Analyze the complexity (in big-O terms) of the following
selection-sort routine. The routine picks out the largest element in
the current list and exchanges it with the last element. It continues
until the list has only one element:

for (i = n-1; i > 0; i--) {
        MaxPosition = i;
        for (j = 0; j < i; j++) {
                if (a[j] > a[MaxPosition])
                        MaxPosition = j;
        }
        exchange(i, MaxPosition);
}

Shouldn't this be the same order as a bubble sort, as it is basically
a reversed bubble sort?  Bubble sort has a worst case and average case
O(n)^2

There are several analyses of bubble sort that you can find by
searching google for "Bubble Sort Complexity"