November Easy is taking place on November the 1-st at 19:00 MSK. Contest duration is 3 hours.
The contest will consist of 6 problems. Each problem is worth 100 points. Most of them have partial scoring. Ties are resolved by penalty time.
I've decided to experiment a bit, so the format of the contest will slightly differ from the previous ones. There will be:
4 usual algorithm problems of various difficulty. In my opinion they might be interesting for complete beginners as well as for enough experienced participants.
2 problems that differ from usual ACM / IOI tasks.
I'd like to thank all the guys who worked on it with me: admins belowthebelt, MDCCXXIX, tester Kmcode and editorialist pkacprzak. Your suggestions and ideas were really useful. I hope I'll work with all of you again in the future :)
As this contest is called Easy it will not require complicated implementations or too hard ideas, but I'm quite sure that participants of almost all the levels will be able to find interesting problems.
Good luck and enjoy it!
And almost forgot about nice bonus: Top 5 beginners (1st year or 2nd year) will receive HackerEarth Tshirts.
UPD: contest starts in less than 2 hours
The contest begins in 100 minutes! Good luck, everyone.
The third problem was nice, also the fifth was nice — but I think that can be standard for many coders.
I can not understand guessing in the fourth and sixth task. Is this programming ?
No, these are just guessing. It was quite unexpected to see such tasks in usual algorithmic contest.
I really want to se editorial for permutation task :)
Is this contests rated, I don't have big experience with HackerEarth ?
I leaked the numbers in memory consumption.. Interesting how to solve in from the first attempt
I was trying to solve a different problem, in which we have shuffled tests — so we need to distinguish between 256 values, not between 16 :)
The problem is — both time and memory usage measurements are note very precise :) I think we can encode 2 bits by memory usage (ML is low, plus memory measurement is not very precise...), 5 bits by time usage and 1 more bit by WA/RE. It gives us 8 bits, should be enough.
Have somebody tried it? :)
I was trying to solve the same version. In the first submit I wanted to get all first numbers, using time usage. I was surprised that time usage created a sorted sequence and later noticed that the i-th test indeed described the i-th number in their permutation. Later, I managed to distinguish almost all 16 values (2 were very close) with one submit, using time usage again.
I remember that in one Clash I got all input files in optimization problem. With memory usage I could get 15 bits from each test (with one submit).
Memory usage became very precise when I added sleep(0.5) to the code. I allocated memory using calloc(1, 128 * 1024 * a)$ where "a" is the value I want to learn (1..16). Then the values I got were (for indexes):
548, 676, 804, 932, 1060, 1188, 1316, 1444, 1572, 1700, 1828, 1956, 2084, 2212, 2340, 2468
So they were very precise, 128k as I was allocating.
Really want to know statement of last program :)
There was answer in the comments. Sum of squares of digits of M, repeat N times.
I've just posted all the editorials. They are available on each problem's page.