Consider as probability matrix a matrix $M \in [0,1]^{n \times n}$ while every row sums up to $1$.
Statement: Consider a $2\times 2$ probability matrix $M' \in [0,1]^{2 \times 2}$. Show, that the following holds: $$\exists \text{ probability matrix } M : M^2 = M' \iff \operatorname{Trace}(M') \geq 1$$
I showed $\implies$, i.e. when $M'$ is square of another probability matrix, then one can show, that the trace must be greater or equal to one.
But I got stuck at the $\impliedby$ part. Any tipps?
So far considering user68061's answer: Let $M' = \begin{pmatrix} x & (1-x) \\ (1-y) & y \end{pmatrix}$ and $M = \begin{pmatrix} a & (1-a) \\ (1-b) & b \end{pmatrix}$, then $M^2 = \begin{pmatrix} a^2+(1-a)(1-b) & a(1-a)+b(1-a) \\ a(1-b)+b(1-b) & b^2+(1-a)(1-b)\end{pmatrix}$.
Hence we need to solve
$x = a^2+(1-a)(1-b),\ y=b^2+(1-a)(1-b)$ where $x+y \geq 1$.
But somehow I think I am missing the trickery in calculation to get this solution..can you help?