![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F12596779%2F2dcd2fef0dc83e2f81d87e08713fb22a%2FScreenshot%20from%202024-10-09%2003-06-24.png?generation=1728432400708855&alt=media\")
It is extremely suspicious that so many recently created accounts suddenly have such similar public score and then TOGETHER go up on the private, I sure do believe in coincidences, but I am convinced is not the one…
\nI hope kaggle admins will take appropriate measures.
It is extremely suspicious that so many recently created accounts suddenly have such similar public score and then TOGETHER go up on the private, I sure do believe in coincidences, but I am convinced is not the one…
\nI hope kaggle admins will take appropriate measures.
0.48 is best public notebook
", "rawMarkdown": "0.48 is best public notebook", "votes": null }, { "id": "3012343", "postDate": "10/09/2024 00:15:31", "content": "the screenshot I shared is actually of the private test, on the public they all had around 0.43
", "rawMarkdown": "the screenshot I shared is actually of the private test, on the public they all had around 0.43", "votes": null }, { "id": "3012346", "postDate": "10/09/2024 00:19:49", "content": "oh, it is my mistake. sorry for that.
", "rawMarkdown": "oh, it is my mistake. sorry for that.", "votes": null }, { "id": "3013020", "postDate": "10/09/2024 15:21:33", "content": "I wouldn't even hope, Ivan. Kaggle team will do very little about it as it has been common.
\nAren't I right @sohier , @addisonhoward ?
From all final leaderboards of competitions without a significant shake up it is apparent that the medal sell-buy business has been flourishing.
\nThere is a lot of evidence from previous competitions that despite having detailed public information about how the business works from decent Kagglers (including Grandmasters) who speek the language, Kaggle team doesn't seem to have reacted anyhow and the leaderboards stayed nearly as they had been.
BTW though the cluster you mentioned is the best visible unfortunately it is not the only one. Here, pretty traditionally, the groups of the same-pattern-teams cover a part of the silver and top bronze zone to the place 107. Their complete number is about 37.
", "rawMarkdown": "I wouldn't even hope, Ivan. Kaggle team will do very little about it as it has been common.\nAren't I right @sohier , @addisonhoward ?\n\nFrom all final leaderboards of competitions without a significant shake up it is apparent that the medal sell-buy business has been flourishing. \nThere is a lot of evidence from previous competitions that despite having detailed public information about how the business works from decent Kagglers (including Grandmasters) who speek the language, Kaggle team doesn't seem to have reacted anyhow and the leaderboards stayed nearly as they had been.\n\nBTW though the cluster you mentioned is the best visible unfortunately it is not the only one. Here, pretty traditionally, the groups of the same-pattern-teams cover a part of the silver and top bronze zone to the place 107. Their complete number is about 37.", "votes": null }, { "id": "3013142", "postDate": "10/09/2024 17:19:52", "content": "Another solution sell case unfortunately
", "rawMarkdown": "Another solution sell case unfortunately", "votes": null }, { "id": "3013327", "postDate": "10/10/2024 00:23:08", "content": "Unfortunately, the scores have been finalized, as I understood, there aren't any bans/investigation incoming(
", "rawMarkdown": "Unfortunately, the scores have been finalized, as I understood, there aren't any bans/investigation incoming(", "votes": null }, { "id": "3013514", "postDate": "10/10/2024 08:19:35", "content": "Kaggle anti-cheating is a joke
", "rawMarkdown": "Kaggle anti-cheating is a joke", "votes": null }, { "id": "3013558", "postDate": "10/10/2024 09:17:17", "content": "Joined two months ago, joined three months ago, joined a month ago… There should be additional verifications.
", "rawMarkdown": "Joined two months ago, joined three months ago, joined a month ago... There should be additional verifications.", "votes": null }, { "id": "3013578", "postDate": "10/10/2024 09:37:48", "content": "yes, quite suspicious.
\nhowever, can't just delete them without evidence.
\nkaggle team always take actions when they find cheating.
\nif competitors were taught or they know how to bypass sensorship, it's hard to find by kaggle.
Sorry @huyidao but it is not true.
\nKaggle team has been given plenty of detailed evidence both publicly and via compliance channel as you can find in discussions to previous competitions. They seem to simply decide to ignore everything.
And this was a code competition and although the scores of all suspicious teams are slightly different it must be very easy to compare the code and prove the same root source of all the submissions.
\nThis should be a natural work after every competition if there is not a public notebook with such a score.
Kaggle's anti-cheating system is already very strong compared to other competition platforms (personal opinion)
", "rawMarkdown": "Kaggle's anti-cheating system is already very strong compared to other competition platforms (personal opinion)", "votes": null }, { "id": "3014219", "postDate": "10/11/2024 03:47:57", "content": "@ivanvybornov In fact, the easiest way is to look at this group of recently created user-submitted solution texts.
", "rawMarkdown": "ivanvybornov In fact, the easiest way is to look at this group of recently created user-submitted solution texts.", "votes": null } ], "comments": [ { "id": 3012341, "author_name": "hengck23", "author_url": "", "post_date": "10/09/2024 00:13:11", "content": "0.48 is best public notebook
", "votes": null, "replies": [ { "id": 3012343, "author_name": "ivanvybornov", "author_url": "", "post_date": "10/09/2024 00:15:31", "content": "the screenshot I shared is actually of the private test, on the public they all had around 0.43
", "votes": null, "replies": [ { "id": 3012346, "author_name": "hengck23", "author_url": "", "post_date": "10/09/2024 00:19:49", "content": "oh, it is my mistake. sorry for that.
", "votes": null, "replies": [] } ] } ] }, { "id": 3013020, "author_name": "blankaf", "author_url": "", "post_date": "10/09/2024 15:21:33", "content": "I wouldn't even hope, Ivan. Kaggle team will do very little about it as it has been common.
\nAren't I right @sohier , @addisonhoward ?
From all final leaderboards of competitions without a significant shake up it is apparent that the medal sell-buy business has been flourishing.
\nThere is a lot of evidence from previous competitions that despite having detailed public information about how the business works from decent Kagglers (including Grandmasters) who speek the language, Kaggle team doesn't seem to have reacted anyhow and the leaderboards stayed nearly as they had been.
BTW though the cluster you mentioned is the best visible unfortunately it is not the only one. Here, pretty traditionally, the groups of the same-pattern-teams cover a part of the silver and top bronze zone to the place 107. Their complete number is about 37.
", "votes": null, "replies": [ { "id": 3013327, "author_name": "ivanvybornov", "author_url": "", "post_date": "10/10/2024 00:23:08", "content": "Unfortunately, the scores have been finalized, as I understood, there aren't any bans/investigation incoming(
", "votes": null, "replies": [ { "id": 3013514, "author_name": "aerdem4", "author_url": "", "post_date": "10/10/2024 08:19:35", "content": "Kaggle anti-cheating is a joke
", "votes": null, "replies": [ { "id": 3014216, "author_name": "oralia", "author_url": "", "post_date": "10/11/2024 03:46:19", "content": "Kaggle's anti-cheating system is already very strong compared to other competition platforms (personal opinion)
", "votes": null, "replies": [] } ] } ] } ] }, { "id": 3013142, "author_name": "aerdem4", "author_url": "", "post_date": "10/09/2024 17:19:52", "content": "Another solution sell case unfortunately
", "votes": null, "replies": [] }, { "id": 3013558, "author_name": "snnclsr", "author_url": "", "post_date": "10/10/2024 09:17:17", "content": "Joined two months ago, joined three months ago, joined a month ago… There should be additional verifications.
", "votes": null, "replies": [] }, { "id": 3013578, "author_name": "huyidao", "author_url": "", "post_date": "10/10/2024 09:37:48", "content": "yes, quite suspicious.
\nhowever, can't just delete them without evidence.
\nkaggle team always take actions when they find cheating.
\nif competitors were taught or they know how to bypass sensorship, it's hard to find by kaggle.
Sorry @huyidao but it is not true.
\nKaggle team has been given plenty of detailed evidence both publicly and via compliance channel as you can find in discussions to previous competitions. They seem to simply decide to ignore everything.
And this was a code competition and although the scores of all suspicious teams are slightly different it must be very easy to compare the code and prove the same root source of all the submissions.
\nThis should be a natural work after every competition if there is not a public notebook with such a score.
@ivanvybornov In fact, the easiest way is to look at this group of recently created user-submitted solution texts.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "3012339": "\n\nIt is extremely suspicious that so many recently created accounts suddenly have such similar public score and then TOGETHER go up on the private, I sure do believe in coincidences, but I am convinced is not the one...\nI hope kaggle admins will take appropriate measures.", "3012341": "0.48 is best public notebook", "3012343": "the screenshot I shared is actually of the private test, on the public they all had around 0.43", "3012346": "oh, it is my mistake. sorry for that.", "3013020": "I wouldn't even hope, Ivan. Kaggle team will do very little about it as it has been common.\nAren't I right @sohier , @addisonhoward ?\n\nFrom all final leaderboards of competitions without a significant shake up it is apparent that the medal sell-buy business has been flourishing. \nThere is a lot of evidence from previous competitions that despite having detailed public information about how the business works from decent Kagglers (including Grandmasters) who speek the language, Kaggle team doesn't seem to have reacted anyhow and the leaderboards stayed nearly as they had been.\n\nBTW though the cluster you mentioned is the best visible unfortunately it is not the only one. Here, pretty traditionally, the groups of the same-pattern-teams cover a part of the silver and top bronze zone to the place 107. Their complete number is about 37.", "3013142": "Another solution sell case unfortunately", "3013327": "Unfortunately, the scores have been finalized, as I understood, there aren't any bans/investigation incoming(", "3013514": "Kaggle anti-cheating is a joke", "3013558": "Joined two months ago, joined three months ago, joined a month ago... There should be additional verifications.", "3013578": "yes, quite suspicious.\nhowever, can't just delete them without evidence. \nkaggle team always take actions when they find cheating. \nif competitors were taught or they know how to bypass sensorship, it's hard to find by kaggle.", "3013856": "Sorry @huyidao but it is not true.\nKaggle team has been given plenty of detailed evidence both publicly and via compliance channel as you can find in discussions to previous competitions. They seem to simply decide to ignore everything.\n\nAnd this was a code competition and although the scores of all suspicious teams are slightly different it must be very easy to compare the code and prove the same root source of all the submissions. \nThis should be a natural work after every competition if there is not a public notebook with such a score.", "3014216": "Kaggle's anti-cheating system is already very strong compared to other competition platforms (personal opinion)", "3014219": "ivanvybornov In fact, the easiest way is to look at this group of recently created user-submitted solution texts." }, "source": "meta" }