Hello everyone!

I would like to invite you to participate in an upcoming HackerEarth contest — HackerEarth March Circuits. It's a long contest that will start on March, 17 21:00 IST (check your timezone). Contest will run for 9 days.

The problemset consists of 7 traditional algorithmic tasks and 1 approximate problem. For traditional algorithmic tasks, you will receive points for every test case your solution passes — so you can get some points with partial solutions as well. For the approximation task, your score depends on the best solution in the contest so far. Please check contest page for more details about in-contest schedule and rules.

I'm a tester of a problemset you'll have to work on. Thanks to harshil for preparing classic part of the problemset, to pkacprzak for preparing approximate task and to r3gz3n for handling technical part of contest preparation.

Tasks should not be very hard for top-level contestants, and I expect them to get full score on classic part of a problemset. However, even if you solved everything — you'll still have to do your best on approximate problem :)

As usually, there will be some nice prizes for those who'll reach top spots, here are prizes for top5 (in case you haven't open contest page):

1. $100 Amazon gift card + HE t-shirt.
2. $75 Amazon gift card + HE t-shirt.
3. $50 Amazon gift card + HE t-shirt.
4. HE t-shirt.
5. HE t-shirt.

Upd. Contest has ended, approximate problem has been judged on final test data, all editorials has been published, all codes are public now.

Congratulations to winners:

1) mugurelionut

1) biginnnner

3) yzyz

4) ceilks

4) SkyFire

Hello Codeforces!

Have you heard about HackerEarth Collegiate Cup? Here is a general invitational post about it — Invitation to HackerEarth Collegiate Cup. In case you missed it, or failed to qualify for the next round, I have a nice surprise for you — there will be a Wild Card Round which wasn't mentioned in original announcement. This contest will take place on October 02, 10:00 IST.

This contest will give another chance for all those who didn't qualify or missed the previous rounds to directly get into the Second Elimination. Top 20 teams will directly qualify for the Second Elimination. Make sure you will have 3 members in your team and all from same college.

This contest will be held using ICPC-style rules (penalties, no partial scoring). You'll have 2 hours to solve problems prepared by Yury_Bandarchuk and kingofnumbers. I worked on it as a tester, and after the contest you'll have editorials by magieNoire provided.

Good luck! See you on the leaderboard. And in the Second Elimination afterwards ;)

Hello everyone!

I would like to invite you to participate in another HackerEarth contest. This time it is HackerEarth July Circuits. It's a long contest that will start on July, 16 21:00 IST (check your timezone). Contest will run for 8 days.

The problemset consists of 7 traditional algorithmic tasks and 1 approximate problem. For traditional algorithmic tasks, you will receive points for every test case your solution passes — so you can get some points with partial solutions as well. For the approximation task, your score depends on the best solution in the contest so far. Check contest page for more details about in-contest schedule and rules.

I'm a tester of a problemset you'll have to work on — thanks to r3gz3n, pkacprzak, Yury_Bandarchuk, magieNoire.

Tasks should not be very hard for top-level contestants, and I expect them to get full score on classic part of a problemset. However, even if you solved everything — you'll still have to do your best on approximate problem :)

As usually, there will be some nice prizes for those who'll reach top spots, here are prizes for top5 (in case you haven't open contest page) :

1. $100 Amazon gift card + HE t-shirt.
2. $75 Amazon gift card + HE t-shirt.
3. $50 Amazon gift card + HE t-shirt.
4. HE t-shirt.
5. HE t-shirt.

Upd. Contest has ended, congratulations to winners:

1) mugurelionut

2) ceilks

3) Rox

4) Sudhanshu Sindhu

5) jtnydv25

All solutions are public, solutions by setter and tester are also available now, all editorials have been published.

Hello everyone!

I want to invite you to participate in July Clash at HackerEarth. Contest is scheduled on July, 9. Contest duration is 24 hours, so there should be some comfortable time for every timezone :)

There will be six tasks in a problemset. Five of them are standard algorithmic problems with partial solutions allowed — you get points for every test that your solution passed. And the last task is an approximate problem — your score for this task depends on how good your solution is comparing to current best solution.

Lewin is author of this problemset. He is experienced contestant and experienced setter as well, you can check his previous contests to know what to expect :)

I was working on this contest as a tester, and I enjoyed solving given problems a lot — I hope you'll have same feeling :) I hope that some tasks will be not too hard for beginners (don't give up and show your best with partial scoring — even very naive solution may give you some decent points; and 24 hours should be enough for you to read all problems and find some tasks which you can solve) while some tasks are challenging enough to make this contest interesting for more experienced contestants. Each problem will have subtasks mentioned, so you can figure out how much your partial solution should get for sure before even implementing it ;) Also I want to thank belowthebelt for handling all technical aspects of contest preparation.

As usual, here is one more reason for you to participate in this contest:

Top5 of leaderboard will also receive some nice prizes:

1. $100 Amazon gift card + HackerEarth T-shirt
2. $75 Amazon gift card + HackerEarth T-shirt
3. $50 Amazon gift card + HackerEarth T-shirt
4. HackerEarth T-shirt
5. HackerEarth T-shirt

Good luck to everybody — I hope to see you at the scoreboard :)

Upd. Contest has ended, thanks to everyone for participating! Congratulations to winners:

1) anta

2) ecnerwala

3) LayCurse

4) y0105w49

5) cjtoribio

All solutions are public, solutions by setter and tester are also available now. Editorials for most of the problems have been published; in a few days remaining editorials will be added / published editorials will be improved and also updated with some alternative approaches from contestants (some of them look quite interesting).

Hello everyone!

I would like to invite you to participate in another HackerEarth contest. This time it is HackerEarth May Circuits. It's a long contest that will start on May, 21, 21:00 IST (check your timezone). Contest will run for 8 days.

The problemset consists of 7 traditional algorithmic tasks and 1 approximate problem. For traditional algorithmic tasks, you will receive points for every test case your solution passes — so you can get some points with partial solutions as well. For the approximation task, your score depends on the best solution in the contest so far. Check contest page for more details about in-contest schedule and rules.

I'm a tester of a problemset you'll have to work on — thanks to xennygrimmato, pkacprzak, himanshujaju, shef_2318 and Sumeet.Varma for preparing these tasks.

The contest will be rated. Exact format of a contest is still under development, so expect to see a lot of changes in future — but this time rules are same as in previous edition (here is an announcement).

Tasks should not be very hard for top-level contestants, and I expect them to get full score on classic part of a problemset. However, even if you solved everything — you'll still have to do your best on approximate problem :)

As usually, there will be some nice prizes for those who'll reach top spots, here are prizes for top5 (in case you haven't open contest page) :

1. $100 Amazon gift card + HE t-shirt.
2. $75 Amazon gift card + HE t-shirt.
3. $50 Amazon gift card + HE t-shirt.
4. HE t-shirt.
5. HE t-shirt.

Upd. Contest has ended :) Thanks to everyone for participating :) Congratulations to winners:

1) Carsten Eilks

2) simonlindholm

3) riadwaw

4) uwi

5) shdut

All solutions are available, also for all classic problems editorials and solutions by setter and tester have been published.

Hello everyone!

I would like to invite you to participate in HackerEarth May Easy Round that will be held on Sunday, 01 of May. The duration of the contest is 3 hours.

The problems were set by pkacprzak and tested by me. The difficulty of this round will be not harder than CodeForces Div. 2 Round (maybe even easier). You'll be provided with 5 classic problems having partial scoring — you get points for every test that your solution passed, and also one approximate problem — your score for this task depends on how good your solution is comparing to current best solution (and don't be afraid of facing an approximate problem in a contest which lasts for 3 hours only ;) ). This contest will be a rated HE Round.

You'll be provided with statements in English, and there will also be a Russian version as well — thanks to shef_2318 for working on a problemset as a translator.

Also I want to thank to belowthebelt for helping with technical aspects of contest preparation.

In any case, Good Luck && Have Fun all of you! I hope to see you in standings :)

UPD. Please notice that problem statement of Bridges in Karak has been changed (there is an announcement about it in a contest as well — output format has been changed). We are sorry for it; in case you already attempted that problem — we strongly recommend you to check updated statement and try to fix your old solution. We'll do our best to minimize effect of this mistake in contest standings (exact decisions about rejudging old submissions and stuff like that will be made later), also contest has been extended by 15 minutes.

Hello everyone!

I want to invite you to participate in March Clash at HackerEarth. Contest is scheduled on March, 26. Contest duration is 24 hours, so there should be some comfortable time for every timezone :)

There will be six tasks in a problemset. Five of them are standard algorithmic problems with partial solutions allowed — you get points for every test that your solution passed. And the last task is an approximate problem — your score for this task depends on how good your solution is comparing to current best solution.

PrinceOfPersia is author of this problemset. You may check some of his previous contests to ensure that he's always providing contestants with interesting tasks :)

I was working on this contest as a tester, and I enjoyed solving given problems a lot :) I hope that several tasks will be not too hard for beginners (don't give up and show your best with partial scoring — even very naive solution may give you some decent points; and 24 hours should be enough for you to read all problems and find some tasks which you can solve) while some tasks are challenging enough to make this contest interesting for more experienced contestants. Even if you think that classic part of problemset is easy — go on, try to beat other contestants on approximate problem :) shef_2318 worked on this contest as translator — you will be provided with problem statements in English and Russian. Also I want to thank belowthebelt for handling all technical aspects of contest preparation.

As usual, here is one more reason for you to participate in this contest:

Top5 of leaderboard will also receive some nice prizes:

1. $100 Amazon gift card + HackerEarth T-shirt
2. $80 Amazon gift card + HackerEarth T-shirt
3. $50 Amazon gift card + HackerEarth T-shirt
4. HackerEarth T-shirt
5. HackerEarth T-shirt

Good luck to everybody — I hope to see you at the scoreboard :)

Upd. Contest has ended :) Thanks to everyone for participating :) Congratulations to winners:

1) Carsten Eilks

2) ffao

3) LayCurse

4) pavel.savchenkov

5) mugurelionut

All solutions are available, also for all classic problems editorials and solutions by setter and tester have been published.

Hello everyone!

I want to invite you to participate in February Clash at HackerEarth. Contest is scheduled on February, 20. Contest duration is 24 hours, so there should be some comfortable time for every timezone :)

There will be six tasks in a problemset. Five of them are standard algorithmic problems with partial solutions allowed — you get points for every test that your solution passed. And the last task is an approximate problem — your score for this task depends on how good your solution is comparing to current best solution.

shef_2318 is author of this problemset. If you want to get an idea what to expect — you may check his previous contests at HackerEarth, CodeChef and HackerRank. Even if you don't need it in order to participate in Clash — I would advice you to take a look at them someday, because there are quite a few really interesting problems there.

I was working on this contest as a tester. I believe it is going to be an interesting one, as usual :) I hope that several problems will be not too hard for beginners (don't give up and show your best with partial scoring — even very naive solution may give you some decent points; and 24 hours should be enough for you to read all problems and find some tasks which you can solve) while some tasks are challenging enough to make this contest interesting for more experienced contestants. Even if you think that classic part of problemset is easy — go on, try to beat other contestants on approximate problem :) shef_2318 also worked on this contest as translator — you will be provided with problem statements in English and Russian. Also I want to thank belowthebelt for providing technical help and doing his best on fixing all issues and improving HackerEarth platform.

As usual, here is one more reason for you to participate in this contest:

Top5 of leaderboard will also receive some nice prizes:

1. $100 Amazon gift card + HackerEarth T-shirt
2. $80 Amazon gift card + HackerEarth T-shirt
3. $50 Amazon gift card + HackerEarth T-shirt
4. HackerEarth T-shirt
5. HackerEarth T-shirt

Good luck to everybody — I hope to see you at the scoreboard :)

Upd. Contest has ended, congratulations to winners:

1) eatmore

2) mugurelionut

3) Carsten Eilks

4) RomaWhite

5) Egor

All editorials have been published.

Hello everyone!

I want to invite you to participate in another contest at HackerEarth — this time it is India Hacks 2016.

Check out the prizes:

Maybe you already saw an announcement of Qualification Round 1 by shef_2318. In case you missed it — there will be two more qualification rounds.

Qualification Round 2 (link) will take place on January, 10 (check your timezone). Top 1000 participants from this round will qualify for the main contest.

If you like prizes on a picture above and still haven't qualified for the next round, or you would like to take part in an interesting competition against a lot of coders from all over the world, or you want to try yourself on a set of not very hard but still interesting and educative problems — this one is for you!

I am a tester of problemset which you'll face during a contest. There will be 5 problems with partial scoring system (you get points for every test which you passed), and you'll have 3 hours to solve them. Thanks to RomaWhite, Baba, subway, kunal93, harshil for preparing problems and to belowthebelt for providing technical help with contest preparation.

Good luck and have fun! Hope to see you in standings :)

UPD. Contest has ended! Congratulations to anta on being fastest to solve all 5 problems, and to 24 other contestants who also managed to get full score. Best luck in next round to all those who qualified to next round, and for all those who didn't get into top1000 — you'll still have one more chance in Qualification Round 3.

Thanks to everyone for participating! Solutions and editorials are public now; feel free to discuss problems.

Hello everyone!

I want to invite you to participate in November Clash at HackerEarth. Contest is scheduled on November, 28. Contest duration is 24 hours, so there should be some comfortable time for every timezone :)

There will be six five tasks in a problemset. Five Four of them are standard algorithmic problems with partial solutions allowed — you get points for every test that your solution passed. And the last task is an approximate problem — your score for this task depends on how good your solution is comparing to current best solution.

witua is author of this problemset. He is experienced setter — I bet you already saw his problems at Codeforces, TopCoder, CodeChef or some other site, and now it is time for him to debute at HackerEarth.

I was working on this contest as a tester. I believe it is going to be an interesting one :) I hope that several problems will be not too hard for beginners (don't give up and show your best with partial scoring — even very naive solution may give you some points; and 24 hours should be enough for you to read all problems and find some tasks which you can solve) while some tasks are challenging enough to make this contest interesting for more experienced contestants. shef_2318 worked on this contest as translator — you will be also provided with problem statements in Russian. Also I want to thank to thank belowthebelt for providing technical help and doing his best on fixing all issues and improving HackerEarth platform.

As usual, here is one more reason for you to participate in this contest:

Top5 of leaderboard will also receive some nice prizes:

1. $100 Amazon gift card + HackerEarth T-shirt
2. $80 Amazon gift card + HackerEarth T-shirt
3. $50 Amazon gift card + HackerEarth T-shirt
4. HackerEarth T-shirt
5. HackerEarth T-shirt

Good luck to everybody — I hope to see you at the scoreboard :)

Upd. Contest has ended :) Thanks to everyone for participating :) Congratulations to winners:

1) anta

2) Carsten Eilks

3) Kmcode

4) HellKitsune

5) mugurelionut

Hello everyone!

I want to invite you to another contest at HackerEarth, and this time it is not Clash :) But ACM ICPC Practice Contest. Contest is scheduled on October, 4.

You'll face ICPC-style problemset with quite a lot of problems) Contest is focused on ICPC preparation, so are the rules — it will be 5h long team event with ICPC scoring rules (only full solutions count, ranking by number of problems and then by penalty time, giving penalty for wrong submissions), and only C/C++/Java in list of allowed languages.

Problemset was prepared for you by FatalEagle, memset123, pkacprzak, shef_2318, xennygrimmato, Sokolov and tanmaysahay94. There will be also statements in Russian, prepared by shef_2318. After contest editorials by pkacprzak will be presented to you. I was working on this contest as a tester. Also I want to thank to belowthebelt for technical help and doing his best on fixing all issues, processing all our feedback and improving HackerEarth platform)

I believe that problemset is quite easy, at least for experienced contestants. And I hope it is actually easy, not just another "I will say that problems are easy to make you feel even more miserable". If I am wrong — please exuse me :) If your team is strong one — it will be a nice warm up for you; otherwise I believe you'll find some challenging problems there and learn a few new things.

Another good reason to try this contest — there are some nice prizes. Easy problems — easy prizes, right? :)

The top 3 Indian teams in this contest, who also qualify for ACM ICPC Regional will have their ACM ICPC travel expenses reimbursed.

The top 3 teams will get $100, $80 and $50 vouchers respectively.

Good luck, waiting to see you in scoreboard!

UPD. Congratulations to winners!

Top3 teams:

1) Anta (anta)

2) bloody_unko (uwi, sugim48)

3) Lewin (Lewin)

Top3 Indian teams:

1) FacelessMen (alecsyde, gsahil, abhilak)

2) DAFruitsSalad (yashkumar18, Sumeet.Varma, kuldeeppatel)

3) Rex_Regum (usaxena95, aditya1495, harshil)

Hello everyone!

I want to invite you to participate in September Clash at HackerEarth. Contest is scheduled on September, 19. Contest duration is 24 hours, so there should be some comfortable time for every timezone :)

There will be six tasks in a problemset. Five of them are standard algorithmic problems with partial solutions allowed — you get points for every test that your solution passed. And the last task is an approximate problem — your score for this task depends on how good your solution is comparing to current best solution.

Lewin is author of this problemset. He already prepared several contests at TopCoder and Codeforces, and now you'll have an opportunity to solve one at HackerEarth.

I was working on this contest as a tester. I believe it is going to be an interesting one :) And I learned some new things while helping with preparation :) I hope that several problems will be not too hard for beginners (don't give up and show your best with partial scoring; 24 hours should be enough for you to read all problems and find some tasks you can solve) while some tasks are challenging enough to make this contest interesting for more experienced contestants. shef_2318 worked on this contest as translator — you will be provided with statements in Russian also. Also I want to thank to belowthebelt for technical help and doing his best on fixing all issues, processing all our feedback and improving HackerEarth platform.

As usual, here is one more reason for you to participate in this contest:

Top5 of leaderboard will also receive some nice prizes:

1. $100 Amazon gift card + HackerEarth T-shirt
2. $80 Amazon gift card + HackerEarth T-shirt
3. $50 Amazon gift card + HackerEarth T-shirt
4. HackerEarth T-shirt
5. HackerEarth T-shirt

Good luck to everybody — I hope to see you at the scoreboard :)

Upd. Contest has ended :) Thanks to everyone for participating :) Congratulations to winners:

1) FatalEagle

2) zeliboba

3) yarrr

4) rnsiehemt

5) anta

Hello everyone!

I want to invite you to participate in August Clash at HackerEarth. Contest is scheduled on August, 15. Contest duration is 24 hours, so there should be some comfortable time for every timezone :)

There will be six tasks in a problemset. Five of them are standard algorithmic problems with partial solutions allowed — you get points for every test that your solution passed. And the last task is an approximate problem — your score for this task depends on how good your solution is comparing to current best solution.

PrinceOfPersia is author of this problemset. Check his blog to get a taste of problems authored by him — he already prepared a lot of different interesting contests. Right now Hunger Games are running :)

I was working on this contest as a tester. I would like to say that I find this problemset interesting, I hope that several problems will be not too hard for beginners (don't give up and show your best with partial scoring) and some tasks are challenging enough to make this contest interesting for more experienced contestants. shef_2318 worked on this contest as translator — you will be provided with statements in Russian also. Also I want to thank to belowthebelt for technical help and doing his best on fixing all issues, processing all our feedback and improving HackerEarth platform.

As usual, here is one more reason for you to participate in this contest:

Top5 of leaderboard will also receive some nice prizes:

1. $100 Amazon gift card + HackerEarth T-shirt
2. $80 Amazon gift card + HackerEarth T-shirt
3. $50 Amazon gift card + HackerEarth T-shirt
4. HackerEarth T-shirt
5. HackerEarth T-shirt

Good luck to everybody — I hope to see you at the scoreboard :)

Upd. Almost 10 hours passed from the start, congratulations to anta on being first to reach full score on classic part of problemset! All of you still have more than 14 hours to reach a top spot by getting 500 on classic part of problemset and beating everybody on approximate problem :)

Upd2. Less than 30 minutes left till the end. We have a close matchup between enot110 and anta; and FatalEagle is third person with full score on classic part of problemset.

Upd3. Congratulations to winners:

1) enot110

2) anta

3) FatalEagle

4) GaryYe

5) azneyes

Hi everybody :)

There were several topics about live streams recently (24h event by Mimino at HackerRank, live streams with enot110 solving Timus...). I already have an experience of making a screencast, and now I am also going to try live streaming.

Streaming real contests isn't a good idea :) Also I am not a huge fan of solving OJ; so the question of choosing idea for first stream is still open.

Right now I think that the best idea would be to open SPOJ (I was never really active at this OJ, there are not a lot of problems solved at my account there — and I have a lot of easy problems remaining) and start solving some tasks there. But I am open to suggestions and ideas)

I am not aiming at making it very interesting/entertaining (so maybe it will be simply about me listening to music and writing code... so boring...); also I am not going to do some lecturing or things like that — I am both not good enough for it and not interested in it:)

Maybe it will be a long one (24 hours does not sound like a limit) — or maybe it will be too boring for me and I'll stop very soon.

I'll update this post with more details later :) Hope to see some nice ideas in comments.

UPD. Probably I'll start a session on Monday morning, I'll try solving easy SPOJ problems (and maybe also some tasks which you had suggested) and I'll use Twitch, as StoneCold_ suggested. Need to get familiar with it now :) I'll post a link and any further updates here.

UPD. #2 Here is Twitch channel. I'll start approximately at 6 A.M. on Monday.

UPD. #3 Approximate schedule of future streams has been published at Twitch channel.

Hello everyone!

I want to invite you to participate in June Clash at HackerEarth. Contest is scheduled on June, 20. Contest duration is 24 hours, so there should be some comfortable time for every timezone :)

There will be five tasks in a problemset. Four of them are standard algorithmic problems with partial solutions allowed — you get points for every test that your solution passed. And the last task is an approximate problem — your score for this task depends on how good your solution is comparing to current best solution.

shef_2318 is author of this problemset. He already prepared January Lunchtime 2015 and a few interesting problems for CodeChef Long contests, and also he was both author of April Clash and co-author of May Clash.

I was working on this contest as a tester. As usual, I would like to say that I find this problemset interesting, I hope that some problems will be not too hard for beginners (don't give up and show your best with partial scoring) and some other tasks are challenging enough to attract more experienced contestants. I was glad to work with shef_2318 again :) Also I want to thank to MDCCXXIX for technical help and doing his best on fixing all issues and improving HackerEarth platform.

There will be quite a lot of different contests over the weekend, and besides interesting problems, I have another reason for you to participate in this one:

Top5 of leaderboard will also receive some nice prizes:

1. $100 Amazon gift card + HackerEarth T-shirt
2. $80 Amazon gift card + HackerEarth T-shirt
3. $50 Amazon gift card + HackerEarth T-shirt
4. HackerEarth T-shirt
5. HackerEarth T-shirt

I hope everything will run smoothly this time. Good luck to everybody — I hope to see you at the scoreboard :)

Hello everyone!

I want to invite you to participate in June Clash at HackerEarth. Contest is scheduled on June, 20. Contest duration is 24 hours, so there should be some comfortable time for every timezone :)

There will be five tasks in a problemset. Four of them are standard algorithmic problems with partial solutions allowed — you get points for every test that your solution passed. And the last task is an approximate problem — your score for this task depends on how good your solution is comparing to current best solution.

shef_2318 is author of this problemset. He already prepared January Lunchtime 2015 and a few interesting problems for CodeChef Long contests, and also he was both author of April Clash and co-author of May Clash.

I was working on this contest as a tester. As usual, I would like to say that I find this problemset interesting, I hope that some problems will be not too hard for beginners (don't give up and show your best with partial scoring) and some other tasks are challenging enough to attract more experienced contestants. I was glad to work with shef_2318 again :) Also I want to thank to MDCCXXIX for technical help and doing his best on fixing all issues and improving HackerEarth platform.

There will be quite a lot of different contests over the weekend, and besides interesting problems, I have another reason for you to participate in this one:

Top5 of leaderboard will also receive some nice prizes:

1. $100 Amazon gift card + HackerEarth T-shirt
2. $80 Amazon gift card + HackerEarth T-shirt
3. $50 Amazon gift card + HackerEarth T-shirt
4. HackerEarth T-shirt
5. HackerEarth T-shirt

I hope everything will run smoothly this time. Good luck to everybody — I hope to see you at the scoreboard :)

Upd. Contest has ended :) Thanks to everyone for participating :) Congratulations to winners:

1) kutengine

2) FatalEagle

3) anta

4) SoMin Mun

5) Kmcode

Hello everyone!

I want to invite you to participate in May Clash at HackerEarth. Contest is scheduled on May, 23. Contest duration was increased comparing to previous editions — now it lasts for 24 hours, to make it even easier for you to find some time at least to try it :)

There will be six tasks in a problemset. Five of them are classic algorithmic problems with partial solutions allowed — you get points for every test that your solution passed. And the last task is an approximate problem — your score for this task depends on how good your solution is comparing to current best solution.

PrinceOfPersia is author of this problemset. Check his blog to get a taste of problems authored by him — he was a writer of recent Codeforces Round 299 (Div. 1), and also prepared few interesting contests in CF gym.

I was working on this contest as a tester. I would like to say that I find this problemset interesting, I hope that some problems will be not too hard for beginners (don't give up and show your best with partial scoring) and some other tasks are challenging enough to attract more experienced contestants. It was nice experience to work with PrinceOfPersia on preparing problemset. He really impressed my by the speed with which he is generating ideas :) Also I want to thank to MDCCXXIX for technical help and doing his best on fixing all issues and improving HackerEarth platform.

If you are still not sure about participation in this contest — I want to add one more thing,

Top5 of leaderboard will also receive some nice prizes:

1. $80 Amazon gift card + HackerEarth T-shirt
2. $60 Amazon gift card + HackerEarth T-shirt
3. $40 Amazon gift card + HackerEarth T-shirt
4. HackerEarth T-shirt
5. HackerEarth T-shirt

I guess some of top contestans will be busy because of ACM ICPC World Finals — it only increases your chances for a prizes :)

Good luck to everybody — I hope to see you at the scoreboard :)

UPD. First announcement said that contest is scheduled on May, 16. Due to unavoidable reasons it has been moved on May, 23 now.

Hello everyone!

I want to invite you to participate in April Clash at HackerEarth. Contest is scheduled on this Saturday. It lasts for 12 hours, so almost everyone can find some time at least to try it :)

There will be five tasks in a problemset. Four of them are standard algorithmic problems with partial solutions allowed — you get points for every test that your solution passed. And the last task is an approximate problem — your score for this task depends on how good your solution is comparing to current best solution.

shef_2318 is author of this problemset. Check January Lunchtime 2015 or problem Little Party from April Challenge 2015 if you want to try some of his tasks as part of preparation for upcoming contest :)

Once again, I was working on this contest as a tester. Time runs fast when you are preparing for ACM ICPC World Finals; it feels like I announced March Clash just few days ago, and now it is time for April Clash already. I would like to say that I find this problemset interesting, I hope that some problems will be not too hard for beginners (don't give up and show your best with partial scoring) and some other tasks are challenging enough to attract more experienced contestants. It was nice to work with shef_2318 on preparing problemset, we discussed a lot of problems while preparing a contest — both related to it and not related at all :) Also I want to thank to MDCCXXIX for technical help and doing his best on fixing all issues and improving HackerEarth platform.

For additional motivation, I have to remind you that top 3 winners will receive HackerEarth T-shirt. (I already got mine for one of previous contests :) )

Good luck to everybody — I hope to see you at the scoreboard :)

(Update)

The contest has ended! Thanks to everyone for participating:)

Congratulations to winners:

1) mugurelionut

2) azukun

3) I_love_Hoang_Yen

Hello everyone!

I want to invite you to participate in March Clash at HackerEarth. Contest is scheduled on this Sunday. It lasts for 12 hours, so almost everyone can find some time at least to try it :)

There will be five tasks in a problemset. Four of them are standard algorithmic problems with partial solutions allowed — you get points for every test that you solution passed. And the last task is an approximation problem — you score for this task depends on how good your solution is comparing to current best solution.

laoriu is author of this problemset — you may know him from preparing Codeforces Round 277 (Div. 2) and December Clash 2014. I was working on this contest as a tester. I would like to say that I find this problemset interesting and it was also a great experience for me to work with laoriu on preparing it. Also I want to thank to MDCCXXIX who helped me a lot with technical part of preparation.

Top 3 winners will receive HackerEarth T-shirt. (I already got mine for one of previous contests :) )

Good luck to everybody — I hope to see you at the scoreboard :)

(Update)

The contest has ended! Thanks to everyone for participating:)

Congratulations to winners:

1) mugurelionut

2) HellKitsune

3) pavel.savchenkov

Недавно ballon в обсуждении SnarkNews winter series сказал, что мне пора делать скринкасты. Не знаю, многим ли интересно что-то такое увидеть, но пусть будет. Как я ответил в той теме, регулярно писать скринкасты я не вижу смысла. Но один контест все же решил записать — для тех, кому интересно; для упрощения ответа всем тем, кто задает вопросы вида "секрет красного цвета в правильной настройке Visual Studio" (да, есть пользователи, которые вроде бы вполне серьезно спрашивают подобные вещи); да и просто испробовать новый опыт.

Ниже — запись моего участия в SNWS 2015 Round 4. Весь экшн оканчивается на 47 минуте, когда все задачи сданы, дальше только обновление монитора.

Предыстория: я уже много раз слышал/читал о том, что double начинает дико тупить, если получаются числа меньше 1e-306. Подобные разговоры всплывали после одного из контестов в зоне NEERC, после одного из раундов TC, еще видел такие фразы в обсуждении некоторых задач на различных online judge. Обычно советуют лечить это с помощью if (value<eps)value=0; Так как это мне всегда помогало, то выглядело вполне правдоподобно. Насколько это правда на самом деле?

Актуальная проблема: недавно столкнулся с похожей проблемой на СF. Сравните 3 сабмита: 7256779, 7256781, 7256786. Они отличаются только в void mult, при этом первое решение работает в десяток раз медленней двух остальных. Один из АС сабмитов действительно с костылем if(value<eps), в другом же разница в том, что две переменные заменены на локальные вместо глобальных. Кто-то может более-менее понятно объяснить, откуда такая разница в продуктивности?

Идем дальше. Выбросим из кода все лишнее) Этот код работает быстро и без проблем. Если переписать его вот так, в стиле кода, из-за которого и была создана эта тема, то продуктивность резко падает после 7 итерации (46 мс для 7 итераций, 1092 мс для 8 итераций, 2698 мс для 9 итераций). Обратите внимание, что в этом коде уже есть if (value<eps), но оно не помогает. Если сделать переменные локальными, как было в одном из примеров выше, то решение начинает проседать на одну итерацию позднее (62 мс для 8 итераций, 1653 мс для 9 итераций). Почему вообще так? И почему в этом случае костыль с <eps абсолютно никак не влияет на ситуацию — хотя в начальной задаче, в, казалось бы, аналогичной ситуации, локализация переменных/eps-костыль — одинаково полезные, с любым одним из этих фиксов решение отрабатывает 70 итераций за 7 секунд, без них — 11 итераций за 1.2 секунды, 12 итераций за 8 секунд, дальше TL.

Если коротко — у меня нету команды для участия в АСМ. И, фактически, никогда не было. Можно считать командой то, что было в самом начале, на первом курсе, в 2010 году — но в тренировочном процессе той команды было множество деталей, которые приводят к мысли, что командой это можно было назвать весьма условно. А после — команды у меня не было. В разное время со мной писали контесты заинтересованные новички (которые забивали на АСМ через неделю), мои одногрупники, знакомые, родственники и вообще какие-то случайные люди. За это время было примерно 8 людей, с которыми повторялась одна и та же история — я вроде бы собираю команду, чтобы готовиться и иметь состав не "на один контест", они вроде бы заинтересованны в занятиях АСМ, я готов помогать им с тренировками и всем остальным, они вроде бы тренируются, через некоторое время (от недели до нескольких месяцев) — они забивают. Некоторые забивали после первого упоминания слова "тренировка", некоторые сами проявляли инициативу, определенное время довольно активно занимались, и забивали уже потом.

Многие из тех, кто мог бы без особых проблем достичь высокого уровня (например, ребята с математической подготовкой на уровень выше моей) — не были заинтересованы в участии в команде по различным причинам: "Я слишком глуп, чем я тебе помогу? Сколько бы я не тренировался, я ничему не научусь", "Ты слишком глуп, я если и буду заниматься этими соревнованиями — то только в команде с кем-то действительно крутым; зачем зря тратить время на неудачников вроде тебя?" (да, было и такое:) ), "это неинтересное занятие", "мне на парах преподаватель рассказывал, что это бесполезная трата времени", ну и так далее.

Команды у меня нету, и она сама по себе не появится в обозримом будущем — ей просто неоткуда появиться. Все, кто имеет хоть какое-то отношение к спортивному программированию — имеют команды, в которых изменения составов не предвидятся. В самом лучшем случае — если приложить достаточно усилий, то получится убедить кого-то на один сезон, без тренировок, а потом гипотетические тиммейты уйдут на full-time работу — и для них трудностью станут не то что тренировки, а даже необходимость прийти на контест (что уж говорить о тренировочных сборах...).

Ближе к делу. У меня осталось еще 3 сезона. И для участия в этих 3 сезонах — нужны тиммейты. И желательно не дюжина различных по отдельности на каждый контест, и не те, кто никогда не слышал о сайте Codeforces и не умеет найти НОД двух чисел (да, таких за последние годы тоже было достаточно). Сейчас я довольно серьезно обсуждаю с одним студентом другого университета его переход в ЛНУ; кроме того, еще с несколькими школьниками обсуждаю вариант ихнего поступления в ЛНУ после окончания 11 класса. Тем не менее, ни один из кандидатов не проявляет особой уверенности в том, что он действительно сделает так, как бы мне хотелось, и с осени станет участником моей команды.

Во-первых, если вдруг есть желающие — пишите в переписку или прямо в эту тему, обсудим. Если речь идет об участнике, который по уровню ощутимо ниже меня — я готов вкладывать в его подготовку столько времени, сколько он сам сможет и захочет на это выделить, и начинать тренировки в онлайн-режиме уже сейчас.

Во-вторых, готов выслушать любые советы и предложения. От тренеров, активистов и просто умных людей. Можно даже и не очень умных, если совет оригинальный и интересный:) Вариант с переходом в КНУ или вообще в какой-то из вузов NEERC/CERC — не предлагать. Я не могу себе позволить это по состоянию здоровья и ряду других причин, а если бы и мог — нету особого желания. Т.е. меня интересует именно то, как собрать команду для выступлений за ЛНУ. Уже несколько раз поднимались темы в стиле "я тренер-активист, как мне собрать команды для участия в соревнованиях?", так вот — здесь почти то же самое, только я не тренер. А немного более заинтересованный практически человек. Да, кстати, все методы, которые описывались в тех темах — мягко говоря, работают плохо. По понятным причинам. Так можно от силы насобирать массовку ради количества участников первого этапа ICPC, но перевести это в качество — вряд ли получится.

Понимаю, что в SEERC вообще мало кто этим занимается, как мне известно — если отбросить Киев, то останется, по различным оценкам, от 0 до 3-4 ВУЗов, где есть хоть одна команда, которая более-менее тренируется. В полном составе тренируется. И не только 2 раза в год на сборах тренируется, а вообще тренируется. И примерно... 1, если не ошибаюсь... ВУЗ, в котором тренировки более-менее регулярны и систематичны, а не вида "что-то делать нечего, давайте что ли тренировку напишем, ведь вроде уже давно не писали". Но ведь все равно как-то выкручиваются?) Так что жду на какие-то советы и предложения.

Сегодня столкнулся с непонятным для меня проседанием по времени на GCC решения, которое нормально заходит под MSC++.

Есть код, в котором что-то примерно такое:

struct point{double x,y;};

vector<pair<point,long > > v;

bool cmp(pair<point, int> a,pair<point, int> b)
{
     if (a.first.x>b.first.x)return true;
     if (a.first.x<b.first.x)return false;
     if (a.first.y>b.first.y)return true;
     return false;
}

int main()
{
sort(v.begin(),v.end(),cmp);
}

Тестирование в "запуске" показало, что при отправке решения под GCC время исполнения превышает время исполнения под студией в 9-12 раз.

Если исправить типы pair<point,long > / pair<point, int> (чтобы пары по описанию точно совпадали в векторе и в сортировке, не важно, будет в обеих случаях int или long), то это ускоряет программу процентов на 30, но все равно время выполнения больше студийного в 6-8 раз.

Кто-то может объяснить мне причины такой ощутимой разницы?

UPD. Если переписать просто на вектор точек (заменить пары на сами точки), то все равно разница сохраняется.

У меня глупый вопрос — как можно открыть свои собственные решения на icpcarchive.ecs.baylor.edu? Я в упор не могу найти там такого функционала. В My Submissions у меня есть только линки на задачу, но не на код. По других разделах сайта полазил — тоже ничего не нашел.

Где оно спрятано? Или это такой уникальный подход к конфиденциальности — и у них нельзя открыть даже свои собственные решения?

Меня устроила бы даже возможность загрузить все файлы одним архивом.

Вопрос: когда же начнется следующий матч?

Не могу понять, если честно. Согласно расписанию, 19:00 по времени сайта, т.е. 17:00 по Киеву. 

Тем не менее, боковая панелька говорит мне, что до начала контеста еще 21:57. С учетом того, что у меня на часах сейчас 20:02, получается, что соревнования все же в 18:00 по Киеву.

Где правда? Какая-то проблема из-за изменений с зимним/летним временем, или я в чем-то туплю?