In 375138399, I got a mysterious Compilation Error with error message

g++.exe: fatal error: cannot execute 'C:/Programs/gcc13-64-winlibs/bin/../lib/gcc/x86_64-w64-mingw32/13.2.0/../../../../x86_64-w64-mingw32/bin/as.exe': CreateProcess: No such file or directory

I submitted basically the same code 3 minutes after and got AC: 375139568

Yes, the existence of an even degree partition can be proved using linear algebra. See, for example, the solution of USAMO 2008/6 here.

There was a youtube/telegram leak.

Actually, I came up with the Souvenirs problem before that problem appeared, and I originally submitted Souvenirs to IOI 2024.

I think that, except for the similar story, they're quite different problems.

I'm the author of Souvenirs.

For Day 1 and Day 2 there are brief explanation slides in Spanish.

Note that time was incorrectly set on Gym (the time at the blog post is the correct one). It has been fixed now.

In fact, the base of the logarithm can be improved to 3. See here.

Edit: sorry, misread the post. The linked problem is about a lower bound for $$$k$$$, this is about an upper bound.

It's similar to this problem 105003C - Equipartition, but a bit harder / more annoying.

I still see it there: https://codeforces.com/gym/105046/attachments/download/24697/editorial_d1.pdf

Great round!

Appreciated the reference to 1656C - Сделай равным по модулю.

You can find the statements in the "Contest materials" panel on the right.

D1C/D2E was googleable: https://www.researchgate.net/publication/231827662_Trees_with_Hamiltonian_square

Participants in the mirror should receive in the following week a link to an upsolving CMS server which will be available for a limited time. After that, the problems will be uploaded to CF Gym at some point.

I think the timeanddate link may not be set properly. It links to 11:00 UTC, not 11:00 UTC+6.

I think this may not actually apply to IOI?

For (low) silver, a good strategy is to go for all the easier subtasks, so it is not wise to spend too much time thinking about the full solution to a problem which seems hard. However, for gold you may actually have to go for 100 points at some problems, so it is probably the right strategy to invest time thinking about full solutions at the risk of not having enough time for scraping subtasks if you end up not finding the solution.

So aiming for gold requires taking risks which may result in not even getting silver, while aiming for silver allows for safer strategies.

Another example of high similarity: 840A - Леха и функция — IMO 1981/2

Thank you. It should be fixed now.

Hello, hope you all enjoyed the problems. Answering some questions:

• There is not a full editorial, but there are some documents here in Spanish explaining very briefly the main ideas behind the solution of the problems, with the code of the official solution.

• Here you can see the standings of the onsite contest if you want to check how you did against the actual contestants. If you participated in the mirror contest, you should have received a copy of the mirror contest standings to your email.

• If you participated in the mirror contest, you also should have access to problems for upsolving. At some point, we will publish them somewhere so everyone can submit.

Problem D is very similar to a problem from Spain's Olympiad in Informatics 2022, which was held just last weekend.

Good luck to everyone participating! You can discuss problems here after the contest.

If constraints were only up to $$$10^{18}$$$, all numbers could be factorized using the following strategy: first, take out all small ($$$\leq 10^6$$$) prime factors, and then each number can have at most $$$2$$$ big prime factors, which if they appear separate then they can be found by calculating gcd with al other numbers, and if they appear together then product of the two factors can be considered as a prime as you said.

Thank you for the detailed feedback. I agree that the statement is somewhat artificial.

Also, I have seen how you have posted those problem-by-problem feedback in each of the rounds and I must say it is very helpful for problem authors.

Hello everybody! Thanks for your feedback!

I would like to invite you to the online mirror of Spain's Olympiad in Informatics which will be held on 2 and 3 April. Check this blog post to see details. Note that if you want to participate, you should register before this Saturday.

If you liked the problems on this round, be sure to participate! And if you didn't — well, there are some additional different problemsetters in the Olympiad other than me, so you should participate anyways.

Cheers!

Please remember that the deadline for registering is this Saturday!

Thank you for your feedback. Could you please elaborate on why you thought F was standard, for example linking to a similar problem?

Such a case was already present in the system tests. However, in that testcase the central vertex is not $$$1$$$, so your solution passes.

Do you know if it is possible to construct a case against your solution if you start from a random vertex instead of vertex $$$1$$$?

Yes. "Statements in English and Spanish".

Auto comment: topic has been updated by FelixMP (previous revision, new revision, compare).

According to webpage: https://swerc.eu/2021/about/ it is on the 23-24 of april.

This is another comment about having received a notification of plagiarism in problem 1566C. I did not share my solution or use online code editors.

Again there are multiple approaches, here is the idea from the official solution:

Here are some hints for one solution (there are multiple possible approaches)

Way too many points given to Div1D1/Div2F1 for its actual difficulty, in my opinion.

Auto comment: topic has been updated by FelixMP (previous revision, new revision, compare).

Some leaders were concerned that the introduction of HMs will affect contestant strategy as they would go for something which is different from maximizing points.

Imagine having a scoring system where solving problems is not the same thing as maximizing points

This post was made by IMO gang

The "task A" situation you describe (a hard task with many points from easy subtasks) should never happen. It pretty much takes away the problem-solving part of the competition for everyone except the ones competing for the very top spots, by failing to reward contestants who choose to spend their time thinking about solving problems instead of thinking about which subtasks give most points. I don't understand why anyone who appreciates IOI being a problem-solving competition would consider it desirable to have this, and yet as you said it happens very often.

Check Anton's On Problemsetting blog (and the sequel).

Wait, is this for real?

Where can we see the individual results?

A similar light blue (#E8F4FF) is already used for "No Medal" (for example here)

On the other hand, it makes sense for it to be similar to "No Medal", since HM is still not a medal.

Seems like "another Nowruz" happened anyways.

You only need to sort the array once if you also keep the original indices of each bottle. Then binary search the answer.

In got WA on 5 by only considering cases where the interval of lucky days of Bob is to the right of the interval of lucky days of Alice (in the optimal case). Not sure if this will help you...

(After considering cases where the interval is to the left I got WA on pretest 8 :( )

What is hack of A?

If one of the player wins for a given position, then the other one loses. Because there is always one player who wins and the other one loses. I don't see why would you need a second dimension.

Unless your state refers to which player starts playing, but I still don't understand why it would be needed since it is specified in the statement that Alice starts playing.

What corner case does test 16 for problem D have? My code passed pretests but gave WA for that one.

Binary searching for the point of tangency. Given a point of tangency, you can compute in O(n) minimum radius and you have that the optimal tangency point can not be farther away from the point that is on the boundary of the minimum circle on the current tangency point.

I also got WA on pretest 4. But I overcame it after increasing the number of iterations of the binary search (unfortunately I got TLE on later tests and didn't have time to optimize the algorithm :( )

edit: nvm, my algorithm was not the same as yours

Note that after each operation, the order of the characters in the string taken cyclically is reversed. But for the zebras the order being reversed doesn't matter, so it suffices to consider the cyclic string.

The phrase "if and only if" is used to declare a coimplicative relationship. That is, ways are considered different if (condition), and only if (condition) they are considered diferent (if (condition) is not satisfied, they aren't different). Its meaning is not related to "only one segment".

I think you need to print output with enough precision.

In problem A my code passed pretests without handling the n=1 case. (Luckily I noticed before getting hacked)

Was Div2C inspired by Problem 2 from International Mathematics Olympiad 1981 ? Probably a coincidence but I was reminded of it just after reading the statement. (Also, not good problem for a programming competition, in my opinion — many people saying that they just guessed the solution).

Did you optimize input/output? I also had the same problem and changing cin/cout to scanf/printf made it pass pretests.

0, since the ratios are already the same.

My solution (apparently correct) was based on Goldbach's conjecture. The answer was always going to be 1, 2 or 3.

Loved the idea behind problem D div. 2 (B div. 1).