Perform a double loop using the following basic structure:

for $($i in $1$:length$($th$))\{$

$...$

for $($j in $1$:B$)\{$

$...$

$\}$

$...$

$\}$

$($a$)$ At the i$-$th iteration of the outer loop you should

$$ define vectors th$.$ML$=$th$.$Bayes$=0$ in which to store your simulated estimates;

$$ perform the inner loop;

$$ store the average squared error for each vector of estimates in the i$-$th element of the

corresponding AVSE.$-$vector.

$($b$)$ At the j$-$th iteration of the inner loop you should

$$ generate a simulated version of X using th$[$i$]$ as the success probability;

$$ compute each version of the two estimators based on that simulated X and store them

in the corresponding th$.-$vector.