import sys

sys.setrecursionlimit(200005)
int1 = lambda x: int(x)-1
p2D = lambda x: print(*x, sep="\n")
def II(): return int(sys.stdin.buffer.readline())
def LI(): return list(map(int, sys.stdin.buffer.readline().split()))
def LI1(): return list(map(int1, sys.stdin.buffer.readline().split()))
def LLI(rows_number): return [LI() for _ in range(rows_number)]
def LLI1(rows_number): return [LI1() for _ in range(rows_number)]
def BI(): return sys.stdin.buffer.readline().rstrip()
def SI(): return sys.stdin.buffer.readline().rstrip().decode()
dij = [(0, 1), (-1, 0), (0, -1), (1, 0)]
# dij = [(0, 1), (-1, 0), (0, -1), (1, 0), (1, 1), (1, -1), (-1, 1), (-1, -1)]
inf = 10**16
# md = 998244353
# md = 10**9+7

from functools import lru_cache

# nullさんの取ったカードがs、kiriさんの取ったカードがtのときに、nullさんの勝つ確率
@lru_cache(None)
def p(s, t):
    used = s | t
    # 全部取り終わった場合
    if used == (1 << n)-1:
        x = sum(a for i, a in enumerate(aa) if s >> i & 1)
        y = tot-x
        return x > y
    res = 0
    for i, ai in enumerate(aa):
        if used >> i & 1: continue
        s1 = s | 1 << i
        p1 = 1
        if popcnt(used) == n-1:
            p1 = p(s | 1 << i, t)
        else:
            for j, aj in enumerate(aa):
                if j == i: continue
                if used >> j & 1: continue
                p1 = min(p1, p(s1, t | 1 << j)/aj+p(s1, t)*(aj-1)/aj)
        cur = 1
        for j, aj in enumerate(aa):
            if used >> j & 1: continue
            p2 = p(s, t | 1 << j)
            cur = min(cur, (aj*p1-p2+ai*p2)/(ai+aj-1))
        res = max(res, cur)
    return res

popcnt = lambda x: bin(x).count("1")
n = II()
aa = LI()
tot = sum(aa)
print(p(0, 0))