In the past 2 blogs, I exposed 2 suspected cheaters: MIDORIYA_ and batats. This blog is a complete summary of all the evidence and suspicious details collected so far.
Previous blogs:
1. batats
The main reason this account became suspicious is the extremely abnormal improvement pattern.
He registered on Codeforces around 5 years ago and stayed at newbie level for a very long time. Before the suspicious contests, his rating had never exceeded roughly 1200, and in most contests he could only solve around 2-3 easy problems.
For example, even after years on the platform, he still performed poorly in many Div2/Div3 contests and struggled with standard implementation and greedy problems. His old submissions and contest history clearly showed a relatively weak level.
Then suddenly, within only about 2 months, his performance changed completely.
He jumped from around 1100 rating to over 2100 rating in only a few contests. In that period, he achieved performances that were completely inconsistent with everything previously shown on the account.
Some suspicious points:
- Solving 6/6 in Div2 contests shortly after struggling with basic Div3 problems.
- Huge rating gain in an extremely short time without gradual progression.
- No visible transition period between low-level and expert-level performance.
- Contest performances suddenly becoming far stronger than his long-term practice history suggested.
One of the most suspicious contests was Codeforces Round 1099, where he achieved rank 2 and gained a massive amount of rating.
Another suspicious detail is that many of the strongest performances happened in contests where multiple suspicious accounts also overperformed.
There are also several skipped contests during the improvement period. The timeline becomes even stranger when looking at the exact sequence:
- Weak performances for years.
- Several skipped contests.
- Suddenly returns with expert/master-level performances almost immediately.
If someone improves naturally, there is usually evidence such as:
- consistent upsolving,
- large training activity,
- strong gym history,
- active problem solving outside contests,
- gradual rating growth.
But here, the jump was almost instantaneous.
Some defenses claimed that he had strong ICPC/ECPC/EOI background experience. However, several of those claims appear misleading or exaggerated.
For example:
- This is reportedly his first university year, so claims about participating in many university ICPC-style competitions do not align well with the timeline.
- Claims about high achievements in ECPC/EOI-related contests could not be verified.
- ACPC Teens participation reportedly happened only once and without notable ranking.
Another important point is the inconsistency between contest skill and general problem-solving history.
A person who can suddenly obtain top ranks in difficult Div2 contests would normally also show:
- strong old submissions,
- difficult solved problems outside contests,
- stable high-level performance afterward.
Instead, the account history looks extremely inconsistent.
There are also suspicious behavioral patterns:
- Very sudden spikes in contest performance.
- Contest-only strength not reflected elsewhere.
- Large differences between contests close in time.
Overall, while rapid improvement is possible, the combination of:
- 5 years at newbie level,
- sudden 1000+ rating jump,
- rank 2 in a major round,
- inconsistent history,
- questionable explanations,
- and multiple suspicious contest patterns
makes this account highly suspicious.
2. MIDORIYA_
This account first appeared mainly to defend batats in the comment section.
However, many of the arguments were extremely weak and sometimes even made the situation more suspicious.
Some examples:
- Using irrelevant comparisons with genuinely talented users.
- Claiming that huge improvement in only a few contests is “normal” without providing convincing evidence.
- Ignoring the inconsistency between old and new performances.
- Avoiding direct responses to stronger evidence.
After additional evidence was presented, the account became much quieter and stopped responding to several important points.
There are also suspicious similarities between the two accounts:
- Similar timing patterns.
- Similar argumentative behavior.
- Defending each other aggressively despite weak evidence.
- Suspicious overlap in contest periods.
Another suspicious point is the attempt to dismiss evidence using obviously bad arguments instead of actually explaining the abnormal contest history.
For example, instead of explaining:
- how a newbie-level account suddenly became expert/master level,
- why there was no gradual progression,
- or why performances were inconsistent,
the defense mostly relied on vague claims like:
- “people can improve fast,”
- “he practiced a lot,”
- or comparisons to unrelated users.
But fast improvement alone is not the issue.
The issue is the combination of:
- extremely long low-level history,
- abrupt transformation,
- lack of supporting practice evidence,
- and highly inconsistent performances.
Conclusion
Of course, only the official Codeforces anti-cheat team can make the final judgment.
However, based on:
- the contest history,
- rating graph,
- performance inconsistencies,
- suspicious timeline,
- weak explanations,
- and overall behavior,
I believe there is more than enough evidence for serious suspicion and investigation.
The purpose of this blog is not harassment. Competitive programming contests should remain fair for everyone who spends years practicing honestly.
I hope that MikeMirzayanov and KAN will hear my words and ban the cheaters mentioned above.
