Given the standard-Arcsin probabilty function, called $Arcsin(0,1)$, looking like this:
$ F_X(x) = \left\{ \begin{matrix} 0 && x < 0\\ \frac{2}{\pi}\arcsin(\sqrt{x}) && 0 \le x \le 1\\ 1 && x > 1 \end{matrix} \right. $
Now say that $W$ is a uniform distribution over the interval $[0, \pi]$. We do a translation called $V$, and this is defined as the following: $V = -\cos(W)$ I have to proof that this is an $Arcsin(-1, 1)$ distribution. To dit this, I would do this:
Let's say that $g(x) = -\cos(x)$, and this is our transformation function.
$f_V(y) = f_W(g^{-1}(y))$
Now we only have to calculate our inverse function right? But when I do this, I don't come near the given answer. $y = -\cos\left({\frac{x}{\pi}}\right) = -\left(\cos\left(2 \cdot \frac{x}{2\pi}\right)\right) = 2\sin^2\left(\frac{x}{2\pi}\right)-1 $
$ y + 1 = 2\sin^2\left( \frac{x}{2\pi} \right) $
$ \sqrt{\frac{y+1}{2}} = \sin\left(\frac{x}{2\pi}\right) $
$ \arcsin\left(\sqrt{\frac{y+1}{2}}\right) = \frac{ x}{2\pi} $
$ 2\pi \arcsin\left(\sqrt{\frac{y+1}{2}}\right) = x $
The resulting function is a function $(-1,1) \rightarrow [0,10]$ So it's a factor 10 too much. My question is, what steps are missing? Or is the question just wrong?