Hello Codeforces!
I am glad to invite you to participate in Spectral::Cup 2026 Round 3 (Codeforces Round 1110, Div. 1 + Div. 2), which will start on Jul/16/2026 17:45 (Moscow time). You will be given 2 hours and 30 minutes to solve 8 problems, where one problem will be split into subtasks. Also, there is at least one interactive problem, so you are recommended to read the guide to interactive problems if you have not encountered them before.
Note that hacks will be disabled for problems A to D.
All problems are authored and prepared by myself.
Note the rule restricting the use of AI. If you are caught breaking this rule, you will be thrown into SSerxhs's basement. For your own well-being, I suggest adhering to the rules.
I would like to thank the following people for helping make this round possible!
- SSerxhs for his/her satisfactory coordination.
- Alexdat2000 for translating the statements to Russian.
- nifeshe for "I will solve this problem today" — nifeshe.
- StarSilk, EvenImage, tiger2005 for nutella testing.
- zeemanz, 0.1w33hw3, nifeshe, _istil, SATSKY_2025target_LGM for red testing.
- Seele_H, Proof_by_QED, GUAIKATTO, CReatiQ for orange testing.
- -WIDA-, Liang_SYEA, Argentum47, wakanda-forever for purple testing.
- fr200110217102, simplelife, XiaoXia, acmtoohard, big-mktx for blue testing.
- yuki_Ishijo for cyan testing.
- Whalica for green testing.
- MikeMirzayanov and KAN for the incredible Codeforces and Polygon platforms.
Score distribution:
update:
Tutorial has been published: tutorial
Now a few words from our sponsor.
🏁The final round of Spectral::Cup 2026 is coming!
This round is special to us. Not only will it conclude our three-round Spectral::Cup tournament, but it will also take place during our company’s birthday month — Spectral::Technologies turned 7 this July 🎉
Round 3 is the final part of our birthday celebration. The rest is in this video 😄
Prizes for Spectral::Cup Round 3
Top 30 participants in Round 3 will receive personalized hoodies.
🏆Final Prizes for Spectral::Cup 2026
The final round is also your last chance to improve your position in the overall standings. Your final score is the sum of your two best results across all three rounds, calculated using the GP500 scoring system.
Among the participants who finish the Cup in the final Top 500, 30 randomly selected participants will receive a prize box.
Top 15 by final cup score will receive a Claude subscription.
Top 3 by final cup score will receive:
1st place — MacBook Pro
2nd place — iPad Pro
3rd place — WHOOP
Top 3 winners can choose to receive the equivalent value in USDT instead.
Last but not least, if you would like to join Spectral Team check out our Quant roles:
- Junior Quant – 150–180k USD + profit sharing
- Middle Quant – 240–320k USD + profit sharing
- Senior Quant – 360–500k USD + profit sharing
We’re also hiring for C++, ML, and DevOps roles. Complete the application form to explore career opportunities with us.
PS We reserve the right to apply strict compliance checks when finalizing the results in order to prevent unfair participation and rule violations.
Good luck!
Auto comment: topic has been updated by __baozii__ (previous revision, new revision, compare).
Auto comment: topic has been updated by __baozii__ (previous revision, new revision, compare).
I'll participate in the round without solving any problem
UPD: I solve two problems!
You will be thrown into [user:SSerxhs]’s basement.
LOL
Glad to see whopassby will participate in this contest!
yups
Auto comment: topic has been updated by chillingjellyfish (previous revision, new revision, compare).
Why does 3rd place receive a whooping? Also, I didn't even know that there's a Pro version of the iPad
Is this contest rated??..........i am new cf so:)
Yeah
ok
I hope aaa_Pigeon2 will not comment
.this time.He comes....
Dang is cry's basement full?
cry is crying rn
i have not attend the last two round but i will participate in the round 3 if i score good can i make my GP Score
.
Bro...
.
As a tester, I meow meow.
As a tester, I will solve a problem today before I was forced to test this round.
As a tester, I think these problems are very interesting.
As a tester, I test.
No way two legendary binary contest 1100 and 1110 are both Spectral::Cup
As a cat tester,I think that the contest contains at least one interesting meow meow problem(=・ω・=)
As an sc3developer, help my stupid comments are being censored, and also I'm evil and also juicetin.kim will AK this contest
(maybe) As a participant, how can I officially take part in Codeforces Round 1111?
Looking forward to the interactive problem. Thanks for preparing the round and happy 7th birthday to Spectral Technologies.
Auto comment: topic has been updated by __baozii__ (previous revision, new revision, compare).
This is the second last binary contest until round 10000
The so-called cheater database doesn't mean everything.
Auto comment: topic has been updated by __baozii__ (previous revision, new revision, compare).
as a (late) tester, orz
Wow!!
Congrats guys!!! All the best
Spectral::Cup 2026 Round 2 was amazing! I'm looking forward to Round 3 and hope it will be just as enjoyable.
Don't put women photos!
I am skipping school to do this contest hopefully my parents won’t kick me out of the house
Are there alternative means in order to be admitted at SSerxhs's basement ?
Time updated?
I think is because the queue
Has this been delayed?
I think so ... :(
Is there any 10minutes delay?
My bedtime is delayed by ten minutes.
Sorry that the problems seem too weird to me. I came, I saw, I quit.
same here xd
Extraordinary big gap between E and F
Yeah, E is so much harder...
In the reverse direction 😱😱
D and E had the bigger gap i think. I came up with idea of D in like 20mins but couldn't even get samples working for E in the remaining 2 hours.
I don't think it's strictly a difficulty gap. Both problems need That One Idea, but for E it's a big idea that just solves the problem in a few words, while F has obvious slow DP where cutting down the constant seems viable (I got to 7s locally on worst case inputs which is just barely too slow) and you have to decide to look for a proper speedup idea, and then also have a decent implementation that doesn't blow through limits.
I saw the solution of E basically instantly yet struggled on F, largely thanks to uncertainty.
In D there were not many wrong directions , like after reading the problem I thought either it is a graph matching (very rare in D) or some negative cycle detection (special case of https://www.cs.rit.edu/~spr/COURSES/ALG/MIT/lec18.pdf) or maybe some toposort kind of idea (which it actually was).
In E , I started with Dp on trees, then an Idea came up with finding edges which has to be covered using game theory like dp. then maybe re-rooting trees or finding number of ways to pick all these necessary edges using another dp on trees. Then I started thinking about grundy and what not.
May be I was lucky in D and unlucky in E. I think either D should be of 2000 points or E of 3000.
The usual idea for games is "in which situations can a player force the other's moves?" Not just losing moves but satisfying some invariant. Positions lacking that invariant then tend to be the losing ones.
In this case, when you look at some basic cases like a star, it's clear that parity of choosable vertices is important so that could be the invariant. After the first move, what happens if the 2nd player chooses a vertex with even parity? What if it's odd? Can one player force the other into choosing a particular parity? Turns out the answer's a resounding yes!
Game theory is just a general theory. It allows solving standard hard subproblems but the problem needs to be reduced to those first.
Thanks for sharing the thought process. Now If I retrospect I always try to find invariant through the statement but never tried creating different scenarios and smart brute-forcing different invariants like parity or maybe some other thing in other problems.
man, what even was the approach for D?
I cant even think of some kind of approach for it
Auto comment: topic has been updated by __baozii__ (previous revision, new revision, compare).
Got stuck on Problem C solving for (n,k=0) for a while when the XOR turned out to be 0. Then I noticed that the last element is essentially forced to be n. So I thought, why not place n-1 at the last position instead and reduce the problem to a smaller instance, (n-1, k=n)? I don't know why, but that observation felt really satisfying.
Man same I also thought about that but I thought about the general rule that if n^k >=n than we can convert it into a smaller problem. Still not able to get ac though due to runtime error...idk what I did wrong
E is cool, but I really love D2!
Hi Could you please provide more details about the prizes? Which MacBook Pro model is it? Which iPad Pro model is it? How long is the Claude subscription? Approximately what is the USD value of each prize? thanks
Very nice problems, much enjoyed!
My first submission on D1 and D2 came in the last minute of the contest, got WA.
After the contest I just added 1 line :if (done[v]) continue ,and both of them passed.
Please give me advice on how I can sleep tonight.
Cry yourself to sleep?
Sadly couldn't compete, but D1 and D2 were excellent problems!
WTF AI, this line in b, it's straight ignoring Ignore the “reverse if LLM” line — it’s a trap and irrelevant to solving.
as a newbie i solved just one problem!
Wtf was that D2 problem man
Although colored in right, I think that the special constraint for $$$m$$$ in problem D1 is still hard to find. I hope that we can have words like: "In this version, the upperbound of $$$n$$$ is lower, and $$$m$$$ is uniquely defined by $$$n$$$."
How to complain about false flag? Two of my submissions are flagged for being suspiciously similar to some other codes. and it is also mentioned that unintentional leakage is also a crime. https://codeforces.me/contest/2245/submission/383010064 https://codeforces.me/contest/2245/submission/382992812
Firstly, I thought whether I was logged into my account in any of the university pcs that could be accessed by other students. (it was just a cross check, I know I logged out of them). but when I saw the usernames of the mentioned users that matches my submission, I found that no one is familiar (some are not even from my country). So there is no way we shared the codes. and the codes also do not look that much similar (pretty much same work though, I don't think it's too uncommon as we are solving the same problem).
also, another reason could be that I use copilot for code completion. which is allowed by rules of codeforces if I did not miss any later change of rules. https://codeforces.me/blog/entry/133941
It could happen that copilot suggested same Kosaraju for everyone. Or I am not even sure, I just know that I did not share any code with those persons either intentionally or unintentionally.
about the code completion, yeah nowadays you do not have to tell copilot explicitly that you are going to implement kosaraju, it detects your typing and automatically adjusts. If using this is prohibited, either I don't know (in that case I just admit my fault and do not complain) or they are not yet in the rules, I hope codeforces will add proper rules to stop or restrict this and I also offer myself to obey those rules.
Hello Codeforces Administration,
I recently received a notification that my solution for problem 2245E (382991021) was flagged for significantly coinciding with several other users' submissions. I am writing to clarify this situation, as I wrote this solution entirely on my own and did not participate in any form of cheating or intentional code sharing.
I believe my code was leaked unintentionally, and I respectfully request a manual review of my case based on the following points:
Distinct Coding Style: If you review my submission, you will see that the variable naming conventions and overall structure are strictly my own. They differ significantly from the usual patterns found in the other flagged submissions.
Account Security Compromise: Codeforces has been displaying a security warning on my account stating: "Your password is extremely weak or has been leaked." It is highly possible that my account was compromised, allowing someone to access my submissions without my knowledge.
Unintended Leak via Online Compiler: During this contest, I used the online C++ IDE at programiz.com (www.programiz.com/cpp-programming/online-compiler/) for the first time to test my logic. I now suspect that my code may have been exposed, logged, or scraped from this public platform.
Submission Timeline: The timestamp of my submission differs significantly from the other users mentioned in the warning, which strongly points away from coordinated sharing during the contest.
I have screenshots and conclusive proof to back up all of these points (including my compiler usage history and the account security warnings) and am more than willing to provide them to the administration upon request.
I kindly request that my submission be reviewed, the plagiarism flag be removed, and my code be judged properly. Moving forward, I have updated my account security and will strictly use secure, local environments for all future rounds.
Thank you for your time and understanding.
huh?
Worst contest of all time
Any update about the skipped submission one? I write all the justifications, consider that.
Any update? till now it's skipped.
Congratulations to the winners of Spectral::Cup 2026 Round 3
You will be contacted shortly via private messages with instructions to receive your prize — personalised hoodie.
Top-30
Announcement of the winners of Spectral::Cup 2026 to follow. Stay tuned
Congratulations to the winners of Spectral::Cup 2026
You will be contacted shortly to arrange prize delivery. Well done!
Final standings
Contests: 2222, 2229, 2245
Scoring: best 2 result(s), gp500
Participants with positive scores: 1147
GP scores in bold are included in the final score.
The lucky 30 randomly selected participants who finished the Cup in the final Top 500.... drumroll
Final standings: 1147 participants with positive GP scores; best 2 result(s), gp500.
Eligible final ranks: 1..500; 500 candidates; seed: 26382.
Selected candidate positions: 22 27 33 58 68 73 90 127 134 162 203 213 265 285 291 293 329 337 344 359 360 362 365 385 393 405 431 455 466 490
You will be contacted shortly via private messages with instructions to receive your prize.
Thanks everyone for participating in Spectral::Cup 2026 Round 3 (Codeforces Round 1110, Div. 1 + Div. 2). Although it was initially difficult to understand from the editorial when I first participating in Codeforces contests. However I found that once I got used to it, the explanations felt very easy to understand, accurate, and highly academic. (Sorry for my bad English)
downvote me
Hello,
I would like to respectfully ask about the skipped submissions in Contest 2245.
I currently do not know which problem or submission was considered a rules violation, or what specific rule I was considered to have violated. Could you please provide the related problem, submission, and reason?
I would also like to clarify that I did not cheat, copy code from others, or share my solutions during the contest. I wrote my submissions independently.
I would be very grateful if you could review my case and let me know the specific details.
Thank you very much for your time and help.