Description

Suppose the effectiveness rate of cocktail therapy is $x$, and the effectiveness rate of a new therapy is $y$. If $y-x$ is greater than $5\%$, the new therapy is better; if $x-y$ is greater than $5\%$, the new therapy is worse; otherwise, they are considered to have similar effectiveness.
Given $n$ groups of clinical control experiments, where the first group uses cocktail therapy and the other $n-1$ groups use various improved therapies, write a program to determine the effectiveness of each improved therapy.

Input Format

The first line contains an integer $n$ ($1<n≤20$);
The next $n$ lines each contain two integers: the first integer is the total number of clinical cases (not exceeding $10000$), and the second integer is the number of effective cases.
Among these $n$ lines of data, the first line contains data for cocktail therapy, and the remaining lines contain data for various improved therapies.

Output Format

Output $n-1$ lines, each indicating the effectiveness of the corresponding improved therapy:
If the effect is better, output "better";
If the effect is worse, output "worse";
Otherwise, output "same".

Sample

5
125 99
112 89
145 99
99 97
123 98

same
worse
better
same