#include <bits/stdc++.h>
#pragma GCC optimize("unroll-loops")
using namespace std;
const int mod=998244353;
#define rep(i,a,b) for (int i=int(a);i<int(b);i++)
#include<atcoder/all>
using namespace atcoder;
using mint=modint998244353;


int main() {
    ios::sync_with_stdio(false);
    cin.tie(nullptr);
    
    int N,K;
	cin>>N>>K;
	vector<int> A(N);
	rep(i,0,N) cin>>A[i];
	sort(A.begin(),A.end());
	vector<mint> D(N);
	rep(i,0,N) D[i]=A[i];
	for(int i=N-1;i>0;i--) D[i]-=D[i-1];
	vector<mint> fact(N+2,1),fact_inv(N+2,1);
	rep(i,0,N+1) fact[i+1]=fact[i]*(i+1);
	fact_inv[N+1]=fact[N+1].inv();
	for(int i=N+1;i>0;i--) fact_inv[i-1]=fact_inv[i]*i;
	auto P=[&](int a,int b) -> mint {
		if(a<0||a<b) return 0;
		return fact[a]*fact_inv[a-b];
	};
	auto C=[&](int a,int b) -> mint {
		if(a<0||a<b) return 0;
		return fact[a]*fact_inv[b]*fact_inv[a-b];
	};
	auto f=[&](int a) -> mint {
		mint res=0;
		rep(j,0,N){
			res+=P(a,j)*D[j]*fact[N-j];
		}
		return res;
	};
	// f(l)+f(l+1)+...+f(r)
	auto g=[&](int l,int r) -> mint {
		mint res=0;
		rep(k,0,N){
			res+=D[k]*fact[N-k]*fact[k]*(C(r+1,k+1)-C(l,k+1));
		}
		return res;
	};
	mint ans=0;
	map<int,int> m;
	int L=N+1,R=-1;
	rep(i,1,N){
		m[N-min(i,K)-min(N-i,K)]++;
	}
	for(auto x:m){
		if(x.second!=2) ans+=f(x.first)*x.second;
		else{
			L=min(L,x.first);
			R=max(R,x.first);
		}
	}
	if(L<=R) ans+=g(L,R)*2;
	cout<<ans.val()<<"\n";
}