We have the following inequality: $$ x < - \sqrt{x^2-8x+4} $$ It is very easy to check that it has no solution (graphically). But I'm having a lot of trouble with a proof. I know for instance that squaring both sides is no good because the $(-)$ sign disappears. Can anybody suggest a way to do it? Thanks.

  • $\begingroup$ Often considering the cases where the LHS/RHS are positive/negative help when you want to square both sides. $\endgroup$ – Shuri2060 Jul 2 '17 at 1:47
  • 1
    $\begingroup$ " I know for instance that squaring both sides is no good because the - sign disappears." That's not a problem at all because you know x is negative. So square it and flip the inequality. $\endgroup$ – fleablood Jul 2 '17 at 4:06

The inequality does not hold for non-negative $x$, because the left hand side is non-negative but the right hand side is non-positive.

For $x <0$, both sides of the inequality is non-positive. Square the original inequality and note that the inequality sign should be reversed,

$$\begin{align*} x &< -\sqrt{x^2-8x+4}\\ x^2 &> x^2 - 8x + 4\\ 0 &> -8x + 4\\ x &> \frac12 \end{align*}$$

The assumption was that $x < 0$, so there is also no solution for the case $x < 0$.


We need to solve $$\sqrt{x^2-8x+4}<-x,$$ which gives $-x>0$ and $x^2-8x+4\geq0$.

Now we can use squaring of the both sides, which gives $$x^2-8x+4<x^2$$ or $$x>\frac{1}{2},$$ which is impossible because we have at least $-x>0$.

Thus, our inequality has no solutions.


To prove $x < - \sqrt{x^2-8x+4}$ is not true for all $x \in \mathbb{R}$, you can instead prove that $x \ge - \sqrt{x^2-8x+4} \,\,\,\,\,\forall\,\,\,\,\, x \in \mathbb{R}$ when the RHS is defined as the two statements are equivalent using the law of trichotomy.

It is easy to see this holds for $x>0$ as the LHS is positive while the RHS is not.

Assume $x<0$.


$$x \le \frac{1}{2}$$

$$\iff 0 \le -8x+4$$

$$\iff 0 \lt x^2 \le x^2-8x+4$$

$$\iff 0 \lt \sqrt{x^2} \le \sqrt{x^2-8x+4}$$


$$\iff -x \le \sqrt{x^2-8x+4}$$

$$\iff x \ge -\sqrt{x^2-8x+4}$$

This is essentially the same method as peterwhy's, but a different way of looking at it.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.