Quicksort

Bob has written a fake QuickSort implementation as below. Can you figure out, if he randomly chooses a permutation p = [p1, p2, . . . , pn] obtained from 1, 2, . . . , n with equal probability, the expected number of inversions of p after calling QuickSort(p, 1, n, k)?

a fake QuickSort implementation
 1: function QuickSort(A, l, r, h)     ▷ Elements in A would be modified
 2:   if h > 1 and l < r then
 3:     m ← Partition(A, l, r)
 4:     QuickSort(A, l, m - 1, h - 1)
 5:     QuickSort(A, m + 1, r, h - 1)
 6: function Partition(A, l, r)        ▷ Elements in A would be modified
 7:   i ← l
 8:   j ← r
 9:   m ← ⌊(l+r)/2⌋ 
10:   pivot ← Am
11:   Am ← Ai
12:   while i < j do
13:     while i < j and Aj ≥ pivot do
14:       j ← j - 1
15:     if i < j then
16:       Ai ← Aj
17:     while i < j and Ai < pivot do
18:       i ← i + 1
19:     if i < j then
20:       Aj ← Ai
21:   Ai ← pivot
22:   return i

The number of inversions of a permutation [p1, p2, . . . , pn] is the number of integer pairs (u, v) such that 1 ≤ u < v ≤ n and pu > pv.

To avoid any precision issue, you are asked to report the product of n!, the factorial of n, and the expected number in modulo 998244353, which ought to be an integer.

The input contains several test cases. The first line contains an integer T indicating the number of test cases. The following describes all test cases. For each test case:

The only line contains two integers n and k.

For each test case, output a line containing “Case #x: y” (without quotes), where x is the test case number starting from 1, and y is the answer to this test case.

• 1 ≤ T ≤ 3 × 105
• 1 ≤ n, k ≤ 6000