I was wondering why top soutions are codechef are no longer ordered by time first and then by memory? I always feel curious to see the most efficient solution unfortunately some of the more efficient ones get pushed to the back! Also the solutions that take the least memory are only those with the least “memory in stack”. Often the top solutions use the same amount of memory that other codes do, only thing is that the memory in heap is not counted! So, I request you to please order the solutions by time again. 
Just a suggestion.
3 Likes
Great contest,learnt use of offline approaches and disjoint sets.
Enjoyed thoroughly!!!
the questions like sereja and random array only require some deep observation and very hard to get the trick.
plz give some focus on classical algorithms…
3 Likes
Really liked QCHEF. Took a lot of attempts to get it to AC, but totally worth it 
1 Like
Problems were good enough for one to enjoy the contest. Problems should be focused more on logic/trick than implementation.
I spent quite a lot of time solving these problems. The contest was more difficult than the previous ones. I’m looking forward to see the editorials for QCHEF, TREECNT2 and RNG.
Good contest. But I really didn’t like EMBED. Challenge problems should be more interesting. In my opinion, each participant may get some points on challenge problem.
1 Like
STRSUB and #QCHEF - Awesome problems (y)
5 Likes
i like this type of contest where the variation in difficulty varies like that…this contest was in the competing mode even though i did not participated cause of exams…i would really like this type of contest in future
The problems were really nice . It would be more fun if more questions were there in the long contest .
1 Like
The contest was better than the previous ones and it had various problems which required very keen observation and logical thinking, managing the time limit and memory limit played the major roles ! I had fun solving them.
My feedback/ views on March Challenge 2015 :
The easiest of this month’s long challenge, solved it at first attempt. Just went through the question and 5-7 minutes thinking and started coding. 
Signwave :
What are those humongous graphs !!! Scared and left it for some later time. After some long thinking I resorted to the old class 11 trigonometry book and found it to be quite a simple problem. Got busy coding and solved it on 15th March, full points in first attempt.
Read it on the first day, seemed to be a problem belonging to easy-medium difficulty, could not solved it. The problem technique seemed somewhat familiar ( now after seeing the editorial I see it is somewhat related to PRLADDU. Didn’t try much as internal/mid semester exam was going on from 04-13 ( bad excuse, others did it, I’m ashamed of my self 
).
Partially-solved, again seemed easy-medium problem but unfortunately I was unable to pass last three sub-subtask of the last subtask ( task 12, 13, 14 ).
At first this problem seemed quite easy, they have already provided the two generator functions with code ( this made me happy, less work in typing and just copy-paste the given functions ). The given generator with some modification and tweaking for faster execution, my solution was accepted with full points.
Matrix :
Started reading and had no idea what to do. finished reading and still no idea what to do !!! Went to the explanation section and saw are those pictures, the problem became more difficult to comprehend. Left it and going through the editorial now.
This problem seemed query based and I thought of implementing segment tree, but again was able to solve it only partially.
Well I am only a beginner/ advance-beginner and this problem was clearly out of my league.
Seemed quite straight forward and I submitted a brute force approach which gave TLE. Disappointed, I googled around for a few hours for matrix exponentiation and FFT but still TLE, so I left the problem.
I did not try this, just went through it and tried to grasp the approach/logic required to solve it.
What I like about March15 :
A few of the question were based on observations,tricks and mathematics. Ohh I loved the problem “signwave”, so easy but seemed quite hard, requires only +2 level trigonometry ). “Counting on a tree” has a very nice solution ( came to know quite a few things after reading the editorial ). “Sereja and random array” problem was also very nice based on tricks and maths. “Random number generator” involves some nice mathematical theories into a single problem ( great work by the problem author ).
What I dislike about March15 :
Most of the question were based on tricks,mathematics and keen observations. “Matrix” was difficult to grasp, “Random number generator” involves a lot of advanced mathematical manipulation ( Generating functions, FFT, number theoretic transform, polynomial mod, polynomial division, that’s 5 medium/advanced topics into a single problem ).
I agree with @dpraveen 
, March15 was comparatively tougher than the last few long challenges. Except CNOTE ( and probably SIGNWAVE ) the average difficulty level of the problems were higher compared to other long contests. In the past few long challenges, the difficulty level increased gradually from cakewalk (very easy), easy, easy-medium, medium then medium-difficult, advanced, hard. For three/four problems, some tricks were required. But this month almost all the problems ( except CNOTE ) required some level of observations and tweaks. The difficulty level jumped from cakewalk (CNOTE) to medium ( SIGNWAVE is based on observation, QCHEF ) and then jumped directly to hard category( SEAPROAR, TREECNT2, MATRIX, RNG, EMBED ).
The The best thing about March 15 long challenge is probably the editorials. It has been a while since we had such elaborate, well written, designed with examples editorials for Long contests. Awesome work by the editorialist @kevinsogo .
EDIT 1 : I agree with @lebron, upper bound of difficulty should not be decreased but intermediate difficulty level was quite high in March15 ( most of us are not familiar with ACM ICPC or any other training as mentioned by @lebron ).
7 Likes
The contest was good. I just loved solving the questions MTRWY and STRSUB and though I could not solve QCHEF even after endless attempts, I learnt a lot on the way!
However, I have a suggestion: there should be less of Mathematics & Number theory and more of Programming. SEAPROAR, SIGNWAVE and RNG all were based on Mathematics, though SIGNWAVE was one very innovative problem. Instead, I would like to see more problems on a variety of topics like Strings, Computational Geometry, DP, Data Structures like BIT, Segment Tree etc.
Tests for TREECNT2 were weak.
questions were good and more tough problem should be added in coming contests
1 Like
I didn’t get to try the other questions because I got stuck at Signwave. I found only a 30 points solution based on observations about how the sin function behaved.
I don’t agree that upper bound of difficulty should be decreased. Problemset this time wasn’t hard, for people used to ACM ICPC trainings it takes no more than an hour to get lineups of solutions (and then it turns out just into finding time for coding), and giving easier problemset will switch all leaderboad competing just to challenge problem. This idea is good only if you are sure that you have a good challenge problem This time it seems to be good, but often they are really awful - like when question was about factorization:) When that problem isn’t very good, and all top50 solved other 9 problems, then board isn’t informative.
And about problems from this contest - it is strange idea to give problems like SEAPROAR, which requires only 10 minutes of using Google to solve it) about last 2 tasks - TREECNT2 is pretty standard ACM task; and RNG is also generally well-known, it also takes not a long time to google details of solution (even if writing the solution itself isn’t that fast&easy), this way of creating problem (take standard task and increase constants) does not look good for long contests where one can simply google a solution for original standard task (and ways to make it run fast).
5 Likes
Difficulty level of challenge can vary like this month’s challenge and previous month’s challenge. Both has its own benefit for different kinds of contestants.
Meanwhile my discussion on three easy problems in march challenge: https://shadekcse.wordpress.com/2015/03/16/codechef-march15-challenge/
It was probably the best long challenge in 2-3 months.