{1}{4}$$$$c(1-a) > \frac{1}{4}$$

Note that if $a,b,c$ are all equal to $\frac{1}{2}$, then the given products are all $\frac{1}{4}$. So it is not true that at least one of $a(1-b)$, $b(1-c)$, and $c(1-a)$ is $\lt \frac{1}{4}$. But we can prove that at least one is $\le \frac{1}{4}$.

If one or more of $1-a$, $1-b$, $1-c$ is $\le 0$, the result is obvious. So suppose they are all positive.

Their product is $a(1-a)(b)(1-b)(c)(1-c)$ (note that we rearranged). Recall, or prove using AM/GM, or in some other way, that if $0\lt x\lt 1$, then $x(1-x)\le \frac{1}{4}$. Thus $a(1-a)(b)(1-b)(c)(1-c)\le \frac{1}{4^3}$, and therefore at least one of the given products is $\le \frac{1}{4}$.

• @user355971: You are welcome. – André Nicolas Jul 24 '16 at 15:46

Follow the idea of contradiction: if $a(1-b)> \frac{1}{4}$ and other two similar inequality hold stimutaneously, than from AM-GM,

$$(\frac{a-b+1}{2})^2\ge a(1-b)> \frac{1}{4},$$ $$a-b+1> 1, \ \ a-b>0$$ the other two lead to $b-c>0$ and $c-a>0$ , summing up gives an contradiction.

Suppose that for all $a,b,c$ $$a(1-b) > \frac{1}{4}\\b(1-c) > \frac{1}{4}\\c(1-a) > \frac{1}{4}$$ We then have $$a+b+c\gt \frac 34+(ab+ac+bc)=\frac 34+\frac{(a+b+c)^2-(a^2+b^2+c^2)}{2}$$ It follows that if $a=b=c$ we have $$12a\gt 15a^2+3\Rightarrow(15a-6)^2+45\lt 0$$ This is absurde.