Here is the question I have to solve:
Find value of $x$ if,
$$ \begin{bmatrix} 1 & x & 1 \end{bmatrix} \begin{bmatrix} 1 & 3 & 2 \\ 2 & 5 & 1 \\ 15 & 3 & 2\\ \end{bmatrix} \begin{bmatrix} 1 \\ 2\\ x\\ \end{bmatrix} =0 $$
My attempt:
I divided on both side of the equation by $\begin{bmatrix}
1 & 3 & 2 \\
2 & 5 & 1 \\
15 & 3 & 2\\
\end{bmatrix}$ to get:
$$\frac{
\begin{bmatrix}
1 & x & 1
\end{bmatrix}
\begin{bmatrix}
1 & 3 & 2 \\
2 & 5 & 1 \\
15 & 3 & 2\\
\end{bmatrix}
\begin{bmatrix}
1 \\
2\\
x\\
\end{bmatrix}}{\begin{bmatrix}
1 & 3 & 2 \\
2 & 5 & 1 \\
15 & 3 & 2\\
\end{bmatrix}} = \frac{0}{\begin{bmatrix}
1 & 3 & 2 \\
2 & 5 & 1 \\
15 & 3 & 2\\
\end{bmatrix}}$$
Which gives:
$$ \begin{bmatrix}
1 & x & 1
\end{bmatrix}
\begin{bmatrix}
1 \\
2\\
x\\
\end{bmatrix} = 0$$
Solving this:
\begin{align}
(1)(1)+(2)(x)+(1)(x) &= 0 \\
1 + 2x + x &= 0 \\
3x &= -1 \\
x &= \frac{-1}{3} \\
\end{align}
... which is the incorrect answer. Is dividing both sides of the equation by the same value invalid in this case or how else is this approach wrong?
The closest thing to dividing by a matrix is multiplying from the left by its inverse. It is true that...
$$ M^{-1}\begin{bmatrix} 1 & x & 1 \end{bmatrix} M \begin{bmatrix} 1 \\ 2\\ x\\ \end{bmatrix} = M^{-1}0=0$$ But it is not valid to change the order of multiplication and conclude that...
$$\begin{bmatrix} 1 & x & 1 \end{bmatrix} M^{-1} M \begin{bmatrix} 1 \\ 2\\ x\\ \end{bmatrix}=0$$
as you seem to have done