You discovered that the problems in the contest were shuffled, so you want to analyze their difficulty.
You assume that the difficulty of each of the $$$n$$$ problems is described by a permutation $$$p = (p_1, \cdots , p_n)$$$ (a permutation of length $$$n$$$ contains each of the numbers from $$$1$$$ to $$$n$$$ exactly once).
For each number $$$k$$$, you want to determine whether the $$$k$$$ easiest problems (numbered $$$1, \cdots , k$$$) appear in the permutation strictly earlier than the $$$k$$$ hardest problems (numbered $$$n, \cdots , n-k+1$$$).
Let the indices of the $$$k$$$ easiest problems be given by the set $$$A_k$$$, and the indices of the $$$k$$$ hardest problems by the set $$$B_k$$$.
Then, for a given number $$$k$$$, it is necessary to check that for any $$$a \in A_k, b \in B_k$$$, the inequality $$$a \lt b$$$ holds.
You need to find the maximum possible $$$k \ge 1$$$ for which the corresponding sets $$$A_k, B_k$$$ satisfy the condition above, or report that no such $$$k$$$ exists.
The first line contains an integer $$$n (1 \le n \le 10^5)$$$ — the number of problems in the contest.
The second line contains $$$n$$$ numbers $$$p_1, p_2, \ldots, p_n\, (1 \le p_i \le n)$$$ — a permutation describing the difficulty of the problems. It is guaranteed that each of the numbers from $$$1$$$ to $$$n$$$ appears exactly once.
In the only line, output the maximum possible $$$k \ge 1$$$ satisfying the condition of the problem.
If no such $$$k$$$ exists, output $$$-1$$$.
71 4 2 3 5 7 6
3
33 2 1
-1