6
For each subproblem: print the year, the subproblem and the number of different accepted submissions.
Use hashes to decide whether two submissions are different or not.
Sort results by year and for the same year lexicographically by subproblem.
Find the 20 longest accepted submissions. (You may assume that there is not a tie for the 20th place.)
Find the set of teams that made these submissions.
Note that this set may contain less than 20 teams.
Print the team names in ascending order. (Each team just once, even if they had more than one submission.)
The "fishiness" of a pair of teams (T1, T2) is defined as the number of different subproblems for which
the two teams received OK within 1 minute of each other (inclusive).
Count the number of pairs of teams with fishiness at least 4.
Note 1: The two teams must be different. We do not count pairs (x,x).
Note 2: The pair (x,y) is the same pair as (y,x), count it just once, not twice.
Note 3: Don't forget to first print the number of results (which is 1) and only then the correct count.
Given a set of problems, the average success rate for that set of problems can be defined
as the number of accepted submissions divided by the number of all submissions for those problems.
For each author: print the author and the average success rate of problems he or she
(co-)authored. Round success rates to exactly 3 decimal places.
Sort results in the order of decreasing success rate.
(There will be no ties.)
Recall that the penalty for a solved subproblem is calculated as "T + 10 * R * D", where
T is the time in minutes (ignoring seconds) from the start of the contest (i.e. for 13:03:47, T=63 minutes),
R is the number of previous rejected (WA) submissions and
D is the difficulty of the subproblem (D is either 1 or 2).
For each subproblem with at least one successful submission, calculate the average penalty
associated with a successful submission.
Print the year, the subproblem (e.g., G1) and the average penalty over all successful submissions.
Round the average penalty to two decimal places. Sort by year and then by subproblem.
Do not print the subproblems that have no successful submissions.
We say that a contestant (as identified by his/her name and his/her team's TIS) is
happy with problem X if and only if his/her team does have an accepted submission for at least one subproblem
of X and his/her name starts with the letter of the task.
(For example, Bob is happy if his team solves at least one of B1 and B2.)
The entertainment value of a problemsetter is computed as follows:
Consider all the tasks he/she (co-)authored. For each task find the count of happy contestants.
Sum all these counts.
For each problemsetter (in lexicographical order), print the problemsetter and his/her entertainment value.
(Note: Comparison between the first letter of contestant's name and task is case-sensitive!)