Hello everyone!
We are back with another Premier round!
Repovive Premier Round 6 will be held this Saturday.
Premier rounds are currently the highest level among our contest series. We hope you enjoy it!
- Start Time: Saturday, 11 July 2026, 14:35 (UTC)
- Duration: 2:30 hours
- Registration: https://repovive.com/contests/16
The contest is rated for everyone.
- Writers: 2sqrt3, ZainUlHaq, RockyYue, Hamed_Ghaffari, _R00T and AmShZ
- Testers: Amoo_Safar, Ali_BBN, Bahamin and mohammadsam
See you on the leaderboard!
UPD1: We are sorry for the issues during the contest.
The website was down for around $$$5$$$ to $$$10$$$ minutes. Also, in Problem E, we forgot to mention the constraint on the sum of $$$k$$$ over test cases. We apologize for these problems.
You can now view the solutions in the Learn tab.
Congratulations to the top participants:
- 1: abc864197532
- 2: Um_nik
- 3: risujiroh
- 4: jeroenodb
- 5: potato167
- 6: Golovanov399
- 7: physics0523
- 8: dezhonger
- 9: WiwiHo
- 10: msab3f
- 11: AmirAli-Asgari
We would also like to thank AmiraliRafiei, Thrb_73, and jigyasu_kalyan for testing the round.
How many problems?
7
Hi again!
Problems are ready now. If you want to test this contest, please send your Codeforces handle and your Repovive registration email to my Telegram account: @H_3014.
For testing, you can either time yourself or solve the problems at your own pace.
All feedback must be sent in Telegram at least 3 hours before the contest starts.
We will update this blog after the contest to include the names of our testers and the top participants.
Good luck ♡ :)
Reminder: About 1 hour to the contest.
plz read my DM(in codeforces) :)
Enjoyed solving problem B.
How can you just forget to write a limitation that is the whole reason this problem even exists?
And is it true that you were not looking at clarifications at all? I think I have checked that nobody answered my clarification after ~40 mins.
And how did people solve it before the clarification came?
Problem F was very nice though.
O(nk) passed with some constant opt, I also sent a clarification but didn't see any popup although it was changed at some point.
Unfortunately, we had a few strong testers, but none of them noticed the missing constraint, and they assumed that the sum of $$$k$$$ is also at most $$$3000$$$. We usually assume the sum of a parameter has the same limit as its regular limit. We are thinking about two solutions to avoid seeing this issue again. Definitely, we will improve our protocol for making sure the problems and statements are correct. Besides, we believe there is a nice way to show these constraints that no website uses. We can show them as a table or something else to make clear whether the limit is on a single test case or over multiple test cases.
About the clarifications, it is not true that we do not check them :). We answer most of them quickly. The reason in your case is that the person who was answering was not sure about the answer to your question. We are going to design a better protocol for clarifications to make sure this never happens again. Thank you for your patience.
I don't have much stake in this fight, but I think the issue with problem E is clearly sufficient reason to unrate this round, and I'm surprised that rating updates were pushed in spite of this error. I imagine the error in the constraints affected many participants; I came up with $$$O( \min(N, K)^2 \log n)$$$ early in the round and assumed this was clearly too slow, leading me to spend the rest of the round coming up with alternate solutions. I refreshed the page about 75m into the round because at that point I suspected that the constraints might be wrong, but at that point no clarification had been pushed. To prevent this issue, I think it's helpful to ask a few testers to carefully read the constraints and make sure they're consistent with what the setters intended. (Maybe have AI do this as well? I imagine LLMs could do a good job of checking that the stated constraints match both the validator and the largest test cases.)
Separately, I don't like that the constraints of this problem leave $$$O(NK)$$$ on the border of passing. My $$$O(NK)$$$ ran in 150% of TL (tested via custom invocation on AtCoder), and according to the other comments, some people were able to get AC with $$$O(NK).$$$
Moreover, when I submitted my $$$O(NK)$$$ solution, the judge gave me a WA verdict, but that verdict has been updated to TLE since then. I'm not sure if this is an issue with the judge or with the test data, but this significantly worsened my contest experience--after seeing WA, I spent quite a while looking for an error in my code (whereas if I saw TLE, I would have decided not to continue attempting to squeeze $$$O(NK)$$$ and would have attempted F and G).
As I mentioned before, unfortunately we had several strong testers, but none of us noticed the missing constraint, and everyone assumed that the sum of $$$k$$$ is also at most $$$3000$$$. During preparation, we were all working with the intended constraint, so we failed to catch that it was not explicitly written in the statement.
We agree that this situation is not acceptable at all. If this happened on CF or a similar established platform, the round would probably be unrated. However, since Repovive is still new and growing, and the rating system has not fully stabilized yet, we decided that it is better to keep the round rated. Also, since the rating changes are mostly positive at this stage, we prefer to let the ratings continue to shape over time. We are also planning to design clearer rules about when a round should become unrated, so that such decisions are not made case by case without a defined policy.
Regarding the WA changing to TLE, we are sorry that this worsened your contest experience. We will investigate this issue as well, since the verdict should not mislead participants during the contest.
We are very sorry about your experience, and we hope to make it up to you in future contests.
Could the WA you saw be from the judging/result popup rather than the submissions page? I noticed there are some places where the site says WA meaning rejected so I'm used to not relying on those and clicking Submissions.
Huh? I was thinking SO hard about how to solve E for stuff like tens of test cases with K and N in the thousands... lol.
very nice contest
I noticed there are some duplicate tests in G, for example
10^5 10^5 11is in 2 tests in a row. Generator bug? Or if the purpose is "test if deterministic", that's better done on judge level and just quantity of different tests should handle correctness anyway.It was a bug, and we'll fix it. Thank you for the notice.
Could you guys please add an option to edit/hide the participant's real name in the profile? For example, mine got pulled from google account and I can't do anything about it. Thanks
Some UX issues I found during the contest:
1) Commentary tab displays the full name, though, handle would be more natural
2) Popups asking for problem feedback during the contest are useless and distracting
3) Problem author info is nice to have, but again, it is a distraction while the contest is running. Consider moving it to editorial or at least hiding under a dropdown
4) Color of TLE is inconsistent between submissions and dashboard
WA instead of TLE on verdict popup in problem E was already mentioned, +1 on that