\n
![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F17484946%2F4909823c2f1ffa23899dc0d88618018e%2F1.png?generation=1751361592636939&alt=media\")
Today I saw a discussion about cheating in the Yale/UNC-CH competition forum. I want to share my thoughts on this as well. Not only the Yale/UNC-CH - Geophysical Waveform Inversion competition, but also the Jane Street Real-Time Market Data Forecasting competition also has a large number of zombie accounts. In addition, there are a lot of suspicious teams in Jane Street Real-Time Market Data Forecasting.
\n
Here is a suspicious example of the Jane Street Real-Time Market Data Forecasting competition
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F17484946%2F716c5b04f39695bfd1a3fe324734aa73%2F1.png?generation=1750829178293377&alt=media\")
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F17484946%2F3aa25c4d76ee3da6dabd532b42b2eb1b%2F2.png?generation=1750829193976281&alt=media\")
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F17484946%2F316d41dba444839e1d236ca3020b687a%2F3.png?generation=1750829205058827&alt=media\")
By reviewing the activity records of certain accounts, we observed the following patterns:
\nSome accounts remained inactive for extended periods but suddenly became active and submitted predictions during the competition;
\nThese accounts display similar activity cycles and contribution graphs.
\nSuch patterns strongly suggest the use of multiple controlled accounts to participate in the competition simultaneously, likely to test various submissions and probe the leaderboard.
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F17484946%2Faaa2855585b85e272776ee738358dcfd%2F5.png?generation=1750829244349315&alt=media\")
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F17484946%2Faa2ea153a7a28a7a0ca55c54186a7cd3%2F6.png?generation=1750829267536275&alt=media\")
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F17484946%2Fb4a48c03786b32f0b8c931b5c61d8bbb%2F7.png?generation=1750829275964778&alt=media\")
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F17484946%2F2d8e5bc1711ae279642be9d9724e8bf5%2F8.png?generation=1750829285171540&alt=media\")
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F17484946%2Fc539781c9a62193f426b575dc6bf03b1%2F9.png?generation=1750829293442316&alt=media\")
I suggest verifying whether the source of their forked open-source notebook is the same, and whether applying the same modifications leads to improved scores.
\nAdditionally, I strongly recommend tracking any changes to the email addresses of winning accounts, as well as their activity patterns after receiving a medal.
\n@ryanholbrook @kaitoneil @mylesoneill @mdesaijs @maggiemd @wcukierski @addisonhoward
", "messages": [ { "id": 3237744, "postDate": "2025-07-01T09:20:43.797Z", "content": "Today I saw a discussion about cheating in the Yale/UNC-CH competition forum. I want to share my thoughts on this as well. Not only the Yale/UNC-CH - Geophysical Waveform Inversion competition, but also the Jane Street Real-Time Market Data Forecasting competition also has a large number of zombie accounts. In addition, there are a lot of suspicious teams in Jane Street Real-Time Market Data Forecasting.
\n
Here is a suspicious example of the Jane Street Real-Time Market Data Forecasting competition
\n
\n
\n
By reviewing the activity records of certain accounts, we observed the following patterns:
\nSome accounts remained inactive for extended periods but suddenly became active and submitted predictions during the competition;
\nThese accounts display similar activity cycles and contribution graphs.
\nSuch patterns strongly suggest the use of multiple controlled accounts to participate in the competition simultaneously, likely to test various submissions and probe the leaderboard.
\n
\n
\n
\n
\n
I suggest verifying whether the source of their forked open-source notebook is the same, and whether applying the same modifications leads to improved scores.
\nAdditionally, I strongly recommend tracking any changes to the email addresses of winning accounts, as well as their activity patterns after receiving a medal.
\n@ryanholbrook @kaitoneil @mylesoneill @mdesaijs @maggiemd @wcukierski @addisonhoward
", "rawMarkdown": "Today I saw a discussion about cheating in the Yale/UNC-CH competition forum. I want to share my thoughts on this as well. Not only the Yale/UNC-CH - Geophysical Waveform Inversion competition, but also the Jane Street Real-Time Market Data Forecasting competition also has a large number of zombie accounts. In addition, there are a lot of suspicious teams in Jane Street Real-Time Market Data Forecasting.\n\n\n\nHere is a suspicious example of the Jane Street Real-Time Market Data Forecasting competition\n\n\n\n\nBy reviewing the activity records of certain accounts, we observed the following patterns:\n\nSome accounts remained inactive for extended periods but suddenly became active and submitted predictions during the competition;\n\nThese accounts display similar activity cycles and contribution graphs.\n\nSuch patterns strongly suggest the use of multiple controlled accounts to participate in the competition simultaneously, likely to test various submissions and probe the leaderboard.\n\n\n\n\n\n\nI suggest verifying whether the source of their forked open-source notebook is the same, and whether applying the same modifications leads to improved scores.\n\nAdditionally, I strongly recommend tracking any changes to the email addresses of winning accounts, as well as their activity patterns after receiving a medal.\n\n@ryanholbrook @kaitoneil @mylesoneill @mdesaijs @maggiemd @wcukierski @addisonhoward ", "votes": 5 } ], "comments": [], "raw_markdown_by_id": { "3237744": "Today I saw a discussion about cheating in the Yale/UNC-CH competition forum. I want to share my thoughts on this as well. Not only the Yale/UNC-CH - Geophysical Waveform Inversion competition, but also the Jane Street Real-Time Market Data Forecasting competition also has a large number of zombie accounts. In addition, there are a lot of suspicious teams in Jane Street Real-Time Market Data Forecasting.\n\n\n\nHere is a suspicious example of the Jane Street Real-Time Market Data Forecasting competition\n\n\n\n\nBy reviewing the activity records of certain accounts, we observed the following patterns:\n\nSome accounts remained inactive for extended periods but suddenly became active and submitted predictions during the competition;\n\nThese accounts display similar activity cycles and contribution graphs.\n\nSuch patterns strongly suggest the use of multiple controlled accounts to participate in the competition simultaneously, likely to test various submissions and probe the leaderboard.\n\n\n\n\n\n\nI suggest verifying whether the source of their forked open-source notebook is the same, and whether applying the same modifications leads to improved scores.\n\nAdditionally, I strongly recommend tracking any changes to the email addresses of winning accounts, as well as their activity patterns after receiving a medal.\n\n@ryanholbrook @kaitoneil @mylesoneill @mdesaijs @maggiemd @wcukierski @addisonhoward " } }