PROGRAMMING:Zzu Alliance
On November 3, 2018, the finals of the League of heroes S8 World Championship opened in Incheon, South Korea. For the first time in eight years, two teams in the same group competed for the championship in the finals. The match between Ig and FNC attracted players' expectation. Finally, the S8 world champion was born! Ig3:0 defeat FNC and win! China lol wins the world champion!
The members of ACM-ICPC Lab (* *) who are interested in programming and algorithms are welcome to add QQ group number: 562888278 * *) are very excited after learning the news, and decide to develop a game of their own - zzu alliance. Little J is responsible for the design of MVP evaluation module after each game. How to select the MVP on the spot according to the performance of each player in the game? It's difficult for him. I need your help. In a real lol game, the selection of MVP involves damage output, ratio of damage to death, average money per minute, 10% damage to tower, control of dragons and monsters, 2.5% control time to enemy, skill dodge and hit, multiple kill, continuous kill, team outfit and team contribution (Group combat efficiency ratio), support efficiency (group participation rate), and the selection of MVP There are many factors such as hero position and proficiency expectation (game default), kDa (kill, death, assist) and so on. In its first version, the zzu alliance decided not to consider the complexity, and adopted the following selection criteria:
(1) Damage output DPS (0 = < DPS < = 50000)
(2) Accumulated money m (0 = < m < = 30000)
(3) Damage to tower, tow (0 = < tow < = 10000)
(4) Support rate s (0 = < s < = 100)
(5) Kill K, death D, assist a (0 = < K, D, a < = 30)
Where each value is an integer.
Finally, there is an MVP candidate value for each player (* * double precision real number * *)
V = Dps/50000 + M / 30000 + Tow / 10000 + S / 100 + (K + A) / (D+1）
The MVP winner is the player with the largest V value in the winner, and the data ensures that the MVP of each game is unique.
###Input format:
An integer t (1 = < T < = 200) in the first line indicates that there are t groups of test data in the next row
For each group of test data,
The first line, win / lose, means the blue army won / lost
The next five lines are the post game index statistics of five players, and the format of each line is
Player ID DPS m tow s k Da
The next action is win / lose, which means the purple army won / lost
The next five lines are also the records of five players in the purple legion, and the format is the same
Player ID DPS m tow s k d a
###Output format:
T line, for each group of test data, output the game MVP player ID
###Input example:
```in
two
win
zxy 23197 13453 8636 61 6 3 3
fxr 12155 9854 8769 57 2 4 10
sgl 21232 23717 5215 67 24 23 25
cy 8268 201 3151 21 0 27 8
lzh 22306 15051 6246 54 22 28 5
lose
chj 16800 14340 4185 90 19 20 23
cwc 2646 1759 6548 60 0 7 5
wjk 11467 951 9615 52 4 16 11
lmc 29038 11256 173 64 8 3 3
lx 4686 11291 2677 77 16 26 22
lose
lzl 7670 14713 4174 42 6 23 7
gy 13955 5514 5087 91 15 29 4
sq 253 4773 7454 13 9 6 1
gxb 32055 7682 6093 83 16 7 26
gxh 24687 12611 3652 70 10 25 29
win
zyj 5074 12640 1414 64 29 8 6
hzy 23484 20429 6293 54 22 16 2
hzp 14823 20893 448 58 9 9 10
sjf 13059 6451 2344 72 5 13 17
tjt 26347 13267 5171 46 6 23 15
```
###Output example:
```out
zxy
zyj
```
**Note: due to the complexity of obtaining the real statistical data of the game, the above data may not be completely consistent with the real situation, but it does not affect the problem solving**
answer：If there is no answer, please comment
The members of ACM-ICPC Lab (* *) who are interested in programming and algorithms are welcome to add QQ group number: 562888278 * *) are very excited after learning the news, and decide to develop a game of their own - zzu alliance. Little J is responsible for the design of MVP evaluation module after each game. How to select the MVP on the spot according to the performance of each player in the game? It's difficult for him. I need your help. In a real lol game, the selection of MVP involves damage output, ratio of damage to death, average money per minute, 10% damage to tower, control of dragons and monsters, 2.5% control time to enemy, skill dodge and hit, multiple kill, continuous kill, team outfit and team contribution (Group combat efficiency ratio), support efficiency (group participation rate), and the selection of MVP There are many factors such as hero position and proficiency expectation (game default), kDa (kill, death, assist) and so on. In its first version, the zzu alliance decided not to consider the complexity, and adopted the following selection criteria:
(1) Damage output DPS (0 = < DPS < = 50000)
(2) Accumulated money m (0 = < m < = 30000)
(3) Damage to tower, tow (0 = < tow < = 10000)
(4) Support rate s (0 = < s < = 100)
(5) Kill K, death D, assist a (0 = < K, D, a < = 30)
Where each value is an integer.
Finally, there is an MVP candidate value for each player (* * double precision real number * *)
V = Dps/50000 + M / 30000 + Tow / 10000 + S / 100 + (K + A) / (D+1）
The MVP winner is the player with the largest V value in the winner, and the data ensures that the MVP of each game is unique.
###Input format:
An integer t (1 = < T < = 200) in the first line indicates that there are t groups of test data in the next row
For each group of test data,
The first line, win / lose, means the blue army won / lost
The next five lines are the post game index statistics of five players, and the format of each line is
Player ID DPS m tow s k Da
The next action is win / lose, which means the purple army won / lost
The next five lines are also the records of five players in the purple legion, and the format is the same
Player ID DPS m tow s k d a
###Output format:
T line, for each group of test data, output the game MVP player ID
###Input example:
```in
two
win
zxy 23197 13453 8636 61 6 3 3
fxr 12155 9854 8769 57 2 4 10
sgl 21232 23717 5215 67 24 23 25
cy 8268 201 3151 21 0 27 8
lzh 22306 15051 6246 54 22 28 5
lose
chj 16800 14340 4185 90 19 20 23
cwc 2646 1759 6548 60 0 7 5
wjk 11467 951 9615 52 4 16 11
lmc 29038 11256 173 64 8 3 3
lx 4686 11291 2677 77 16 26 22
lose
lzl 7670 14713 4174 42 6 23 7
gy 13955 5514 5087 91 15 29 4
sq 253 4773 7454 13 9 6 1
gxb 32055 7682 6093 83 16 7 26
gxh 24687 12611 3652 70 10 25 29
win
zyj 5074 12640 1414 64 29 8 6
hzy 23484 20429 6293 54 22 16 2
hzp 14823 20893 448 58 9 9 10
sjf 13059 6451 2344 72 5 13 17
tjt 26347 13267 5171 46 6 23 15
```
###Output example:
```out
zxy
zyj
```
**Note: due to the complexity of obtaining the real statistical data of the game, the above data may not be completely consistent with the real situation, but it does not affect the problem solving**
answer：If there is no answer, please comment