Hi Codeforces!
The 27th Central-European Olympiad in Informatics takes place from 23-29 August in Nagykanizsa, Hungary. The mirror contests will be held on 26.08.2020 15:05 (Московское время) and 28.08.2020 15:05 (Московское время) (one day after each official contest day) here on Codeforces in the usual IOI format (it is unrated, the best scoring submission counts, scoreboard is not visible during the contest and there's full feedback throughout the whole contest).
In this year's CEOI, the 4 best problem solvers going to secondary schools of Austria, Croatia, Czech Republic, Germany, Hungary, Italy, Lithuania, Poland, Romania, Slovakia, Slovenia, Switzerland, and The Netherlands will compete. Due to the current situation, unfortunately not everyone will be able to make it onsite, but we hope the offsite participants will also enjoy the event, although without the magical week in Nagykanizsa.
Each day will feature a 5-hour contest. In both contests, you will be given 3 problems with a maximum of 100 points to score on each. Every task has subtasks of varying difficulties, so we hope everyone from newbie to legendary grandmaster will have fun solving the problems (well, legendary grandmasters might want to refer to the instructions here).
We would like to thank MikeMirzayanov for creating Codeforces and Polygon platforms and his invaluable help in making this mirror contest possible, and all of the organizers and sponsors.
We wish you will enjoy the contest just as much as we enjoyed preparing it!
CEOI 2020 Scientific Committee
Gyula Horváth, Attila Gáspár, birka0, busamate, csandras, Emarci15, leanil, mraron, lacito, uno20001
UPD1: The tutorial for day 1 is here.
UPD2: The tutorial for day 2 is here.
Sorry for the issues with the queue and with the bug interactor of potion causing python solutions to fail. Hope you still enjoyed the problems!
Congratulations to everyone who participated, here are the top 5 combined scores:
The rest of the combined scores can be found here.
You can also check how would you fare against the official contestants on this ranklist.
Auto comment: topic has been updated by mraron (previous revision, new revision, compare).
Btw Indian TST tie breaker = this mirror. We'll have the legendary battle between SleepyShashwat and RestingRajarshi
That doesn't sound safe at all, especially when contestants know that the contest took place yesterday.
India doesn't do "safe".
They don't have any contact in Europe. So, it will work.
Results?
SleepyShashwat $$$175$$$
RestingRajarshi $$$160$$$
One day or two days combined? (If one day,which?
Those were day 1 results.
Ustani bane, Hrvatska te zove!
Thanks for this Contest,I'll try my best.
Can I participate in this competition
Yes of course, this mirror is open to everyone except those who already know the tasks.
Why hide standings for whole time , make it visible for first 3 hour and freeze standings in last two hours.
Standing serves as inspiration to compete whole 5 hours. Making it not visible will reduce excitement and fun of problem solving.
While the topic of the IOI and scoreboards is an interesting one, we did not want to change the format too drastically from the official one.
What kind of difficulty level should i expect?? If possible plz ans like div 1 or in codeforces rating...
The contest was prepared for the brightest high school students in Central Europe. I would say div 1, but you have much more time than a div 1 contest.
Ok thanks
Yes and there are subtasks as well :)
How do our final score being calculated? By the sum of max score for each subtask or max score for the problem?
Max score for the problem
for a submission i get 55 point but it doesn't say why i didn't get rest of it, is this normal???
Click on the Verdict..
It would be much better if you added subtask number in the beginning of input file, when importing to Codeforces.
It's not that hard to separate subtasks. And it's also not common for CEOI.
Where is test case 2 on subtask 1 (sample) in problem A? it says TLE on subtask 1 test case 2?
I got it, I didn't see the downloadable test cases, I was seeing only the ones present on the website.
read the announcements
Well-prepared tasks. Will there be an editorial?
how to solve B
There is a well-known algorithm to check if $$$n$$$ segments don't intersect. Sweep from left to right, maintain a set of segments that are intersected by a current sweep line, sorted by $$$y$$$ of the intersection. When a new segment $$$B$$$ comes and it's inserted between segments $$$A$$$ and $$$C$$$ in the set, you can connect the start of $$$B$$$ with the most recent (so, rightmost) point from area between $$$A$$$ and $$$C$$$. We can find such point if for every segment $$$s$$$ we additionally find the most recent point below it and above it (that is, neighboring with $$$s$$$).
My screencast, full score in 3h20m https://youtu.be/Tjv78ZThV5c
I liked the problems, especially B. The subtasks were also well chosen and the official leaderboard shows that. The only annoying thing was the implementation of all-direction dp in C.
Will you do a screencast for day2 ?
Yes, the recording will be up in Youtube in 1-2 hours because of processing time.
I haven't solved C though.
EDIT, here's the screencast link https://youtu.be/tYuikvBoV40 and btw. I'm streaming 1v1 lockout round tomorrow in the morning (link)
Can anyone post an editorial for this contest or can anybody link a blog that contains the solutions to these questions
Bruh it's in announcement.
Yeah I just saw it now
The limits on the actual contest were:
A: 0.1 s, 32 MB
B: 0.3 s, 32 MB
C: 0.2 s, 32 MB
It's right, but we had stronger computers. Codeforces time limits were also based on the running times of our model solutions.
I remember benchmarking on practice session and seeing that modulos are about as fast there and on codeforces, though maybe it's possible that there's a more visible difference in actual solutions. But still, having stronger computers doesn't change memory consumption(I think memory was mainly a problem on C).
You had computers that use parallel universes to let programs with larger memory requirements run in the 32 MB limit? Because that kind of limit is fucking insane in $CURRENT_YEAR.
I was on CEOI when it was also in Hungary. It had stupid limits like "you can't ever read the input for this practice problem in time". You're still doing this shit.
I totally agree with you, that's why I modified the tasks' limits in the first place to be at least 1sec/256mb for the cf mirror, these very low limits are just not fun for the participants and 90% of the time they don't add anything to the strength of the tests in a sense that they don't catch more bad solutions just more slightly less good ones :/
I am with you on the content of your comment, but the way you say it is offensive IMO. "You're still doing this shit" after the sentence about Hungary feels like you associate this with our country, which hurts me because I suggested higher limits, but it was downvoted because the people leading competition development here are stuck 25 years ago. Of course, you didn't know this, but you could have said "Some people there still haven't changed" — skipping the bad language. I regret that I couldn't push further, and I definitely will next time. (We are hosting IOI 2023...) At least we could set the limits here on Codeforces with mraron
be thankful it's not Um_nik responding :D
You don't like it? Good. It means I managed to communicate how it pisses me off to see shit limits there again.
Yeah as busamate said.
I'm also personally not a fan of very tight limits and wouldn't want block something like conqueror_of_tourist's geniosity python solutions :D
Why I can't read others' code?
When will be the test data open?