partial points in cookoff84 : Chang and the Perfect Function giving full marks !!

I submitted a Brute solution to that problem and got 20 points out of 100 ( Partial Acceptance ) Which should not be allowed in cookoff ( This is the Rule right? ) . Later I submitted a correct solution and got 100/100 but it did not reflect in standings .

I belive there are many solutions which got full marks (1 point ) in the standings but only could pass partial test cases .

Please Rejugde !

Thanks and Do not unrate the contest please :slight_smile: !!

18 Likes

want to upvote…

4 Likes

You are absolutely right. There was glitch during contest that had not been noticed by @admin.

There are many peoples who got full score even after submitting a brute force solution of CHNGFUNC question. You can cross check this by looking into this solution. If you notice then you will find that this user got partial score of 20, but contest was not partial graded(No sub-tasks were there). So how this solution could be possible?

Therefore I would like to request @admin, please look into this issue ASAP.

1 Like

I checked the comments where @dpraveen himself said that-

rasensuriken- @admin there is an issue with test evaluation format. It is evaluating in IOI style instead of ACM style.
drpaveen- @rasensuriken Thanks, its been corrected.

Anyone has any leads on rejudge thing? The Q was not hard, but accepting that this problem had 1.2 k AC solutions (just a little more than cakewalk) makes me a bit uneasy…

3 Likes

The real Q is, why were even subtasks made for Cook-Off question in first place? There was no need to make separate sub task and label assign weightage 20,80 to them.

3 Likes

We are looking into the situation. I got to know about the issue from a user around 30 minutes into the round. The contest will remain rated. We will soon let you know what will be decided. A possible scenario is awarding ACs to all the users who pass even a single subtask, as that wouldn’t take away any accepted solution of a participant during the contest.

2 Likes

Wouldn’t awarding ACs to all the users who pass even a single subtask be unfair to people who actually solved the full problem? Writing a brute force solution doesn’t take time sir.

8 Likes

Awarding ACs to all users is unfair :frowning:

14 Likes

@dpraveen -

A possible scenario is awarding ACs to all the users who pass even a single subtask, as that wouldn't take away any accepted solution of a participant during the contest.

There are much better scenarios available. I understand that you dont want that someone who got AC during contest gets TLE and WA now, but we also need to look at the cost. Short contest is more of time and accuracy. If a user submitting a brute force solution is getting ahead of users submitting correct solution, then its really wrong.

That Q has 1262 correct submissions, and thats ~66% of entire population. Ranks 175-1262 were determined by this problem. A huge margin. If someone genuine gets his ratings reduced because other people’s brute force got accepted, then this wont be fair to him either. Infact its a lot unfairer than giving WA to brute forcers after contest.

Not only that, this system of AC on passing 1 subtask might result in other complications too. Someone getting WA in higher subtask but AC in lower one, or vice versa, and deserving WA has got AC.

The contest will remain rated.

Whatever you do, just remember, its better for contest be unrated than unfair. Its tough to get a “fair-to-all” solution, especially when population involved is large percentage wise.

4 Likes

Some users got AC with partially correct solutions in problem CHNGSUM also.

You can check this link. This user’s score is 3 but best solution submitted by him/her for problem CHNGSUM is partially correct(And He/She got AC for that).

I request admin to look into the matter.

9 Likes

I think this is not not limited to CHNGSUM. It’s for each problem in the July Cook-Off 2017. There are many users who got partial scores but were awarded ACs in problems other than CHNGSUM as well.
I even mailed at this e-mail ID about the same, around at 1am today and still awaiting the reply.

Check this out…
solution
his/her ranking

And yeah, please do-not unrate this .

Thanks :slight_smile:

2 Likes

If i submitted my solution in the first 30 minutes, I would have known my mistake too! Got WA just because of using int instead of long long! This is unfair :frowning:

2 Likes

See rank 19! No offence but his first solution got 20 points and he could easily know that his mistake was using int instead of long long! I had the same problem with my code but got WA because they changed the format until then! Isn’t it unfair??

2 Likes

@dishant_18 If you have some experience with cook-offs then you must know that cook off does not allow partial acceptance of solutions , and bugs in codechef are too common now :stuck_out_tongue: , You could have try to make it for 100 on 100 =) ,

1 Like

Ya I know but what I mean is many figured out whats wrong with their solution because of this bug. This makes the contest unfair!!

2 Likes

Thanks, we’ll check this up.

1 Like
bugs in codechef are too common now :P

This bug was really not expected. If the issue wa reported to admin, he should have been more careful. There were things that could had been done to avoid complications now. But whatever happened, has happened, i hope that whatever solution/decision the admins arrive is fair and satisfying to all.

1 Like

Just a suggestion: Instead of making the round unrated a free roll can be done as it was done a few days ago in a round at csacademy. In a free roll a person’s rating does not decrease. It is like rating = max(rating,rating after the contest). This would be fair to everyone imo :slight_smile:

15 Likes

Thats actually a nice idea. The chief issue is a fair/honest user getting his ratings reduced while brute forcers getting ratings increased. Rating is motivation to more people than i actually thought, so honestly, your idea seems really nice in this respect.

2 Likes

I feel that awarding AC to a guy who has partially solved is not fair at all.

I tried the third question for about an hour, but i could not come up with a solution less than O(n^2). I knew the solution wont pass, hence i did not code it at all, because i knew it would fail.

Most coders just know what time complexity will pass for a given question from it constraints and hence they would try to make a solution within that time limit.

Hence, it would just be unfair to all those people who did not brute force the question, even though they had the knowledge to brute force it.

I hope this is considered while taking the final verdict of the challenge.

3 Likes
//