#include<bits/stdc++.h>
#pragma optimize ("g",on)
#pragma GCC optimize ("inline")
#pragma GCC optimize ("Ofast")
#pragma GCC optimize ("unroll-loops")
#pragma GCC optimize ("03")
#pragma GCC target ("sse,sse2,sse3,ssse3,sse4,popcnt,abm,avx2,mmx,fma,avx,tune=native")
#pragma comment(linker, "/stack:200000000")
//01001101 01100001 01101011 01101000 01100001 01100111 01100001 01111001
using namespace std;
#define all(a) a.begin(), a.end()
#define rall(a) a.rbegin(), a.rend()
#define ll long long
#define pb push_back
#define sz(a) a.size()
#define nl '\n'
#define popb pop_back()
#define ld double
#define ull unsigned long long
#define ff first
#define ss second
#define fix fixed << setprecision
#define pii pair<int, int>
#define E exit (0)
#define int long long
const int inf = 1e18, N = 5e5 + 1, mod = 998244353;
vector<pii> dir = {{1, 0}, {0, 1}, {-1, 0}, {0, -1}};
main() {
//freopen("cowrect.in", "r", stdin);
//freopen("cowrect.out", "w", stdout);
ios_base::sync_with_stdio(NULL);
cin.tie(NULL);
int n, ans = inf;
cin >> n;
int a[n];
for (int i = 0; i < n; i++) cin >> a[i];
for (int mask = 0; mask < (1 << (n - 1)); mask++) {
vector<pii> s;
vector<int> h;
int l = 0, cnt = 0, is = 1;
for (int i = 1; i < n; i++) {
if (mask >> (i - 1) & 1) {
s.pb({l, i - 1});
l = i;
cnt++;
continue;
}
if (a[i] < a[i - 1]) is = 0;
}
s.pb({l, n - 1});
cnt++;
for (auto e: s) {
if (h.empty()) {
for (int i = e.ff; i <= e.ss; i++) h.pb(a[i]);
continue;
}
if (h[0] >= a[e.ss]) {
for (int i = e.ss; i >= e.ff; i--) h.insert(h.begin(), a[i]);
continue;
}
if (h.back() <= a[e.ff]) {
for (int i = e.ff; i <= e.ss; i++) h.pb(a[i]);
continue;
}
bool ok = 0;
for (int i = 0; i < sz(h) - 1; i++) {
if (h[i] <= a[e.ff] && a[e.ss] <= h[i + 1]) {
for (int j = e.ss; j >= e.ff; j--) h.insert(h.begin() + i + 1, a[j]);
ok = 1;
break;
}
}
if (!ok) is = 0;
}
if (is) ans = min(ans, cnt);
}
cout << ans;
}
