I'm trying to find a condition on a, b and c for the quartic $P(x)=x^4+ax^3+bx+c$ to have a triple root.
Using the Multiple Root Theorem, it's easy to show that if it has a triple root, it must be $\alpha=-\frac{a}{2}$.
So the usual method of finding the condition is to substitute this back into the original polynomial. In other words, $P\left ( -\frac{a}{2} \right )=0$ should yield the (necessary?) condition for the polynomial to have a triple root.
After some computation, I get
- $a^4+8ab=16c \qquad (1)$
Here is where I come across issues.
In order to check that this actually works, I set $a=1, b=1$ and found $c=\frac{9}{16}$, which does NOT have a triple root from the graphing calculator.
My thought process then was "Oh! It must be because all triple roots must also be double roots as well, so perhaps I need to substitute $x=-\frac{a}{2}$ back into $P'(x)$ to get a second condition that must be satisfied!"
So I worked with
$P'\left ( -\frac{a}{2} \right )=0 $ to get
- $a^3+4b=0 \qquad (2)$
My thought process afterwards was "Okay, so if I find a,b,c satisfying both (1) AND (2), then that should yield a polynomial with a triple root!"
I set $a=2$, which gave me $b=-2$ using (2) which then using (1) yielded $c=-1$. So this triplet should satisfy both conditions (1) and (2).
But the graph does not have a triple root unfortunately.
What is going on? Is it perhaps to do with the fact that the conditions are necessary, but not sufficient?

In that case, why does a triplet (a,b,c) satisfying (1) not yield a triple root, whereas (a,b,c) satisfying (1) AND (2) work?
For example, (2,-2,-1) works as it's derived from (1) and (2), whereas (1,1,9/16) doesn't work, and it's derived purely from (1).
– Trogdor Sep 11 '14 at 16:12P''(k)=0 ---> potential inflexion point.
P'(k)=0 ---> potential horizontal inflexion point.
P(k)=0 ---> lies on the x axis.
– Trogdor Sep 11 '14 at 16:29