#pragma GCC optimize ("O3")
#pragma GCC optimize ("unroll-loops")
//#pragma GCC target("sse,sse2,sse3,ssse3,sse4,popcnt,abm,mmx,avx,tune=native")
#include <bits/stdc++.h>
typedef long long ll;
typedef long double ld;
using namespace std;
const int N = 2e5 + 5;
int n, p, k, a[N];
pair<int, int> pos[N];
ll sum[N];
bool check(int i, int x) {
    ll cnt = sum[n - 1] * (x / n);
    if(i + x % n >= n)
        cnt += sum[n - 1] - sum[i - 1] + sum[(i + x % n) % n];
    else {
        cnt += sum[i + x % n];
        if(i)
            cnt -= sum[i - 1];
    }
    return cnt > p;
}
int main() {
    ios_base::sync_with_stdio(0);cin.tie(0);cout.tie(0);
    cin >> n >> p >> k;
    for(int i = 0;i < n;i++) {
        cin >> a[i];
        sum[i] = a[i];
        if(i)
            sum[i] += sum[i - 1];
    }
    for(int i = 0;i < n;i++) {
        int low = 0, high = 1e9, mid, ans;
        while(low <= high) {
            mid = low + high >> 1;
            if(check(i, mid))
                ans = mid, high = mid - 1;
            else
                low = mid + 1;
        }
        pos[i] = {(i + ans) % n, ans};
    }
    int cur = 0;
    int vis[n + 5];
    vector<pair<int, ll>> v;
    memset(vis, -1, sizeof vis);
    while(vis[cur] == -1) {
        v.push_back({cur, pos[cur].second});
        vis[cur] = v.size() - 1;
        cur = pos[cur].first;
    }
    for(int i = 1;i < v.size();i++)
        v[i].second += v[i - 1].second;
    if(k < v.size())
        return cout << v[k - 1].second / n, 0;
    k -= v.size();
    ll cycLen = v.size() - vis[cur];
    ll cycCnt = v.back().second;
    if(vis[cur])
        cycCnt -= v[vis[cur] - 1].second;
    ll ans = v.back().second + (k / cycLen) * cycCnt;
    int rem = k % cycLen;
    if(rem) {
        ans += v[vis[cur] + rem - 1].second;
        if(vis[cur])
            ans -= v[vis[cur] - 1].second;
    }
    cout << ans / n;
    return 0;
}