Both roots of the equation $$(x-b) (x-c) +(x-a) (x-c) +(x-a) (x-b) = 0$$ are always positive , negative or real. Prove your result.

By solving this equation I got $3x^2 - 2(a+b+c)x +ab + bc + ca = 0$

Now by inspection of this quadratic : $a>0$ and $D = 4(a^2 + b^2 + c^2 -ab - bc - ca)$

I'm stuck here and unable to conclude whether $D>0$ , $D<0$ or $D = 0$

Also is there an alternative approach to solve these types of questions. Kindly suggest any other method (if any) too.

By various answers here now I know where I was stuck. But I just want to know alternate method to approach this question without having to do this much algebra.

  • $\begingroup$ What does "positive, negative or real" mean ? $\endgroup$ – Captain Lama Apr 21 '16 at 8:48
  • $\begingroup$ @CaptainLama the question asks whether roots of the given quadratic equation are "positive, negative or real" $\endgroup$ – Saksham Apr 21 '16 at 8:49
  • $\begingroup$ Yes but that doesn't make a lot of sense : you have to be real for words such as "positive" and negative to make sense. $\endgroup$ – Captain Lama Apr 21 '16 at 8:53
  • 3
    $\begingroup$ Consider $(a-b)^2+(a-c)^2+(b-c)^2$. $\endgroup$ – Gerry Myerson Apr 21 '16 at 8:53
  • 1
    $\begingroup$ I agree that I have a linear algebra "trick" to obtain rather easily this kind of terms' grouping called "Cholesky decomposition" $\endgroup$ – Jean Marie Apr 21 '16 at 14:43

Edit : I divide this answer into 2 parts. The first one originates as a quick answer to the question; the second one has my preference because it provides an intuitive understanding of the situation.

Part 1 :

With your notations ($D$ being the discriminant):

$$\dfrac{D}{4} = a^2 + b^2 + c^2 -ab - bc - ca=\dfrac{1}{4}(2a-b-c)^2+\dfrac{3}{4}(b-c)^2$$

Thus, in general $D\geq 0$ with:

  • $D=0$ when simultaneously $b=c$ and $2a-b-c=0$, which means if and only if all $a=b=c$. (case of a single root).

  • $D>0$ otherwise. (two real roots).

Part 2 :

Before beginning we assume that $a,b,c$ are distinct and WLOG that $a<b<c$.

Dividing by $(x-a)(x-b)(x-c)$, we have the following equivalent equation .

$$\dfrac{1}{x-a}+\dfrac{1}{x-b}+\dfrac{1}{x-c}=0 \ \ \ (1)$$

In fact, set $f(x)$ for the LHS of (1); it is immediate that f'(x)<0, thus $f$ is a strictly decreasing function (see graphical representation example below). Consider interval $(a,b)$: $lim_{x \rightarrow a+0} f(x) =+\infty$ and $lim_{x \rightarrow b-0} f(x) =-\infty$. Thus there is a root $x_1$ in this interval. Besides, this root is unique because $f$ is strictly decreasing. The same for interval $(b,c)$.

Thus we have two real roots such that:


No more than two real roots may occur as we know from the first form of the equation.

enter image description here

  • $\begingroup$ Is there any other easy approach without having to do this much algebra? $\endgroup$ – Saksham Apr 21 '16 at 9:07
  • $\begingroup$ What do you think about this other approach ? $\endgroup$ – Jean Marie Apr 21 '16 at 13:31
  • $\begingroup$ I don't know what WLOG is? $\endgroup$ – Saksham Apr 21 '16 at 14:33
  • 1
    $\begingroup$ Without Loss Of Generality $\endgroup$ – Jean Marie Apr 21 '16 at 14:40

$$f(x)=(x-a)(x-b)+(x-b)(x-c)+(x-c)(x-a)$$ Then, $$f(a)=(a-b)(a-c)$$ $$f(b)=(b-c)(b-a)$$ $$f(c)=(c-a)(c-b)$$

Clearly, the coefficient of $x^2$ in $f(x)$ is positive. Suppose there are no real roots. Then, $f(x)\gt0 \forall x\in\mathbb R$. $$(a-b)(a-c)\gt0$$ Without loss of generality, let $a\gt b$ and $a\gt c$. $$(b-c)(b-a)\gt 0\Rightarrow c\gt b$$ $$(c-a)(c-b)\gt 0\Rightarrow b\gt c$$ This leads to a contradiction. Thus, the graph must meet the $x$-axis and thus the roots are real.

  • 1
    $\begingroup$ Yeah a good approach. $\endgroup$ – Saksham Apr 21 '16 at 9:32

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.