#include <map>
#include <tuple>
#include <cstdio>
#include <vector>
#include <algorithm>
using namespace std;
int n, a[500006], dist[500006], res;
vector<int> playgrounds;
int pv[500006], nx[500006];
map<tuple<int, int, int, int, int>, int> mem;
int f(int x, int y, int z, int w, int p) {
if (y > 2 * n - 2) return 0;
if (mem.find({ x, y, z, w, p }) != mem.end()) return mem[{ x, y, z, w, p }];
if (w && a[x] == -1) return f(x, y, z, 0, p);
int res = 0;
if (w && a[x] != -1) res = max(res, f(x + 1, y - 1, min(n, max(a[x + 1], 0)), 1, p));
if (x < n - 1) res = max(res, f(x + 1, y + 1, min(n, max(a[x + 1], 0)), 0, p));
if (pv[x] != -1 && z > 0 && y <= 2 * x) {
int dist = x - pv[x];
int N = min(z, (2 * x - y) / (2 * dist) + 1);
res = max(res, f(x, y + 2 * dist * N, z - N, 0, p) + N);
if (!p) for (int i = 1; i <= N; i++) res = max(res, f(x, y + 2 * dist * i, z - i, 0, 1) + i);
}
if (nx[x] != -1 && z > 0 && y <= 2 * x) {
int dist = nx[x] - x;
int N = min(z, (2 * x - y) / (2 * dist) + 1);
res = max(res, f(x, y + 2 * dist * N, z - N, 1, p) + N);
if (!p) for (int i = 1; i <= N; i++) res = max(res, f(x, y + 2 * dist * i, z - i, 0, 1) + i);
}
return mem[{ x, y, z, w, p }] = res;
}
int main() {
scanf("%d", &n);
long long sum = 0;
for (int i = 0; i < n; i++) {
scanf("%d", a + i);
sum += max(a[i], 0);
dist[i] = (int)1e9;
if (a[i] < 0) playgrounds.push_back(i);
}
for (int i = 0; i < n; i++) {
int it = lower_bound(playgrounds.begin(), playgrounds.end(), i) - playgrounds.begin();
if (it == (int)playgrounds.size()) nx[i] = -1;
else nx[i] = playgrounds[it];
it = upper_bound(playgrounds.begin(), playgrounds.end(), i) - playgrounds.begin() - 1;
if (it == -1) pv[i] = -1;
else pv[i] = playgrounds[it];
}
{
int tmp = (int)1e9;
for (int i = 0; i < n; i++) {
if (a[i] == -1) tmp = 0;
dist[i] = min(dist[i], tmp);
tmp++;
}
tmp = (int)1e9;
for (int i = n - 1; i >= 0; i--) {
if (a[i] == -1) tmp = 0;
dist[i] = min(dist[i], tmp);
tmp++;
}
}
printf("%lld", sum - f(0, 0, min(n, max(a[0], 0)), 0, 0));
}
