1: \begin{proof}

2:  From (\ref{eq:outage_function}), for $\lambda>1$, we have 

3: \begin{eqnarray}

4: %\array{l}

5: F(\Delta,\ell)=\Pr\left\{a>\lambda\delta, b_1<\frac{\delta (a+1)}{(a-\delta)}, \dots, b_\ell< \frac{\delta (a+1)}{(a-\delta)} \right \}\leq \left(\Pr\left\{b_1<\frac{\lambda\delta+1}{\lambda-1}\right\}\right)^\ell

6: %\endarray

7: \end{eqnarray}

8: since $(a-\delta)>0$  and $\frac{a+1}{a-\delta}$ is decreasing  w.r.t. $a$ when $a>\delta$.  Apparently, $F(\Delta,\ell)$ can attain the same order of $\rho^{-\ell}$. As for the case of $\lambda<1$, due to the fact that $\frac{ab}{a+b+1}\leq \min [a,b]\leq a$,  $F(\Delta,\ell)$ is lower bounded by $\Pr\{\delta>a>\lambda\delta\}$ whose order is equal to 1, and thus suffers from  diversity losses. A rigorous proof is provided in Appendix\ref{appendix:F_out_function_diversity_order}.

9: \end{proof}