The podium for the best programmers. CC BY-SA 4.0 by Santeri Viinamäki on Wikimedia Commons

Audrey is playing a game ranking her favourite $$$100$$$ competitive programmers in a top $$$10$$$ list. She already has objectively ranked all $$$100$$$ of them. Now, one by one, she is given $$$10$$$ distinct programmers uniformly at random, and must rank them among each other, in a relative top $$$10$$$ list. Once she has made a choice, she is not allowed to move them any more, and she must stick to her choice. She cannot bear her assignments list being incorrect, so she tries to give a correct list as often as possible. Of course, she knows that it is impossible to always be right, so she is satisfied if she is correct at least $$$1\%$$$ of the time. She is not allowed to assign the same spot to two different people. Help Audrey fill in her top $$$10$$$ lists.

This is an interactive problem. Your submission will be run against an interactor, which reads from the standard output of your submission and writes to the standard input of your submission. This interaction needs to follow a specific protocol:

Each game, exactly $$$10$$$ times, the interactor will send one line with an integer $$$n$$$ ($$$1\leq n \leq{100}$$$), the objective rank Audrey assigned to a competitive programmer. Then, your program should print one line with an integer $$$r$$$ ($$$1 \leq r \leq 10$$$), the spot in the relative ranking Audrey should assign this programmer to.

Writing the same value of $$$r$$$ twice in the same game will result in a wrong answer.

The $$$10$$$ objective ranks given in each game are pairwise distinct.

After each game, the interactor will print one line with either "win" or "loss", depending on whether your list is correct or not. Note that, in each game, exactly $$$10$$$ programmers need to be assigned a relative rank, even when it is clear that your list can no longer be correct.

Your submission will be run on exactly one test case, in which the interactor will play exactly $$${10\,000}$$$ games. Winning fewer than $$${100}$$$ of these games will result in a wrong answer. Note that the sample interaction on the next page shows only two games for illustration purposes only.

The interactor is not adaptive: for each of your submissions, the $$$10$$$ objective ranks of each independent game are uniformly randomly generated, and do not depend on your queries.

Make sure you flush the buffer after each write.

A testing tool is provided to help you develop your solution. Read the comments in this testing tool for explanation on how to use it.

Sample interaction

<1
>1
<2
>2
<3
>3
<4
>4
<5
>5
<6
>6
<7
>7
<8
>8
<9
>9
<10
>10
<win
<26
>3
<3
>1
<8
>2
<50
>5
<46
>4
<52
>6
<92
>10
<95
>9
<84
>8
<42
>7
<loss