Some feedback for September Long 2017

First of all, thank you Codechef for another long challenge. These contests are great. There were a few things i disliked and i would like to point them out. Perhaps others agree with me and it will help improve the quality of the contest even more.

  • Problems were not balanced: 3 easy, 1 medium but apparently having weak test cases (SEACO), another medium, and 5 hard problems. Huge jump between 5 and 6 problems solved.

  • Complex problem statements: WEASELTX. It could have been explained much better with a few more sentences instead of those formulas and Delta signs :). Also WEASELSC problem statement was not clear enough, which i understood from the comments section where some great coders asked for explanation.

  • Too much math in some problems: SEAFUNC and SUMCUBE. I just lost interest after seeing all those formulas honestly.

This is just my humble opinion :).
Have a nice day!

13 Likes

I too agree with you. Even i was surprised to see the successful submissions of SEACO.

For the problems SUMCUBE and SEAFUNC, it took about 5 seconds to realize that i have no chance cracking them, although i used brute force in SUMCUBE for 8 points!

2 Likes
  1. Point 1 is correct. 1 simple and 2 cakewalk problems. I can imagine a lot of ties for those who solved on 3 problems…
  2. That was the exact problem coming in my mind when i red the first few words. That statement took me so long to interpret, that too at last at courtesy of a fellow contestant who I must thank for the explanation. That unnecessary use of that triangle and stuff. Gosh! It took me days to get what that statement went :frowning:
  3. Asking or not asking maths more or not depends on setter and his interests. Some people are really crafty when it comes to making mathematical problems. So I wont say any comments here.
1 Like

I agree too with the above points.
By the way is there some issue like rejudging etc since the ratings have not been updated yet.

1 Like

To me, although I didn’t solve the problem completely, such clear definition of problems are easy to understand and start working than a problem having “Ramayan”, “Mahabharat”, “History”, “Drama”, per se. I found myself nowhere in such questions. I prefer questions that are defined directly. I enjoyed all of the questions in this contest because it was easy to me to understand the problem in just one reading. Well solving a math problem is a different scenario altogether. But at least I understood the problem, to begin with.

For example, in the last month’s contest, I missed STRINGRA, just because the language seemed very cryptic to me and I didn’t understand the problem.

3 Likes

I don’t understand why so many people seem to be afraid of the challenge problem. There have been instances where you got basically no points at all unless you had a pretty sophisticated solution. But in most long challenges even a pretty naive and simple approach will get you quite a few points. Just have a look at this months editorial for the challenge problem. The solutions presented there ( for 40+ points) would at most fit in the EASY category for the algorithmic problems.

13 Likes

I suppose atleast top 4 problems to be of same difficulty range as previous long challenges. Since in this sept challenge, there was a gap between problem 1,2,3 and problem 4.

first 4 questions solved by 2000 people then 200 ,literally that’s not fair. first 3 questions were too easy ,i think codechef changed their difficulty level after snackdown 2k17

1 Like

Many thanks for the organization of those mind challenges.
It’s hard to fit everybody’s needs and wishes.

Usually I spend hours (translating, reading 10 times or more) trying to figure what sort of graph is expected and how it can be built.
(Clearly I’m weak with this category, discovering things I’ve never learned ; the focus for my future efforts.)

The most encouraging graph problem I had to solve in a long challenge this year had an URL pointing to the definition of the graph. It was a rooted graph iirc. It was nice to learn things this way. An external URL. Nice.

September Challenge was an Interesting and informative challenge for me.I believe all long challenges should be of this quality.
This challenge forced me improve my skills by working hard on some of the hard problems.

There must be 2 to 3 problems of Easy Level to motivate beginners and others Medium to Hard Problems.
The more the number of challenging( Medium - hard ) problems the participants will try to work hard on them ( since this is a Long Challenge) thereby upgrading their skill set.
Participants will try to experiment with different algorithms and Data structures in order to optimize their solution.
100 points per problem is only a motivation.
People must understand that.
The main aim must always be gaining the ability to solve hard problems by applying efficient algorithms and data structures.
Patience is the Key. Try to have a Never Give Up Attitude.

I believe the main aim of CodeChef is to promote Programming in India and help in developing quality programmers with the ultimate aim to bring a Gold Medal to India in this field.

I think the ranking is just a motivation.
But the Need Must Be Gold.

The easiness will not motivate Hard Work and practice ( Thats what i believe ).

As for the stories I think the stories in the problems are necessary as in real world the problems are not always direct ( The World is not Kind enough ). The problems are reduced to smaller subproblems or another well defined problem.
Thats the Nature of Problem Solving.

Thanks to the Problem setters we get to encouter such a diverse set of problems that force us to deduce them into Mathematical problems by analysing them.

The Cook Off must mantain the standard and should be set in a way so as to serve as a bridge for The ACM ICPC.

You can motivate few top emerging Programmers (of Low rating) who are working hard ( Solve more problems than their perfomance during previous contest) between or during contests by rewarding them with Codechef Laddus.

5 Likes
  1. Agree with this.
  2. This is a long contest. You have like 10 days of time to read and understand a problem statement.
  3. I beleive maths is a part of competitive programming.
1 Like

Actually, I got a bug in one of the questions. I don’t know if it will qualify as a bug or not, but it was related to a missing test case. The problem code is CHEFPDIG (Chef and Pick digits). The question has partial marking. In my code, I never got the output Z in the output string, but still, I received partial marks as Subtask 1 didn’t include check for letter Z. It had the test cases for letters other than Z but not Z. Although the test case for letter Z was present in Subtask 2 so, I received full points after correcting it for Z.

Here is my solution.
https://www.codechef.com/viewsolution/15314187

You can see that I have written the line for j in range(0,25). So in Python, j will never be equal to 25 so Z will never be printed. And just by luck, I didn’t get output as wrong answer, but I got partial marks.

I am new to Competitive Programming, so please don’t judge me on the quality of code which I have written.

I hope that you will correct the same. Looking forward to see some laddus in my account.

1 Like

Thanks, we will keep this feedback in mind. From next contests, we have focus on making first 5 problems relatively easier than usual, and remaining ones making a bit harder than usual. All your feedback is well taken, we will make sure to improve it up.

7 Likes

I don’t mind math in problems, but perhaps try to cover it behind some storyline :). I freak out when i see half the Greek alphabet in the formulas of a problem statement :).

2 Likes

Yes, thats a valid point. Giving away straight formula (much worse, giving away all those symbols which we dont know) is bad.

One of his point was Huge jump between 5 and 6 problems solved.. If you make first 5 easier and last 5 harder, you are just increasing the gap.

What you guys can do is, first 5 should be easy, 1st=cakewalk and 5th= EASY-MEDIUM. The difficulty should gradually increase, not like 5 problems done by 2000 and 6th problem only by 200.

8 Likes

Use the rating predictor, it’s extremely accurate:

https://discuss.codechef.com/questions/100887/codechef-rating-predictor

3 Likes

Next time I’ll try to come up with interesting story lines… It’s not that easy XD

4 Likes

Yes, I understand that. But anything will work. Some humor, something related to some issue, or any random story/situation coming to your mind. Problem story is one thing where setter can unleash the writer within :stuck_out_tongue:

1 Like

thanks a ton! Awesome scrip!!
Dont have enough credits to upvote :expressionless:

1 Like