I am writing this blog to report a competitive-integrity issue involving the Codeforces user [user:NotOnlySuccess,2026-08-09] and a website associated with their content.↵
↵
The main issue is simple: **problem-specific solution material and hints for an ongoing Codeforces contest were made available before the contest ended.**↵
↵
# 1. Solution material published before the contest ended↵
↵
During [Codeforces Round 1116 (Div. 1 +, Div. 2)](https://codeforces.me/blog/entry/155840668), an article titled [CF Round 1116 (Div. 1+2)](https://mp.weixin.qq.com/s/P3QJqmQFRsLlW9GX_y_8YQ) was published on a WeChat public account at August 10, 2026, 00:37 (UTC +8). The Codeforces contest ended at August 10, 2026, 01:05 (UTC +8).↵
↵
Therefore, the article was publicly available **28 minutes before the contest ended**.↵
↵
↵
↵
The article already contained problem-specific solution material.↵
↵
For readers who do not understand Chinese, here is a direct translation of the relevant part of the article for **Div. 2A — Three Numbers on the Blackboard**:↵
↵
> **Greedy**↵
>↵
> - In one operation, one number is replaced by the sum of the other two numbers. What relationship do the three numbers satisfy after the operation?↵
> - Suppose the two numbers that are kept are $u \le v$. Then the new triple is $(u, v, u+v)$. Therefore, the largest number is exactly equal to the sum of the other two numbers. After the first operation, the state becomes a degenerate triangle.↵
>↵
> - Suppose the current triple is $p \le q \le r$ and $r \le p+q$. What is the new range after replacing $p$, $q$, or $r$, respectively?↵
> - The three new ranges are $r$, $r$, and $q$, respectively. The original range is $r-p$, and $q \ge r-p$. Therefore, after entering the triangle state, no subsequent operation can reduce the range.↵
>↵
> - Since subsequent operations cannot further improve the answer, which candidates need to be compared in the initial state?↵
> - Compare $z-x$, obtained by doing nothing, with $z$, $z$, and $y$, obtained by replacing each of the three positions in the first operation. Therefore, after sorting the numbers as $x \le y \le z$, the answer is $\min(z-x, y)$.↵
↵
This is substantially more than a generic hint such as "try greedy". It gives the key structural observation, proves why further operations cannot improve the answer, and derives the final formula, all while the official contest was still running.↵
↵
Regardless of whether one calls this a "hint", an "editorial", or something else, this is problem-specific reasoning that can materially help a contestant solve the problem.↵
↵
The important point is that it was available **while the official Codeforces contest was still running**.↵
↵
# 2. This was not an isolated incident↵
↵
The early publication of the Codeforces Round 1116 material does not appear to be an isolated case.↵
↵
There are also earlier examples of solution material being published before other programming contests had ended. In particular, solution articles for AtCoder Beginner Contests had previously been published while the corresponding contests were still in progress.↵
↵
For example, [this article](https://mp.weixin.qq.com/s/T4oFYbFTb3e8AUJlKqLeSQ) was published before the corresponding AtCoder Beginner Contest had ended.↵
↵
↵
↵
The article was published at July 11, 2026, 21:04 (UTC +8). The official contest ended at July 11, 2026, 21:40 (UTC +8).↵
↵
More importantly, the material in this case went beyond heuristic hints or high-level observations: it directly included **accepted source code** for contest problems while the contest was still running.↵
↵
This makes the issue materially more serious than an accidentally premature publication of a single editorial. At minimum, it shows that the Codeforces Round 1116 incident should not be considered in isolation, but as part of a repeated pattern of publishing contest-specific solution material before the corresponding contests had ended.↵
↵
# 3. Why I associate this content with [user:NotOnlySuccess,2026-08-09]↵
↵
Another article on the same WeChat public account, titled roughly [How to Learn Algorithms in a Planned Way?](https://mp.weixin.qq.com/s/EZ9pyFOrrgW6ON4aYp2rLw), displays the names:↵
↵
↵
↵
This is why I associate the account and its content with the Codeforces user [user:NotOnlySuccess,2026-08-09].↵
↵
# 4. A mirror-contest website providing hints during official contests↵
↵
The content also links to a website called [Shortest Path](https://shortestpath.cn/mirrors), which provides what it describes as "mirror contests".↵
↵
The website states that it follows official AtCoder and Codeforces schedules and provides Chinese statements and independent judging.↵
↵
More importantly, it explicitly advertises **contest-time guidance**.↵
↵
Its description includes:↵
↵
> During the contest, heuristic hints are progressively unlocked.↵
↵
It also describes the feature roughly as:↵
↵
> When you are stuck, continue thinking along the questions. The hints first use questions to help locate where you are stuck, and then you decide whether to view further information.↵
↵
↵
↵
The mirror contest was scheduled to start only five minutes after the official contest started.↵
↵
The hints were not necessarily available immediately. They were unlocked after waiting for a certain amount of time.↵
↵
However, this does not remove the competitive-integrity issue.↵
↵
A participant in the official Codeforces contest can open the mirror contest, wait until a hint becomes available, and then use that information while the official contest is still in progress.↵
↵
Here is an example of the hint interface:↵
↵
↵
↵
The hints shown there also appear to correspond closely to the reasoning later published in the WeChat article.↵
↵
In other words, the website appears to make parts of the solution reasoning available during the official contest, before the corresponding post-contest explanation is published.↵
↵
# 5. Why this is a serious problem↵
↵
There is an important difference between:↵
↵
- publishing editorials, hints, explanations, or solutions **after a contest has ended**, and↵
- providing problem-specific assistance **while the official contest is still running**.↵
↵
The first is completely normal and useful.↵
↵
The second can provide an unfair advantage to contestants who use it.↵
↵
A hint does not need to contain complete code or the entire solution to be valuable.↵
↵
For competitive programming problems, revealing a critical observation, invariant, greedy choice, reduction, construction, or intended data structure may be enough to substantially reduce the difficulty of the problem.↵
↵
Calling such information a "heuristic hint" does not change its effect on an ongoing contest.↵
↵
If the information is specifically designed for the current contest problem and is accessible before the official contest ends, contestants can use it as external assistance.↵
↵
# 6. What I am NOT claiming↵
↵
I want to make the scope of this report explicit.↵
↵
**I am not claiming that I can prove that [user:NotOnlySuccess,2026-08-09] personally cheated in Codeforces contests.**↵
↵
I do not have sufficient evidence to make that accusation.↵
↵
Similarly, I do not consider an unusual rating graph or rapid rating improvement, by itself, to be proof of cheating.↵
↵
That is not the point of this blog.↵
↵
The issue I am reporting is directly observable:↵
↵
**problem-specific assistance for an ongoing Codeforces contest was made available to other people before the contest ended.**↵
↵
Whether we can identify a particular contestant who actually used the assistance is a separate question.↵
↵
In my opinion, deliberately providing such assistance during a rated contest is itself a serious competitive-integrity issue.↵
↵
# 7. Request to Codeforces↵
↵
I would appreciate it if the Codeforces team could review this situation and determine whether providing problem-specific hints during an ongoing Codeforces contest violates Codeforces rules or competitive-integrity policies.↵
↵
In particular, I believe the following should be reviewed:↵
↵
1. the publication of solution material before Codeforces Round 1116 ended;↵
2. the mirror-contest system that makes problem-specific hints available while the corresponding official contest is still running;↵
3. the relationship between the website, the WeChat account, and the Codeforces account [user:NotOnlySuccess,2026-08-09].↵
↵
I am providing the screenshots and links so that the evidence can be independently checked.↵
↵
# 8. Request to [user:NotOnlySuccess,2026-08-09]↵
↵
To [user:NotOnlySuccess,2026-08-09]:↵
↵
The purpose of this blog is not to start harassment against you or to damage your reputation beyond what the evidence supports.↵
↵
However, I strongly ask you to:↵
↵
- stop publishing solution material before the corresponding Codeforces contest has ended;↵
- disable any feature that provides problem-specific hints while the corresponding official contest is still running.↵
↵
Mirror contests, editorials, hints, solution explanations, and training resources can all be valuable educational tools.↵
↵
There is no problem with providing them **after the official contest ends**.↵
↵
The problem is making information that materially helps solve the current contest problems available **during the contest itself**.
↵
The main issue is simple: **problem-specific solution material and hints for an ongoing Codeforces contest were made available before the contest ended.**↵
↵
# 1. Solution material published before the contest ended↵
↵
During [Codeforces Round 1116 (Div. 1
↵
Therefore, the article was publicly available **28 minutes before the contest ended**.↵
↵
↵
↵
The article already contained problem-specific solution material.↵
↵
For readers who do not understand Chinese, here is a direct translation of the relevant part of the article for **Div. 2A — Three Numbers on the Blackboard**:↵
↵
> **Greedy**↵
>↵
> - In one operation, one number is replaced by the sum of the other two numbers. What relationship do the three numbers satisfy after the operation?↵
> - Suppose the two numbers that are kept are $u \le v$. Then the new triple is $(u, v, u+v)$. Therefore, the largest number is exactly equal to the sum of the other two numbers. After the first operation, the state becomes a degenerate triangle.↵
>↵
> - Suppose the current triple is $p \le q \le r$ and $r \le p+q$. What is the new range after replacing $p$, $q$, or $r$, respectively?↵
> - The three new ranges are $r$, $r$, and $q$, respectively. The original range is $r-p$, and $q \ge r-p$. Therefore, after entering the triangle state, no subsequent operation can reduce the range.↵
>↵
> - Since subsequent operations cannot further improve the answer, which candidates need to be compared in the initial state?↵
> - Compare $z-x$, obtained by doing nothing, with $z$, $z$, and $y$, obtained by replacing each of the three positions in the first operation. Therefore, after sorting the numbers as $x \le y \le z$, the answer is $\min(z-x, y)$.↵
↵
This is substantially more than a generic hint such as "try greedy". It gives the key structural observation, proves why further operations cannot improve the answer, and derives the final formula, all while the official contest was still running.↵
↵
Regardless of whether one calls this a "hint", an "editorial", or something else, this is problem-specific reasoning that can materially help a contestant solve the problem.↵
↵
The important point is that it was available **while the official Codeforces contest was still running**.↵
↵
# 2. This was not an isolated incident↵
↵
The early publication of the Codeforces Round 1116 material does not appear to be an isolated case.↵
↵
There are also earlier examples of solution material being published before other programming contests had ended. In particular, solution articles for AtCoder Beginner Contests had previously been published while the corresponding contests were still in progress.↵
↵
For example, [this article](https://mp.weixin.qq.com/s/T4oFYbFTb3e8AUJlKqLeSQ) was published before the corresponding AtCoder Beginner Contest had ended.↵
↵
↵
↵
The article was published at July 11, 2026, 21:04 (UTC +8). The official contest ended at July 11, 2026, 21:40 (UTC +8).↵
↵
More importantly, the material in this case went beyond heuristic hints or high-level observations: it directly included **accepted source code** for contest problems while the contest was still running.↵
↵
This makes the issue materially more serious than an accidentally premature publication of a single editorial. At minimum, it shows that the Codeforces Round 1116 incident should not be considered in isolation, but as part of a repeated pattern of publishing contest-specific solution material before the corresponding contests had ended.↵
↵
# 3. Why I associate this content with [user:NotOnlySuccess,2026-08-09]↵
↵
Another article on the same WeChat public account, titled roughly [How to Learn Algorithms in a Planned Way?](https://mp.weixin.qq.com/s/EZ9pyFOrrgW6ON4aYp2rLw), displays the names:↵
↵
↵
↵
This is why I associate the account and its content with the Codeforces user [user:NotOnlySuccess,2026-08-09].↵
↵
# 4. A mirror-contest website providing hints during official contests↵
↵
The content also links to a website called [Shortest Path](https://shortestpath.cn/mirrors), which provides what it describes as "mirror contests".↵
↵
The website states that it follows official AtCoder and Codeforces schedules and provides Chinese statements and independent judging.↵
↵
More importantly, it explicitly advertises **contest-time guidance**.↵
↵
Its description includes:↵
↵
> During the contest, heuristic hints are progressively unlocked.↵
↵
It also describes the feature roughly as:↵
↵
> When you are stuck, continue thinking along the questions. The hints first use questions to help locate where you are stuck, and then you decide whether to view further information.↵
↵
↵
↵
The mirror contest was scheduled to start only five minutes after the official contest started.↵
↵
The hints were not necessarily available immediately. They were unlocked after waiting for a certain amount of time.↵
↵
However, this does not remove the competitive-integrity issue.↵
↵
A participant in the official Codeforces contest can open the mirror contest, wait until a hint becomes available, and then use that information while the official contest is still in progress.↵
↵
Here is an example of the hint interface:↵
↵
↵
↵
The hints shown there also appear to correspond closely to the reasoning later published in the WeChat article.↵
↵
In other words, the website appears to make parts of the solution reasoning available during the official contest, before the corresponding post-contest explanation is published.↵
↵
# 5. Why this is a serious problem↵
↵
There is an important difference between:↵
↵
- publishing editorials, hints, explanations, or solutions **after a contest has ended**, and↵
- providing problem-specific assistance **while the official contest is still running**.↵
↵
The first is completely normal and useful.↵
↵
The second can provide an unfair advantage to contestants who use it.↵
↵
A hint does not need to contain complete code or the entire solution to be valuable.↵
↵
For competitive programming problems, revealing a critical observation, invariant, greedy choice, reduction, construction, or intended data structure may be enough to substantially reduce the difficulty of the problem.↵
↵
Calling such information a "heuristic hint" does not change its effect on an ongoing contest.↵
↵
If the information is specifically designed for the current contest problem and is accessible before the official contest ends, contestants can use it as external assistance.↵
↵
# 6. What I am NOT claiming↵
↵
I want to make the scope of this report explicit.↵
↵
**I am not claiming that I can prove that [user:NotOnlySuccess,2026-08-09] personally cheated in Codeforces contests.**↵
↵
I do not have sufficient evidence to make that accusation.↵
↵
Similarly, I do not consider an unusual rating graph or rapid rating improvement, by itself, to be proof of cheating.↵
↵
That is not the point of this blog.↵
↵
The issue I am reporting is directly observable:↵
↵
**problem-specific assistance for an ongoing Codeforces contest was made available to other people before the contest ended.**↵
↵
Whether we can identify a particular contestant who actually used the assistance is a separate question.↵
↵
In my opinion, deliberately providing such assistance during a rated contest is itself a serious competitive-integrity issue.↵
↵
# 7. Request to Codeforces↵
↵
I would appreciate it if the Codeforces team could review this situation and determine whether providing problem-specific hints during an ongoing Codeforces contest violates Codeforces rules or competitive-integrity policies.↵
↵
In particular, I believe the following should be reviewed:↵
↵
1. the publication of solution material before Codeforces Round 1116 ended;↵
2. the mirror-contest system that makes problem-specific hints available while the corresponding official contest is still running;↵
3. the relationship between the website, the WeChat account, and the Codeforces account [user:NotOnlySuccess,2026-08-09].↵
↵
I am providing the screenshots and links so that the evidence can be independently checked.↵
↵
# 8. Request to [user:NotOnlySuccess,2026-08-09]↵
↵
To [user:NotOnlySuccess,2026-08-09]:↵
↵
The purpose of this blog is not to start harassment against you or to damage your reputation beyond what the evidence supports.↵
↵
However, I strongly ask you to:↵
↵
- stop publishing solution material before the corresponding Codeforces contest has ended;↵
- disable any feature that provides problem-specific hints while the corresponding official contest is still running.↵
↵
Mirror contests, editorials, hints, solution explanations, and training resources can all be valuable educational tools.↵
↵
There is no problem with providing them **after the official contest ends**.↵
↵
The problem is making information that materially helps solve the current contest problems available **during the contest itself**.
