> [!definition]
>
> Let $\bracs{X_n}_0^\infty$ be a discrete-time, $\bracs{\cf_n}_0^\infty$-[[Filtration|adapted]] [[Stochastic Process|stochastic process]], and let $\bracs{C_n}_1^\infty$ be the *investment* into the process, such that in each unit of time, the profit/loss is $C_{n + 1}(X_{n + 1} - X_n)$.
>
> The process $\bracs{C_n}_1^\infty$ is $\bracs{\cf_n}_0^\infty$-**previsible** if $C_{n + 1} \in L^1(\Omega, \cf_n, \bp)$ for all $n \ge 0$. In other words, future investments are chosen based on past observations, in which case
> $
> \begin{align*}
> \Delta_{n + 1} &= \ev\braks{C_{n + 1}(X_{n + 1} - X_{n})} \\
> &= \int \ev\bracs{C_{n + 1}(X_{n + 1} - X_{n})|\cf_n} \\
> &= \ev(C_{n + 1}\ev\bracs{X_{n + 1}|\cf_n}) - \ev(C_{n + 1}X_n)
> \end{align*}
> $