If f is defined as $f(t,p) = g^{-1}(E(g(t,p))$ for any random variable $p$ and natural number $t$, what properties does $g$ have to have to make $f$ log concave in $t$?
Here, $g^{-1}$ is the inverse function of $g$. We can think of $p$ as a random variable that occurs at time $t$. And we want conditions on $g$ to be $f$ log concave in $t$.
