#include <bits/stdc++.h>
using namespace std;

using vi = vector<int>; using vvi = vector<vi>; using vvvi = vector<vvi>;
using ll = long long int;
using vll = vector<ll>; using vvll = vector<vll>; using vvvll = vector<vvll>;
using vd = vector<double>; using vvd = vector<vd>; using vvvd = vector<vvd>;
using P = pair<int, int>;
using Pll = pair<ll, ll>;
using cdouble = complex<double>;

const double INFD = numeric_limits<double>::infinity();
const double PI = 3.14159265358979323846;
#define Loop(i, n) for(int i = 0; i < (int)(n); i++)
#define Loop1(i, n) for(int i = 1; i <= (int)(n); i++)
#define Loopr(i, n) for(int i = (int)(n) - 1; i >= 0; i--)
#define Loopr1(i, n) for(int i = (int)(n); i >= 1; i--)
#define Foreach(buf, container) for(auto buf : container)
#define Loopdiag(i, j, h, w, sum) for(int i = ((sum) >= (h) ? (h) - 1 : (sum)), j = (sum) - i; i >= 0 && j < (w); i--, j++)
#define Loopdiagr(i, j, h, w, sum) for(int j = ((sum) >= (w) ? (w) - 1 : (sum)), i = (sum) - j; j >= 0 && i < (h); j--, i++)
#define Loopdiagsym(i, j, h, w, gap) for (int i = ((gap) >= 0 ? (gap) : 0), j = i - (gap); i < (h) && j < (w); i++, j++)
#define Loopdiagsymr(i, j, h, w, gap) for (int i = ((gap) > (h) - (w) - 1 ? (h) - 1 : (w) - 1 + (gap)), j = i - (gap); i >= 0 && j >= 0; i--, j--)
#define Loopitr(itr, container) for(auto itr = container.begin(); itr != container.end(); itr++)
#define printv(vector) Loop(ex_i, vector.size()) { cout << vector[ex_i] << " "; } cout << endl;
#define printmx(matrix) Loop(ex_i, matrix.size()) { Loop(ex_j, matrix[ex_i].size()) { cout << matrix[ex_i][ex_j] << " "; } cout << endl; }
#define quickio() ios::sync_with_stdio(false); cin.tie(0);
#define bitmanip(m,val) static_cast<bitset<(int)m>>(val)
#define Comp(type_t) bool operator<(const type_t &another) const
#define fst first
#define snd second
bool nearlyeq(double x, double y) { return abs(x - y) < 1e-9; }
bool inrange(int x, int t) { return x >= 0 && x < t; }
bool inrange(vi xs, int t) { Foreach(x, xs) if (!(x >= 0 && x < t)) return false; return true; }
ll rndf(double x) { return (ll)(x + (x >= 0 ? 0.5 : -0.5)); }
ll floorsqrt(ll x) { ll m = (ll)sqrt((double)x); return m + (m * m <= x ? 0 : -1); }
ll ceilsqrt(ll x) { ll m = (ll)sqrt((double)x); return m + (x <= m * m ? 0 : 1); }
ll rnddiv(ll a, ll b) { return (a / b + (a % b * 2 >= b ? 1 : 0)); }
ll ceildiv(ll a, ll b) { return (a / b + (a % b == 0 ? 0 : 1)); }
ll gcd(ll m, ll n) { if (n == 0) return m; else return gcd(n, m % n); }
ll lcm(ll m, ll n) { return m * n / gcd(m, n); }

/*******************************************************/

vvd solve_p_table(int n, vi &a, double pa) {
	vvd ret(n, vd(n));
	vd ps(1 << n);
	vi done(1 << n);
	queue<P> que;
	ps[(1 << n) - 1] = 1;
	que.push({ 0, (1 << n) - 1 });
	done[(1 << n) - 1] = 1;
	while (que.size()) {
		int turn = que.front().fst;
		int mask = que.front().snd;
		que.pop();
		int minv = INT_MAX, minid = 0, cnt = 0;
		Loop(i, n) {
			if ((1 << i) & mask) {
				cnt++;
				if (a[i] < minv) {
					minv = a[i];
					minid = i;
				}
			}
		}
		Loop(i, n) {
			if ((1 << i) & mask) {
				if (i == minid) {
					if (cnt > 1) {
						ps[mask ^ (1 << i)] += ps[mask] * pa;
						ret[turn][i] += ps[mask] * pa;
					}
					else if (cnt == 1) {
						ps[mask ^ (1 << i)] += ps[mask];
						ret[turn][i] += ps[mask];
					}
				}
				else {
					if (cnt > 1) {
						ps[mask ^ (1 << i)] += ps[mask] * double(1 - pa) / (cnt - 1);
						ret[turn][i] += ps[mask] * double(1 - pa) / (cnt - 1);
					}
				}
				if (done[mask ^ (1 << i)] == 0) {
					que.push({ turn + 1, mask ^ (1 << i) });
					done[mask ^ (1 << i)] = 1;
				}
			}
		}
	}
	return ret;
}

double solve(int n, vvd &pa, vvd &pb, vi &a, vi &b) {
	vd ret(n);
	Loop(i, n) {
		Loop(ja, n) {
			Loop(jb, n) {
				if (a[ja] > b[jb]) {
					ret[i] += (a[ja] + b[jb]) * pa[i][ja] * pb[i][jb];
				}
			}
		}
	}
	return accumulate(ret.begin(), ret.end(), 0.0);
}

int main() {
	int n; cin >> n;
	double pa, pb; cin >> pa >> pb;
	vi a(n), b(n);
	Loop(i, n) cin >> a[i];
	Loop(i, n) cin >> b[i];
	vvd table_a = solve_p_table(n, a, pa);
	vvd table_b = solve_p_table(n, b, pb);
	cout << setprecision(12) << solve(n, table_a, table_b, a, b) << endl;
}