#include <bits/stdc++.h>
using namespace std;
using ll = long long;
#define ALL(v) v.begin(),v.end()
#define dbg(x) cerr << #x << ": " << (x) << endl;
template<class F, class S>
ostream& operator<<(ostream& os, pair<F,S>& p) {
    os << '(' << p.first << ',' << p.second << ')';
    return os;
}
template<class Iter>
void print(Iter beg, Iter end) {
    for (Iter itr = beg; itr != end; ++itr) {
        cerr << *itr << ' ';
    }
    cerr << '\n';
}
struct BIT {
    int n;
    vector<ll> d;
    BIT (int n) : n(n), d(n+1) {}

    void add(int idx, ll val) {
        ++idx;
        for (int x = idx; x <= n; x += x&-x) {
            d[x] += val;
        }
    }
    ll sum(int idx) {
        ll res = 0;
        for (int x = idx; x > 0; x -= x&-x) {
            res += d[x];
        }
        return res;
    }
    ll sum(int l, int r) {
        return sum(r) - sum(l);
    }
};
int n;
vector<int> x,y;
int main() {
    cin >> n;
    x.resize(n); y.resize(n);
    for (int i = 0; i < n; ++i) cin >> x[i] >> y[i];

    auto tmp = x;
    sort(ALL(tmp));
    tmp.erase(unique(ALL(tmp)), tmp.end());
    for (int i = 0; i < n; ++i) x[i] = lower_bound(ALL(tmp), x[i]) - tmp.begin();
    int xmax = tmp.size();
    tmp = y;
    sort(ALL(tmp));
    tmp.erase(unique(ALL(tmp)), tmp.end());
    for (int i = 0; i < n; ++i) y[i] = lower_bound(ALL(tmp), y[i]) - tmp.begin();
    int ymax = tmp.size();

    ll p=0, q=0, r=0, s=0;
    BIT bit(ymax);
    vector<vector<int>> ys(xmax);
    for (int i = 0; i < n; ++i) ys[x[i]].push_back(y[i]);
    for (int i = 0; i < xmax; ++i) {
        for (int j : ys[i]) {
            p += bit.sum(j);
            q += bit.sum(j+1, ymax);
        }
        for (int j : ys[i]) {
            bit.add(j, 1);
        }
    }
    vector<int> mx(xmax),my(ymax);
    for (int i = 0; i < n; ++i) {
        mx[x[i]]++;
        my[y[i]]++;
    }
    for (int i = 0; i < xmax; ++i) {
        r += (ll)mx[i] * (n - mx[i]);
    }
    for (int i = 0; i < ymax; ++i) {
        s += (ll)my[i] * (n - my[i]);
    }
    r >>= 1;
    s >>= 1;
    double ans = (p - q) / (sqrt(r) * sqrt(s));
    cout << fixed << setprecision(20) << ans << '\n';
}