I trying to find the roots of the equation
$$ax^3 + bx^2 + cx + d = 0$$
By using some changes of variable (which does not really matter now) I was able to rewrite this equation as
$$z^3 - \frac{\Delta_{_{0}}}{3a^2}z+\frac{\Delta_{_{1}}}{27a^3} = 0$$
I used then the Vieta's substitution
$z = w + \frac{\Delta_{_{0}}}{9a^2w}$
to obtain the following equation:
$$(w^3)^2 + \frac{\Delta_{_{1}}}{27a^3}w^3 + \frac{\Delta_{_{0}}^{3}}{729a^6} = 0$$
This is a quadratic equation on $w^3$ with roots:
$$ w^3 = \frac{1}{27a^3}\,\frac{-\Delta_{_{1}} \pm \sqrt{\Delta_{_{1}}^{2}-4\Delta_{_{0}}^{3}}}{2}$$
There are then six possible values for $w$. Denoting
$$ C=\sqrt[3]{\frac{\Delta_{_{1}}+\sqrt{\Delta_{_{1}}^{2}-4\Delta_{_{0}}^{3}}}{2}} \text{ and } \bar{C}=\sqrt[3]{\frac{\Delta_{_{1}}-\sqrt{\Delta_{_{1}}^{2}-4\Delta_{_{0}}^{3}}}{2}}$$
they are
$$ w = -\frac{\mu_{_{k}}}{3a}C \text{ and } w = -\frac{\mu_{_{k}}}{3a}\bar{C} $$
where $\mu_{_{k}} = \left\{ 1, \frac{-1 \pm i\sqrt{3}}{2} \right\}$ for $k=1,2,3.$
Finally... my point is: I can substitute any of the six $w$'s into the Vieta's substitution formula and obtain six values for $z$ when only three should exist. What am I missing?
UPDATE
Since $u_{_{k}}^{3} = 1$ there are three equivalent solutions for $C$ and three for $\bar{C}$ and so there are only two (possibly) distinct solutions for $w$:
$ w = -\frac{1}{3a}C \,\,\,\,\,\,\,\,\,\, \mathrm{and} \,\,\,\,\,\,\,\,\,\, w = \frac{1}{3a}\bar{C}$
which will give me only two possible values for z. A short while ago I had three solutions too many and now I'm one short. The confusion goes on...
Thank you very much.
Regardless of whether all six are distinct (and I suspect not), it should be clear that Vieta's substitution does not typically have a unique $w$ that yields a given $z$. In that substitution, $w$ is a quadratic function of $z$.