I had taken an image of the old leaderboard.
\n As you said, it looks like the 4th and 14th places on the original leaderboard are gone. Once again, the 13th and 14th places are exactly the same as the shake up.
![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F2930293%2Fc61e87fd16812275da0ec861e385e956%2Fleaderboard_old.png?generation=1661567322162440&alt=media\")
I just started my kaggle career and this is the second competition that I joined. I was at 12nd place but today when I again logged in, I was surprised to see that I was removed from the leaderboard. I never open any discussion before and just focus on improving my own model performance. After I submitted statement in compliance sector, I received the reply as follows without any other evidence:
\nWe have completed our review and have determined that a member of your team violated competition rules through private sharing or using multiple accounts. We will not be reinstating your team.
\nwhich is a meaningless and perfunctory reply. I don't even know what it means since I only have one account and never share anything with anyone. As a individual competitor in this competition, I don't even get in touch with anyone else. However, I received the email which told me that I received an ' unspecified' medal in competition. That's why I double check my account and knew I was removed without any reasons.
\nI spent quite a lot time on this competition but even the staff is not responsible for the issue. Is it how you treat you participants? If kaggle is not even a fair and just platform, who will stay?
", "messages": [ { "id": 1920098, "postDate": "2022-08-30T22:08:02.793Z", "content": "I am fairly certain that Kaggle staff don't remove participants randomly. While I sympathize with your disappointment, raving about it and calling out hard-working admins will never work in your favor, either with them or with general community.
\nI have no problem that they don't disclose their methodology, because the moment they do it someone will share that reason, the same way you shared their responses. That will help future offenders avoid detection, so of course they don't want to do it.
\nIt is not difficult, however, to guess how they detect offenders. If multiple accounts are accessed from the same IP address, or if multiple accounts have similar submission names, similar submission times, it is likely that they belong to the same person. Please don't take this the wrong way, but I know there is no way that you didn't do anything wrong and they simply canned you. If that were the case, there would be many random competitors eliminated. Maybe you created an account 3 months ago, submitted a few times, then forgot the password. If you created a new account and accessed it from the same computer, that is a violation, even if you didn't do it for the purposes of cheating.
", "rawMarkdown": "I am fairly certain that Kaggle staff don't remove participants randomly. While I sympathize with your disappointment, raving about it and calling out hard-working admins will never work in your favor, either with them or with general community.\n\nI have no problem that they don't disclose their methodology, because the moment they do it someone will share that reason, the same way you shared their responses. That will help future offenders avoid detection, so of course they don't want to do it.\n\nIt is not difficult, however, to guess how they detect offenders. If multiple accounts are accessed from the same IP address, or if multiple accounts have similar submission names, similar submission times, it is likely that they belong to the same person. Please don't take this the wrong way, but I know there is no way that you didn't do anything wrong and they simply canned you. If that were the case, there would be many random competitors eliminated. Maybe you created an account 3 months ago, submitted a few times, then forgot the password. If you created a new account and accessed it from the same computer, that is a violation, even if you didn't do it for the purposes of cheating.", "votes": 6 }, { "id": 1919068, "postDate": "2022-08-30T05:27:38.580Z", "content": "I just started my kaggle career and this is the second competition that I joined. I was at 12nd place but today when I again logged in, I was surprised to see that I was removed from the leaderboard. I never open any discussion before and just focus on improving my own model performance. After I submitted statement in compliance sector, I received the reply as follows without any other evidence:
\nWe have completed our review and have determined that a member of your team violated competition rules through private sharing or using multiple accounts. We will not be reinstating your team.
\nwhich is a meaningless and perfunctory reply. I don't even know what it means since I only have one account and never share anything with anyone. As a individual competitor in this competition, I don't even get in touch with anyone else. However, I received the email which told me that I received an ' unspecified' medal in competition. That's why I double check my account and knew I was removed without any reasons.
\nI spent quite a lot time on this competition but even the staff is not responsible for the issue. Is it how you treat you participants? If kaggle is not even a fair and just platform, who will stay?
", "rawMarkdown": "I just started my kaggle career and this is the second competition that I joined. I was at 12nd place but today when I again logged in, I was surprised to see that I was removed from the leaderboard. I never open any discussion before and just focus on improving my own model performance. After I submitted statement in compliance sector, I received the reply as follows without any other evidence:\n\nWe have completed our review and have determined that a member of your team violated competition rules through private sharing or using multiple accounts. We will not be reinstating your team. \n\nwhich is a meaningless and perfunctory reply. I don't even know what it means since I only have one account and never share anything with anyone. As a individual competitor in this competition, I don't even get in touch with anyone else. However, I received the email which told me that I received an ' unspecified' medal in competition. That's why I double check my account and knew I was removed without any reasons.\n\nI spent quite a lot time on this competition but even the staff is not responsible for the issue. Is it how you treat you participants? If kaggle is not even a fair and just platform, who will stay? ", "votes": 6 }, { "id": 1919278, "postDate": "2022-08-30T09:35:01.503Z", "content": "This is a post from another discussion:
\n\n\nI had taken an image of the old leaderboard.
\n
\n As you said, it looks like the 4th and 14th places on the original leaderboard are gone. Once again, the 13th and 14th places are exactly the same as the shake up.\n
It shows that you and place 14 had an identical score. Oddly enough, also a somewhat similar team name. Apparently, the scores were also identical on the public leaderboard. How do you think this happened?
", "rawMarkdown": "This is a post from [another discussion](https://www.kaggle.com/competitions/amex-default-prediction/discussion/348093#1915476):\n\n> I had taken an image of the old leaderboard.\n> As you said, it looks like the 4th and 14th places on the original leaderboard are gone. Once again, the 13th and 14th places are exactly the same as the shake up.\n> \n> \n\nIt shows that you and place 14 had an identical score. Oddly enough, also a somewhat similar team name. Apparently, the scores were also identical on the public leaderboard. How do you think this happened?", "votes": 3, "replies": [ { "id": 1919320, "postDate": "2022-08-30T10:12:59.247Z", "content": "Thanks for sharing this image with me. Same score is common in this competition, as you can see now, many team have same score with other teams i.e. team rank 29th and team rank 30th, team rank 12th and team rank 13th ,team rank 22th and team rank 23th etc. I just listed some teams that have higher rank. More examples are in those who stay lower rank. It’s not odd in this competition since there are many examples.
\nFor name, how can you identify ‘ossas’ and ‘hh_oas’ as the similar name? Just because we all have letter ‘o’,’a’ and ‘s’? If so, many teams have similar names. Obviously, that’s ridiculous. Besides, I don’t know who is in team ‘hh_oas’ and I don’t have any connection with them even the email or team merge request. I don’t care the members in any other team.
\nFor the identical score on public leaderboard,currently there are still many teams have same socre in public leaderboard. Team rank 29th and team rank 30th also have the identical scores both in public leaderboard and private leaderboard within teams ranking top 30. Maybe there are more in silver medals and bronze medals zone etc but I don't verify this. I don’t regard it as a strange phenomena.
\nI don’t know how others build their model and get the output. For my final submission, I add a little random factor in my last 5 csv files (i.e output+rand()* a small multiplier). Theoretically speaking, the random factor might cause the similarity although the probability is extremely low. If you have to ask, this might be the only factor that I suspect.
", "rawMarkdown": "Thanks for sharing this image with me. Same score is common in this competition, as you can see now, many team have same score with other teams i.e. team rank 29th and team rank 30th, team rank 12th and team rank 13th ,team rank 22th and team rank 23th etc. I just listed some teams that have higher rank. More examples are in those who stay lower rank. It’s not odd in this competition since there are many examples. \n\nFor name, how can you identify ‘ossas’ and ‘hh_oas’ as the similar name? Just because we all have letter ‘o’,’a’ and ‘s’? If so, many teams have similar names. Obviously, that’s ridiculous. Besides, I don’t know who is in team ‘hh_oas’ and I don’t have any connection with them even the email or team merge request. I don’t care the members in any other team. \n\nFor the identical score on public leaderboard,currently there are still many teams have same socre in public leaderboard. Team rank 29th and team rank 30th also have the identical scores both in public leaderboard and private leaderboard within teams ranking top 30. Maybe there are more in silver medals and bronze medals zone etc but I don't verify this. I don’t regard it as a strange phenomena.\n\nI don’t know how others build their model and get the output. For my final submission, I add a little random factor in my last 5 csv files (i.e output+rand()* a small multiplier). Theoretically speaking, the random factor might cause the similarity although the probability is extremely low. If you have to ask, this might be the only factor that I suspect. " }, { "id": 1919353, "postDate": "2022-08-30T10:35:00.987Z", "rawMarkdown": "", "isDeleted": true }, { "id": 1919376, "postDate": "2022-08-30T10:58:37.770Z", "content": "I understand that, I can't know what was the reason but that was my assumption. Also, some teams have the same score on public and private lb, thats right. But you have the same score on public, and private lb, which is much more likely. Also, kaggle staff does not need to rely on the few lb digits we are given, but can precisely look for patterns between submission. Perhaps they found something there?
", "rawMarkdown": "I understand that, I can't know what was the reason but that was my assumption. Also, some teams have the same score on public and private lb, thats right. But you have the same score on public, and private lb, which is much more likely. Also, kaggle staff does not need to rely on the few lb digits we are given, but can precisely look for patterns between submission. Perhaps they found something there?", "votes": 1 }, { "id": 1919463, "postDate": "2022-08-30T12:37:49.230Z", "content": "After seeing your sharing did I realize that I have the same score with the team right behind me. Besides, there are also some other teams have the same score in the private leaderboard now as you can see. In addition, there are also other teams have both identical public score and private score(i.e: team rank 29th and team rank 30th as I mention before). I agree that they have more methods and information to identify, but it doesn't mean that they do not make mistake. I follow all the rules about the submission(like 5 submissions a day etc). I don't have many submissions in total since I prefer to submit my files once I have a significant improvement of my model performance. Therefore, I don't submit the files everyday. I assume that no submission in some days is also allowed. Maybe they compare my submission time with other teams(just my guess), but how can I know when will other teams submit their files and obviously it's impossible for me to stop others from submitting. By the way, after compliance, no kaggle staff or team member is responsible for further feedback or provide their evidence or tell me which rule that I violate and reasons. Then what' the role of kaggle compliance if I can only get the somewhat 'auto-reply'? A air-talk to 'proof' that they are somehow 'just and fair'? Really disappointed about kaggle.
", "rawMarkdown": "After seeing your sharing did I realize that I have the same score with the team right behind me. Besides, there are also some other teams have the same score in the private leaderboard now as you can see. In addition, there are also other teams have both identical public score and private score(i.e: team rank 29th and team rank 30th as I mention before). I agree that they have more methods and information to identify, but it doesn't mean that they do not make mistake. I follow all the rules about the submission(like 5 submissions a day etc). I don't have many submissions in total since I prefer to submit my files once I have a significant improvement of my model performance. Therefore, I don't submit the files everyday. I assume that no submission in some days is also allowed. Maybe they compare my submission time with other teams(just my guess), but how can I know when will other teams submit their files and obviously it's impossible for me to stop others from submitting. By the way, after compliance, no kaggle staff or team member is responsible for further feedback or provide their evidence or tell me which rule that I violate and reasons. Then what' the role of kaggle compliance if I can only get the somewhat 'auto-reply'? A air-talk to 'proof' that they are somehow 'just and fair'? Really disappointed about kaggle." } ] }, { "id": 1920861, "postDate": "2022-08-31T12:58:27.123Z", "rawMarkdown": "", "votes": -9, "replies": [ { "id": 1920949, "postDate": "2022-08-31T13:43:13.863Z", "content": "It may not be a scam, but it is a spam.
", "rawMarkdown": "It may not be a scam, but it is a spam.", "votes": 4 } ] }, { "id": 1919685, "postDate": "2022-08-30T15:51:53.943Z", "content": "Some updates here. I email to kaggle staff Addison Howard and here is the response that I get.
\nHi there,
\nWe do not divulge details of how we determine if you shared privately or held multiple accounts - you may not share outside of Kaggle teams (including sharing before you merged), and you may not use multiple accounts to submit to a Kaggle competition.
\nRegards,
\nAddison Howard
\nStill talking to air and not even a valid response. Also feel like a robot reply. If you can not disclose anything, then how can any participant regard kaggle competition as a fair and justified platform which has clear judgement criteria instead of a whole blackbox and bearing the risk of being removed from the leaderboard without any reasons after spending quite a lot of time and effort and the whole team is not responsible for it.
", "rawMarkdown": "Some updates here. I email to kaggle staff Addison Howard and here is the response that I get. \n\nHi there,\n\nWe do not divulge details of how we determine if you shared privately or held multiple accounts - you may not share outside of Kaggle teams (including sharing before you merged), and you may not use multiple accounts to submit to a Kaggle competition.\n\nRegards,\n\nAddison Howard\n\nStill talking to air and not even a valid response. Also feel like a robot reply. If you can not disclose anything, then how can any participant regard kaggle competition as a fair and justified platform which has clear judgement criteria instead of a whole blackbox and bearing the risk of being removed from the leaderboard without any reasons after spending quite a lot of time and effort and the whole team is not responsible for it. ", "votes": -10, "replies": [ { "id": 1919875, "postDate": "2022-08-30T18:04:24.050Z", "content": "For the record, are you maintaining that these two accounts have nothing to do with you?
\nhttps://www.kaggle.com/ssssswchen
\nhttps://www.kaggle.com/songweichenss
Yes and seems that first one is ''hh_oas'.
", "rawMarkdown": "Yes and seems that first one is ''hh_oas'. ", "votes": -4 }, { "id": 1920248, "postDate": "2022-08-31T03:13:59.243Z", "content": "We have overwhelming evidence, and I mean overwhelming evidence, to suggest otherwise.
", "rawMarkdown": "We have overwhelming evidence, and I mean _overwhelming_ evidence, to suggest otherwise.", "votes": 16 }, { "id": 1920803, "postDate": "2022-08-31T12:21:38.647Z", "content": "… and he's gone :)
", "rawMarkdown": "... and he's gone :)", "votes": 5 }, { "id": 1921128, "postDate": "2022-08-31T15:48:44.800Z", "content": "The other account, supposedly unrelated, is gone as well. World is full of coincidences ¯_(ツ)_/¯
", "rawMarkdown": "The other account, supposedly unrelated, is gone as well. World is full of coincidences ¯\\_(ツ)_/¯", "votes": 2 }, { "id": 1921299, "postDate": "2022-08-31T17:39:49.393Z", "content": "I wouldn't assume that was the choice of the owner(s) of those accounts. It could be Kaggle Staff choice based on his response above and their evidence plus criteria for banning. This situation, with \"for the record\" question from Kaggle Staff seems similar to the below, from their compliance page:
\n\"You should NOT use this form if you or your teammates violated the competition rules. Doing so may result in further punitive action against your account(s).\"
", "rawMarkdown": "I wouldn't assume that was the choice of the owner(s) of those accounts. It could be Kaggle Staff choice based on his response above and their evidence plus criteria for banning. This situation, with \"for the record\" question from Kaggle Staff seems similar to the below, from their compliance page:\n\n\"You should NOT use this form if you or your teammates violated the competition rules. Doing so may result in further punitive action against your account(s).\"", "votes": 1 }, { "id": 1921369, "postDate": "2022-08-31T19:10:41.823Z", "content": "I don't think Kaggle admins will remove an account just because the owner complained, either publicly or privately, though I am sure they remove accounts for more egregious violations. I have read many complaints in previous competitions where the offenders retained their accounts.
\nGiven that world is full of coincidences, maybe that's the case here with two accounts deleted within hours of an admin mentioning overwhelming evidence of wrongdoing. By the same token, it may not be a coincidence.
", "rawMarkdown": "I don't think Kaggle admins will remove an account just because the owner complained, either publicly or privately, though I am sure they remove accounts for more egregious violations. I have read many complaints in previous competitions where the offenders retained their accounts.\n\nGiven that world is full of coincidences, maybe that's the case here with two accounts deleted within hours of an admin mentioning overwhelming evidence of wrongdoing. By the same token, it may not be a coincidence.", "votes": 2 }, { "id": 1921503, "postDate": "2022-08-31T22:05:17.890Z", "content": "Oh, I'm not arguing against the possiblity the owner deleted the account. I'm still arguing that perhaps the admins would ban an account for lying to them during an official investigation. Not for the complaining! For the on-record \"yes\" response.
", "rawMarkdown": "Oh, I'm not arguing against the possiblity the owner deleted the account. I'm still arguing that perhaps the admins would ban an account for lying to them during an official investigation. Not for the complaining! For the on-record \"yes\" response. ", "votes": 1 }, { "id": 1921520, "postDate": "2022-08-31T22:26:42.007Z", "content": "Those who cheat are already lying on the record by breaking the agreement we all have made to obey the competition rules, but it is possible that logging a complaint when one doesn't have any case, and thus creating extra work for admins, is grounds for banning.
", "rawMarkdown": "Those who cheat are already lying on the record by breaking the agreement we all have made to obey the competition rules, but it is possible that logging a complaint when one doesn't have any case, and thus creating extra work for admins, is grounds for banning." }, { "id": 1922723, "postDate": "2022-09-01T17:07:35.903Z", "content": "Complaining will never get someone banned. A few things that will get you banned: Abusing resources, spam, threats, egregious cheating, repeated (or egregious) violations of Kaggle's code of conduct, etc.
\nIf a user submits a Compliance appeal when there is significant evidence of cheating, we will ban the user's account.
\nThere have been instances when a user has made an appeal, and, upon further investigation, we've reversed the decision to remove that team from the leaderboard.
", "rawMarkdown": "Complaining will never get someone banned. A few things that will get you banned: Abusing resources, spam, threats, egregious cheating, repeated (or egregious) violations of Kaggle's code of conduct, etc.\n\nIf a user submits a Compliance appeal when there is significant evidence of cheating, we will ban the user's account.\n\nThere have been instances when a user has made an appeal, and, upon further investigation, we've reversed the decision to remove that team from the leaderboard. ", "votes": 5 }, { "id": 1922989, "postDate": "2022-09-01T22:24:46.287Z", "content": "Don't know why but it was so satisfying to see the cheater lie then get banned right after 😄
", "rawMarkdown": "Don't know why but it was so satisfying to see the cheater lie then get banned right after 😄", "votes": 1 } ] } ], "comments": [ { "id": 1920098, "author_name": "Tilii", "author_url": "", "post_date": "2022-08-30T22:08:02.793000", "content": "I am fairly certain that Kaggle staff don't remove participants randomly. While I sympathize with your disappointment, raving about it and calling out hard-working admins will never work in your favor, either with them or with general community.
\nI have no problem that they don't disclose their methodology, because the moment they do it someone will share that reason, the same way you shared their responses. That will help future offenders avoid detection, so of course they don't want to do it.
\nIt is not difficult, however, to guess how they detect offenders. If multiple accounts are accessed from the same IP address, or if multiple accounts have similar submission names, similar submission times, it is likely that they belong to the same person. Please don't take this the wrong way, but I know there is no way that you didn't do anything wrong and they simply canned you. If that were the case, there would be many random competitors eliminated. Maybe you created an account 3 months ago, submitted a few times, then forgot the password. If you created a new account and accessed it from the same computer, that is a violation, even if you didn't do it for the purposes of cheating.
", "votes": 6, "replies": [] }, { "id": 1919278, "author_name": "Fritz Cremer", "author_url": "", "post_date": "2022-08-30T09:35:01.503000", "content": "This is a post from another discussion:
\n\n\nI had taken an image of the old leaderboard.
\n
\n As you said, it looks like the 4th and 14th places on the original leaderboard are gone. Once again, the 13th and 14th places are exactly the same as the shake up.\n
It shows that you and place 14 had an identical score. Oddly enough, also a somewhat similar team name. Apparently, the scores were also identical on the public leaderboard. How do you think this happened?
", "votes": 3, "replies": [ { "id": 1919320, "author_name": "", "author_url": "", "post_date": "2022-08-30T10:12:59.247000", "content": "Thanks for sharing this image with me. Same score is common in this competition, as you can see now, many team have same score with other teams i.e. team rank 29th and team rank 30th, team rank 12th and team rank 13th ,team rank 22th and team rank 23th etc. I just listed some teams that have higher rank. More examples are in those who stay lower rank. It’s not odd in this competition since there are many examples.
\nFor name, how can you identify ‘ossas’ and ‘hh_oas’ as the similar name? Just because we all have letter ‘o’,’a’ and ‘s’? If so, many teams have similar names. Obviously, that’s ridiculous. Besides, I don’t know who is in team ‘hh_oas’ and I don’t have any connection with them even the email or team merge request. I don’t care the members in any other team.
\nFor the identical score on public leaderboard,currently there are still many teams have same socre in public leaderboard. Team rank 29th and team rank 30th also have the identical scores both in public leaderboard and private leaderboard within teams ranking top 30. Maybe there are more in silver medals and bronze medals zone etc but I don't verify this. I don’t regard it as a strange phenomena.
\nI don’t know how others build their model and get the output. For my final submission, I add a little random factor in my last 5 csv files (i.e output+rand()* a small multiplier). Theoretically speaking, the random factor might cause the similarity although the probability is extremely low. If you have to ask, this might be the only factor that I suspect.
", "votes": 0, "replies": [] }, { "id": 1919353, "author_name": "", "author_url": "", "post_date": "2022-08-30T10:35:00.987000", "content": "", "votes": 0, "replies": [] }, { "id": 1919376, "author_name": "Fritz Cremer", "author_url": "", "post_date": "2022-08-30T10:58:37.770000", "content": "I understand that, I can't know what was the reason but that was my assumption. Also, some teams have the same score on public and private lb, thats right. But you have the same score on public, and private lb, which is much more likely. Also, kaggle staff does not need to rely on the few lb digits we are given, but can precisely look for patterns between submission. Perhaps they found something there?
", "votes": 1, "replies": [] }, { "id": 1919463, "author_name": "", "author_url": "", "post_date": "2022-08-30T12:37:49.230000", "content": "After seeing your sharing did I realize that I have the same score with the team right behind me. Besides, there are also some other teams have the same score in the private leaderboard now as you can see. In addition, there are also other teams have both identical public score and private score(i.e: team rank 29th and team rank 30th as I mention before). I agree that they have more methods and information to identify, but it doesn't mean that they do not make mistake. I follow all the rules about the submission(like 5 submissions a day etc). I don't have many submissions in total since I prefer to submit my files once I have a significant improvement of my model performance. Therefore, I don't submit the files everyday. I assume that no submission in some days is also allowed. Maybe they compare my submission time with other teams(just my guess), but how can I know when will other teams submit their files and obviously it's impossible for me to stop others from submitting. By the way, after compliance, no kaggle staff or team member is responsible for further feedback or provide their evidence or tell me which rule that I violate and reasons. Then what' the role of kaggle compliance if I can only get the somewhat 'auto-reply'? A air-talk to 'proof' that they are somehow 'just and fair'? Really disappointed about kaggle.
", "votes": 0, "replies": [] } ] }, { "id": 1920861, "author_name": "", "author_url": "", "post_date": "2022-08-31T12:58:27.123000", "content": "", "votes": -9, "replies": [ { "id": 1920949, "author_name": "Tilii", "author_url": "", "post_date": "2022-08-31T13:43:13.863000", "content": "It may not be a scam, but it is a spam.
", "votes": 4, "replies": [] } ] }, { "id": 1919685, "author_name": "", "author_url": "", "post_date": "2022-08-30T15:51:53.943000", "content": "Some updates here. I email to kaggle staff Addison Howard and here is the response that I get.
\nHi there,
\nWe do not divulge details of how we determine if you shared privately or held multiple accounts - you may not share outside of Kaggle teams (including sharing before you merged), and you may not use multiple accounts to submit to a Kaggle competition.
\nRegards,
\nAddison Howard
\nStill talking to air and not even a valid response. Also feel like a robot reply. If you can not disclose anything, then how can any participant regard kaggle competition as a fair and justified platform which has clear judgement criteria instead of a whole blackbox and bearing the risk of being removed from the leaderboard without any reasons after spending quite a lot of time and effort and the whole team is not responsible for it.
", "votes": -10, "replies": [ { "id": 1919875, "author_name": "inversion", "author_url": "", "post_date": "2022-08-30T18:04:24.050000", "content": "For the record, are you maintaining that these two accounts have nothing to do with you?
\nhttps://www.kaggle.com/ssssswchen
\nhttps://www.kaggle.com/songweichenss
Yes and seems that first one is ''hh_oas'.
", "votes": -4, "replies": [] }, { "id": 1920248, "author_name": "inversion", "author_url": "", "post_date": "2022-08-31T03:13:59.243000", "content": "We have overwhelming evidence, and I mean overwhelming evidence, to suggest otherwise.
", "votes": 16, "replies": [] }, { "id": 1920803, "author_name": "raddar", "author_url": "", "post_date": "2022-08-31T12:21:38.647000", "content": "… and he's gone :)
", "votes": 5, "replies": [] }, { "id": 1921128, "author_name": "Tilii", "author_url": "", "post_date": "2022-08-31T15:48:44.800000", "content": "The other account, supposedly unrelated, is gone as well. World is full of coincidences ¯_(ツ)_/¯
", "votes": 2, "replies": [] }, { "id": 1921299, "author_name": "Robert Hatch", "author_url": "", "post_date": "2022-08-31T17:39:49.393000", "content": "I wouldn't assume that was the choice of the owner(s) of those accounts. It could be Kaggle Staff choice based on his response above and their evidence plus criteria for banning. This situation, with \"for the record\" question from Kaggle Staff seems similar to the below, from their compliance page:
\n\"You should NOT use this form if you or your teammates violated the competition rules. Doing so may result in further punitive action against your account(s).\"
", "votes": 1, "replies": [] }, { "id": 1921369, "author_name": "Tilii", "author_url": "", "post_date": "2022-08-31T19:10:41.823000", "content": "I don't think Kaggle admins will remove an account just because the owner complained, either publicly or privately, though I am sure they remove accounts for more egregious violations. I have read many complaints in previous competitions where the offenders retained their accounts.
\nGiven that world is full of coincidences, maybe that's the case here with two accounts deleted within hours of an admin mentioning overwhelming evidence of wrongdoing. By the same token, it may not be a coincidence.
", "votes": 2, "replies": [] }, { "id": 1921503, "author_name": "Robert Hatch", "author_url": "", "post_date": "2022-08-31T22:05:17.890000", "content": "Oh, I'm not arguing against the possiblity the owner deleted the account. I'm still arguing that perhaps the admins would ban an account for lying to them during an official investigation. Not for the complaining! For the on-record \"yes\" response.
", "votes": 1, "replies": [] }, { "id": 1921520, "author_name": "Tilii", "author_url": "", "post_date": "2022-08-31T22:26:42.007000", "content": "Those who cheat are already lying on the record by breaking the agreement we all have made to obey the competition rules, but it is possible that logging a complaint when one doesn't have any case, and thus creating extra work for admins, is grounds for banning.
", "votes": 0, "replies": [] }, { "id": 1922723, "author_name": "inversion", "author_url": "", "post_date": "2022-09-01T17:07:35.903000", "content": "Complaining will never get someone banned. A few things that will get you banned: Abusing resources, spam, threats, egregious cheating, repeated (or egregious) violations of Kaggle's code of conduct, etc.
\nIf a user submits a Compliance appeal when there is significant evidence of cheating, we will ban the user's account.
\nThere have been instances when a user has made an appeal, and, upon further investigation, we've reversed the decision to remove that team from the leaderboard.
", "votes": 5, "replies": [] }, { "id": 1922989, "author_name": "JM", "author_url": "", "post_date": "2022-09-01T22:24:46.287000", "content": "Don't know why but it was so satisfying to see the cheater lie then get banned right after 😄
", "votes": 1, "replies": [] } ] } ], "raw_markdown_by_id": { "1920098": "I am fairly certain that Kaggle staff don't remove participants randomly. While I sympathize with your disappointment, raving about it and calling out hard-working admins will never work in your favor, either with them or with general community.\n\nI have no problem that they don't disclose their methodology, because the moment they do it someone will share that reason, the same way you shared their responses. That will help future offenders avoid detection, so of course they don't want to do it.\n\nIt is not difficult, however, to guess how they detect offenders. If multiple accounts are accessed from the same IP address, or if multiple accounts have similar submission names, similar submission times, it is likely that they belong to the same person. Please don't take this the wrong way, but I know there is no way that you didn't do anything wrong and they simply canned you. If that were the case, there would be many random competitors eliminated. Maybe you created an account 3 months ago, submitted a few times, then forgot the password. If you created a new account and accessed it from the same computer, that is a violation, even if you didn't do it for the purposes of cheating.", "1919068": "I just started my kaggle career and this is the second competition that I joined. I was at 12nd place but today when I again logged in, I was surprised to see that I was removed from the leaderboard. I never open any discussion before and just focus on improving my own model performance. After I submitted statement in compliance sector, I received the reply as follows without any other evidence:\n\nWe have completed our review and have determined that a member of your team violated competition rules through private sharing or using multiple accounts. We will not be reinstating your team. \n\nwhich is a meaningless and perfunctory reply. I don't even know what it means since I only have one account and never share anything with anyone. As a individual competitor in this competition, I don't even get in touch with anyone else. However, I received the email which told me that I received an ' unspecified' medal in competition. That's why I double check my account and knew I was removed without any reasons.\n\nI spent quite a lot time on this competition but even the staff is not responsible for the issue. Is it how you treat you participants? If kaggle is not even a fair and just platform, who will stay? ", "1919278": "This is a post from [another discussion](https://www.kaggle.com/competitions/amex-default-prediction/discussion/348093#1915476):\n\n> I had taken an image of the old leaderboard.\n> As you said, it looks like the 4th and 14th places on the original leaderboard are gone. Once again, the 13th and 14th places are exactly the same as the shake up.\n> \n> \n\nIt shows that you and place 14 had an identical score. Oddly enough, also a somewhat similar team name. Apparently, the scores were also identical on the public leaderboard. How do you think this happened?", "1920861": "", "1919685": "Some updates here. I email to kaggle staff Addison Howard and here is the response that I get. \n\nHi there,\n\nWe do not divulge details of how we determine if you shared privately or held multiple accounts - you may not share outside of Kaggle teams (including sharing before you merged), and you may not use multiple accounts to submit to a Kaggle competition.\n\nRegards,\n\nAddison Howard\n\nStill talking to air and not even a valid response. Also feel like a robot reply. If you can not disclose anything, then how can any participant regard kaggle competition as a fair and justified platform which has clear judgement criteria instead of a whole blackbox and bearing the risk of being removed from the leaderboard without any reasons after spending quite a lot of time and effort and the whole team is not responsible for it. " } }