[step:Apply the simple Markov property at each $s_n$ and pass to the [limit](/page/Limit)]
Fix $A \in \mathcal{F}_s^+$ and let $F : \mathbb{R}^{dk} \to \mathbb{R}$ be bounded and continuous. Since $\mathcal{F}_s^+ = \bigcap_{t > s} \mathcal{F}_t^B$ and $s_n > s$, we have $A \in \mathcal{F}_{s_n}^B$ for every $n$. By the [Simple Markov Property](/theorems/1175)(iii), the process $(B_{t + s_n} - B_{s_n})_{t \geq 0}$ is independent of $\mathcal{F}_{s_n}^B$, so
\begin{align*}
\mathbb{E}\bigl[F(B_{t_1 + s_n} - B_{s_n}, \ldots, B_{t_k + s_n} - B_{s_n}) \, \mathbb{1}_A\bigr] = \mathbb{E}\bigl[F(B_{t_1 + s_n} - B_{s_n}, \ldots, B_{t_k + s_n} - B_{s_n})\bigr] \cdot \mathbb{P}(A).
\end{align*}
The integrand on the left is bounded by $\|F\|_\infty$ and converges almost surely to $F(B_{t_1+s} - B_s, \ldots, B_{t_k+s} - B_s) \, \mathbb{1}_A$. By the [Dominated Convergence Theorem](/theorems/4), the left-hand side converges to $\mathbb{E}[F(B_{t_1+s} - B_s, \ldots, B_{t_k+s} - B_s) \, \mathbb{1}_A]$. The same argument (without $\mathbb{1}_A$) shows the first factor on the right converges to $\mathbb{E}[F(B_{t_1+s} - B_s, \ldots, B_{t_k+s} - B_s)]$. Therefore
\begin{align*}
\mathbb{E}\bigl[F(B_{t_1+s} - B_s, \ldots, B_{t_k+s} - B_s) \, \mathbb{1}_A\bigr] = \mathbb{E}\bigl[F(B_{t_1+s} - B_s, \ldots, B_{t_k+s} - B_s)\bigr] \cdot \mathbb{P}(A).
\end{align*}
Since this holds for every bounded continuous $F$ and every $A \in \mathcal{F}_s^+$, the process $(B_{t+s} - B_s)_{t \geq 0}$ is independent of $\mathcal{F}_s^+$.
[/step]