Nine hours before the end of the competition, someone announced the notebook that can win the silver medal in the public ranking list. I am really disappointed about this. I feel that my efforts have been wasted. I believe many people will have the same idea. I hope that the employees of kaggle can make some actions, and we need a better environment
", "messages": [ { "id": 1039417, "postDate": "2020-10-06T15:04:42.183Z", "content": "Nine hours before the end of the competition, someone announced the notebook that can win the silver medal in the public ranking list. I am really disappointed about this. I feel that my efforts have been wasted. I believe many people will have the same idea. I hope that the employees of kaggle can make some actions, and we need a better environment
", "rawMarkdown": " Nine hours before the end of the competition, someone announced the notebook that can win the silver medal in the public ranking list. I am really disappointed about this. I feel that my efforts have been wasted. I believe many people will have the same idea. I hope that the employees of kaggle can make some actions, and we need a better environment", "votes": 12 }, { "id": 1039448, "postDate": "2020-10-06T15:25:58.697Z", "content": "I assume you're talking about this notebook. It is a fork of another public ensemble notebook which is not even properly done and filled with errors. I believe it wouldn't be within top 500 in private leaderboard. However, It's really disappointing to see how people publish high scoring kernel on the very last day/week to ruin the ongoing competitions.
", "rawMarkdown": "I assume you're talking about [this](https://www.kaggle.com/thebigd8ta/osic-ensemble-iv) notebook. It is a fork of another public ensemble notebook which is not even properly done and filled with errors. I believe it wouldn't be within top 500 in private leaderboard. However, It's really disappointing to see how people publish high scoring kernel on the very last day/week to ruin the ongoing competitions. ", "votes": 4 }, { "id": 1039454, "postDate": "2020-10-06T15:32:52.620Z", "content": "It's best if kaggle doesn't allow publishing such kernels in the last week.
", "rawMarkdown": "It's best if kaggle doesn't allow publishing such kernels in the last week.", "votes": 2 }, { "id": 1039800, "postDate": "2020-10-06T19:47:17.943Z", "content": "Don't be sad about something non-existent, that notebook is not gonna win any medals……👀👀
", "rawMarkdown": "Don't be sad about something non-existent, that notebook is not gonna win any medals......👀👀", "votes": 2, "replies": [ { "id": 1040783, "postDate": "2020-10-07T10:50:08.893Z", "content": "wisdom word!
", "rawMarkdown": "wisdom word!", "votes": 1 } ] }, { "id": 1039508, "postDate": "2020-10-06T16:17:49.750Z", "content": "We understand this is frustrating. Kaggle can't control the behavior of thousands of people and no amount of automatic checking can prevent somebody who is truly determined to publish a late idea. If we attempt to punish/block the behavior on the site, we risk creating an incentive for external publication (i.e. on Github), which is an even less fair and democratic outcome. If it's here, at least everyone sees it and can run it.
\nFor the record, our informal policy (https://www.kaggle.com/spoiler-alert) on last-minute notebooks is linked from a warning message on the notebooks page during the final week of competitions. It says this:
\nPublic sharing of high-scoring notebooks is highly disruptive and is strongly discouraged when a competition is nearing its deadline. We've placed warning messages to encourage participants to reconsider this behavior in the last week of a competition.\n\nKaggle created Notebooks for the purpose of educating and learning. However, when a high-scoring notebook is shared late in the competition, it leaves too little time for other competitors to understand those shared ideas. These scripts then become too easy for others to fork and submit, discrediting hard-earned time and work invested by others who become overshadowed in leaderboard placement.\nAgain, we completely agree this is frustrating and imperfect. We don't have a foolproof solution to stop it and are always open to ideas on how to improve adherence to this messaging.
", "rawMarkdown": "We understand this is frustrating. Kaggle can't control the behavior of thousands of people and no amount of automatic checking can prevent somebody who is truly determined to publish a late idea. If we attempt to punish/block the behavior on the site, we risk creating an incentive for external publication (i.e. on Github), which is an even less fair and democratic outcome. If it's here, at least everyone sees it and can run it.\n\nFor the record, our informal policy (https://www.kaggle.com/spoiler-alert) on last-minute notebooks is linked from a warning message on the notebooks page during the final week of competitions. It says this:\n```\nPublic sharing of high-scoring notebooks is highly disruptive and is strongly discouraged when a competition is nearing its deadline. We've placed warning messages to encourage participants to reconsider this behavior in the last week of a competition.\n\nKaggle created Notebooks for the purpose of educating and learning. However, when a high-scoring notebook is shared late in the competition, it leaves too little time for other competitors to understand those shared ideas. These scripts then become too easy for others to fork and submit, discrediting hard-earned time and work invested by others who become overshadowed in leaderboard placement.\n```\n\nAgain, we completely agree this is frustrating and imperfect. We don't have a foolproof solution to stop it and are always open to ideas on how to improve adherence to this messaging.", "votes": 2, "replies": [ { "id": 1039537, "postDate": "2020-10-06T16:28:43.270Z", "content": "@wcukierski I think there's a workaround. Sharing kernels anywhere outside Kaggle, should fall under private sharing as the rules state that:
\nIt's okay to share code if made available to all participants on the forums.\nIf a user uploads their kernel to an external platform other than Kaggle, they might not be complying with the rules stated above. This might be considered equivalent to private sharing.
", "rawMarkdown": "@wcukierski I think there's a workaround. Sharing kernels anywhere outside Kaggle, should fall under private sharing as the rules state that:\n\n```\nIt's okay to share code if made available to all participants on the forums.\n```\n\nIf a user uploads their kernel to an external platform other than Kaggle, they might not be complying with the rules stated above. This might be considered equivalent to private sharing." }, { "id": 1039580, "postDate": "2020-10-06T16:45:20.887Z", "content": "@aadhavvignesh agreed in principle, but the enforcement of such a standard is not feasible. We can't keep eyes on the entire web, in multiple languages, scanning for external notebooks, cross checking publication timelines against competition timelines, arbitrating reports where we have no join between a user on Kaggle to an associated user on some other platform.
", "rawMarkdown": "@aadhavvignesh agreed in principle, but the enforcement of such a standard is not feasible. We can't keep eyes on the entire web, in multiple languages, scanning for external notebooks, cross checking publication timelines against competition timelines, arbitrating reports where we have no join between a user on Kaggle to an associated user on some other platform." }, { "id": 1039587, "postDate": "2020-10-06T16:51:45.780Z", "content": "\n\ndisabling public sharing of kernels last week
\n
Totally agree, solves the problem! Lets implement and try it.
\nAnd the concern:
\n\n\nwe risk creating an incentive for external publication (i.e. on Github)
\n
That problem may already exist(?), we have seen in previous competitions that those who want to break the rules, cheat, share solutions outside Kaggle with or without payment will find all possible solution, is regardless another problem, let's solve this first and we all help to solve the others too, no one wants it, and many Kaggle users report what they find, but now we need your help with this behavior.
", "rawMarkdown": "> disabling public sharing of kernels last week \n\nTotally agree, solves the problem! Lets implement and try it.\n\nAnd the concern:\n> we risk creating an incentive for external publication (i.e. on Github)\n\nThat problem may already exist(?), we have seen in previous competitions that those who want to break the rules, cheat, share solutions outside Kaggle with or without payment will find all possible solution, is regardless another problem, let's solve this first and we all help to solve the others too, no one wants it, and many Kaggle users report what they find, but now we need your help with this behavior.", "votes": 4 }, { "id": 1039588, "postDate": "2020-10-06T16:52:41.090Z", "content": "Hi @wcukierski , do you think there's some way that members could be restricted to not publish such kernels once the last-minute warning message on the notebooks page during the final week of competitions is displayed. I know this would restrict exchanging ideas too in some way but there is plenty of time given for that in the competition.
", "rawMarkdown": "Hi @wcukierski , do you think there's some way that members could be restricted to not publish such kernels once the last-minute warning message on the notebooks page during the final week of competitions is displayed. I know this would restrict exchanging ideas too in some way but there is plenty of time given for that in the competition." }, { "id": 1039592, "postDate": "2020-10-06T16:54:46.420Z", "content": "@wcukierski I understand your concern. Let's hope people don't publish their high-scoring kernels at the competition, just for getting upvotes. Ethics matter; votes don't.
", "rawMarkdown": "@wcukierski I understand your concern. Let's hope people don't publish their high-scoring kernels at the competition, just for getting upvotes. Ethics matter; votes don't." }, { "id": 1039622, "postDate": "2020-10-06T17:18:21.827Z", "content": "Our original reasoning for not disabling publishing was that we don't truly know if a notebook is affiliated with a competition (we currently use the presence or absence of the official competition dataset for this purpose). In other words, if you made a notebook and were blocked from publishing, unmounting the competition dataset would allow you to publish it, albeit without it showing in the competition's notebook listing.
\nNonetheless, we will reconsider if we should trial this idea. If it provides a roadblock that prevents the people from accidentally breaking this policy, it could be worthwhile.
", "rawMarkdown": "Our original reasoning for not disabling publishing was that we don't truly know if a notebook is affiliated with a competition (we currently use the presence or absence of the official competition dataset for this purpose). In other words, if you made a notebook and were blocked from publishing, unmounting the competition dataset would allow you to publish it, albeit without it showing in the competition's notebook listing.\n\nNonetheless, we will reconsider if we should trial this idea. If it provides a roadblock that prevents the people from accidentally breaking this policy, it could be worthwhile.", "votes": 3 }, { "id": 1039636, "postDate": "2020-10-06T17:24:27.627Z", "content": "Thanks @wcukierski 👍👍, much appreciated😄
", "rawMarkdown": "Thanks @wcukierski 👍👍, much appreciated😄", "votes": 2 } ] }, { "id": 1039646, "postDate": "2020-10-06T17:34:48.980Z", "content": "I agree with @wcysysu, the problem of the last week high score public notebook is that some new users just fork and submit one notebook and simply going to top position of LB and it's not good for PPL who are working hard at this competition.
", "rawMarkdown": "I agree with @wcysysu, the problem of the last week high score public notebook is that some new users just fork and submit one notebook and simply going to top position of LB and it's not good for PPL who are working hard at this competition." }, { "id": 1041646, "postDate": "2020-10-07T21:23:33.323Z", "content": "Some Grandmasters said in there Kaggle days videos.\n- If Notebook perform best in Public Ranking does mean it won the Competition\n- Public Data may or may not be same as Private (Don't Worry)\n- Sometimes it can go more worse than that No Pattern between the\nPublic and Private Dataset
\n\nAlso One last Quote >> It may help u
\n\n\n", "rawMarkdown": "Some Grandmasters said in there Kaggle days videos.\n- If Notebook perform best in Public Ranking does mean it won the Competition\n- Public Data may or may not be same as Private (Don't Worry)\n- Sometimes it can go more worse than that No Pattern between the\nPublic and Private Dataset\n\nAlso One last Quote >> It may help u\n> The best yardstick for our progress is not other people, but ourselves. Am I better than I was yesterday? This is the only question worth asking. As long as you go to bed at night a better practitioner than the one who woke up that morning, you have succeeded. Your worth should have nothing to do with how your progress stacks up relative to another \nGood Luck for your Journey🎉 \nGet to Work 🔥 🔥 " }, { "id": 1039517, "postDate": "2020-10-06T16:20:58.920Z", "content": "The best yardstick for our progress is not other people, but ourselves. Am I better than I was yesterday? This is the only question worth asking. As long as you go to bed at night a better practitioner than the one who woke up that morning, you have succeeded. Your worth should have nothing to do with how your progress stacks up relative to another
\n
\n Good Luck for your Journey🎉 \n Get to Work 🔥 🔥
@wcysysu Don't worry. I've tagged the mods, and I know that they would take this seriously.
", "rawMarkdown": "@wcysysu Don't worry. I've tagged the mods, and I know that they would take this seriously." }, { "id": 1039944, "postDate": "2020-10-06T23:13:52.157Z", "rawMarkdown": "", "votes": 1, "isDeleted": true }, { "id": 1039828, "postDate": "2020-10-06T20:10:44.947Z", "rawMarkdown": "", "votes": -1, "isDeleted": true }, { "id": 1039601, "postDate": "2020-10-06T17:00:56.347Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 1039448, "author_name": "Nischay Dhankhar", "author_url": "", "post_date": "2020-10-06T15:25:58.697000", "content": "I assume you're talking about this notebook. It is a fork of another public ensemble notebook which is not even properly done and filled with errors. I believe it wouldn't be within top 500 in private leaderboard. However, It's really disappointing to see how people publish high scoring kernel on the very last day/week to ruin the ongoing competitions.
", "votes": 4, "replies": [] }, { "id": 1039454, "author_name": "NoMalady", "author_url": "", "post_date": "2020-10-06T15:32:52.620000", "content": "It's best if kaggle doesn't allow publishing such kernels in the last week.
", "votes": 2, "replies": [] }, { "id": 1039800, "author_name": "sin", "author_url": "", "post_date": "2020-10-06T19:47:17.943000", "content": "Don't be sad about something non-existent, that notebook is not gonna win any medals……👀👀
", "votes": 2, "replies": [ { "id": 1040783, "author_name": "Ji Wong Park", "author_url": "", "post_date": "2020-10-07T10:50:08.893000", "content": "wisdom word!
", "votes": 1, "replies": [] } ] }, { "id": 1039508, "author_name": "Will Cukierski", "author_url": "", "post_date": "2020-10-06T16:17:49.750000", "content": "We understand this is frustrating. Kaggle can't control the behavior of thousands of people and no amount of automatic checking can prevent somebody who is truly determined to publish a late idea. If we attempt to punish/block the behavior on the site, we risk creating an incentive for external publication (i.e. on Github), which is an even less fair and democratic outcome. If it's here, at least everyone sees it and can run it.
\nFor the record, our informal policy (https://www.kaggle.com/spoiler-alert) on last-minute notebooks is linked from a warning message on the notebooks page during the final week of competitions. It says this:
\nPublic sharing of high-scoring notebooks is highly disruptive and is strongly discouraged when a competition is nearing its deadline. We've placed warning messages to encourage participants to reconsider this behavior in the last week of a competition.\n\nKaggle created Notebooks for the purpose of educating and learning. However, when a high-scoring notebook is shared late in the competition, it leaves too little time for other competitors to understand those shared ideas. These scripts then become too easy for others to fork and submit, discrediting hard-earned time and work invested by others who become overshadowed in leaderboard placement.\nAgain, we completely agree this is frustrating and imperfect. We don't have a foolproof solution to stop it and are always open to ideas on how to improve adherence to this messaging.
", "votes": 2, "replies": [ { "id": 1039537, "author_name": "Aadhav Vignesh", "author_url": "", "post_date": "2020-10-06T16:28:43.270000", "content": "@wcukierski I think there's a workaround. Sharing kernels anywhere outside Kaggle, should fall under private sharing as the rules state that:
\nIt's okay to share code if made available to all participants on the forums.\nIf a user uploads their kernel to an external platform other than Kaggle, they might not be complying with the rules stated above. This might be considered equivalent to private sharing.
", "votes": 0, "replies": [] }, { "id": 1039580, "author_name": "Will Cukierski", "author_url": "", "post_date": "2020-10-06T16:45:20.887000", "content": "@aadhavvignesh agreed in principle, but the enforcement of such a standard is not feasible. We can't keep eyes on the entire web, in multiple languages, scanning for external notebooks, cross checking publication timelines against competition timelines, arbitrating reports where we have no join between a user on Kaggle to an associated user on some other platform.
", "votes": 0, "replies": [] }, { "id": 1039587, "author_name": "Kirderf", "author_url": "", "post_date": "2020-10-06T16:51:45.780000", "content": "\n\ndisabling public sharing of kernels last week
\n
Totally agree, solves the problem! Lets implement and try it.
\nAnd the concern:
\n\n\nwe risk creating an incentive for external publication (i.e. on Github)
\n
That problem may already exist(?), we have seen in previous competitions that those who want to break the rules, cheat, share solutions outside Kaggle with or without payment will find all possible solution, is regardless another problem, let's solve this first and we all help to solve the others too, no one wants it, and many Kaggle users report what they find, but now we need your help with this behavior.
", "votes": 4, "replies": [] }, { "id": 1039588, "author_name": "NoMalady", "author_url": "", "post_date": "2020-10-06T16:52:41.090000", "content": "Hi @wcukierski , do you think there's some way that members could be restricted to not publish such kernels once the last-minute warning message on the notebooks page during the final week of competitions is displayed. I know this would restrict exchanging ideas too in some way but there is plenty of time given for that in the competition.
", "votes": 0, "replies": [] }, { "id": 1039592, "author_name": "Aadhav Vignesh", "author_url": "", "post_date": "2020-10-06T16:54:46.420000", "content": "@wcukierski I understand your concern. Let's hope people don't publish their high-scoring kernels at the competition, just for getting upvotes. Ethics matter; votes don't.
", "votes": 0, "replies": [] }, { "id": 1039622, "author_name": "Will Cukierski", "author_url": "", "post_date": "2020-10-06T17:18:21.827000", "content": "Our original reasoning for not disabling publishing was that we don't truly know if a notebook is affiliated with a competition (we currently use the presence or absence of the official competition dataset for this purpose). In other words, if you made a notebook and were blocked from publishing, unmounting the competition dataset would allow you to publish it, albeit without it showing in the competition's notebook listing.
\nNonetheless, we will reconsider if we should trial this idea. If it provides a roadblock that prevents the people from accidentally breaking this policy, it could be worthwhile.
", "votes": 3, "replies": [] }, { "id": 1039636, "author_name": "NoMalady", "author_url": "", "post_date": "2020-10-06T17:24:27.627000", "content": "Thanks @wcukierski 👍👍, much appreciated😄
", "votes": 2, "replies": [] } ] }, { "id": 1039646, "author_name": "", "author_url": "", "post_date": "2020-10-06T17:34:48.980000", "content": "I agree with @wcysysu, the problem of the last week high score public notebook is that some new users just fork and submit one notebook and simply going to top position of LB and it's not good for PPL who are working hard at this competition.
", "votes": 0, "replies": [] }, { "id": 1041646, "author_name": "Walia", "author_url": "", "post_date": "2020-10-07T21:23:33.323000", "content": "Some Grandmasters said in there Kaggle days videos.\n- If Notebook perform best in Public Ranking does mean it won the Competition\n- Public Data may or may not be same as Private (Don't Worry)\n- Sometimes it can go more worse than that No Pattern between the\nPublic and Private Dataset
\n\nAlso One last Quote >> It may help u
\n\n\n", "votes": 0, "replies": [] }, { "id": 1039517, "author_name": "Aadhav Vignesh", "author_url": "", "post_date": "2020-10-06T16:20:58.920000", "content": "The best yardstick for our progress is not other people, but ourselves. Am I better than I was yesterday? This is the only question worth asking. As long as you go to bed at night a better practitioner than the one who woke up that morning, you have succeeded. Your worth should have nothing to do with how your progress stacks up relative to another
\n
\n Good Luck for your Journey🎉 \n Get to Work 🔥 🔥
@wcysysu Don't worry. I've tagged the mods, and I know that they would take this seriously.
", "votes": 0, "replies": [] }, { "id": 1039944, "author_name": "", "author_url": "", "post_date": "2020-10-06T23:13:52.157000", "content": "", "votes": 1, "replies": [] }, { "id": 1039828, "author_name": "", "author_url": "", "post_date": "2020-10-06T20:10:44.947000", "content": "", "votes": -1, "replies": [] }, { "id": 1039601, "author_name": "", "author_url": "", "post_date": "2020-10-06T17:00:56.347000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "1039417": " Nine hours before the end of the competition, someone announced the notebook that can win the silver medal in the public ranking list. I am really disappointed about this. I feel that my efforts have been wasted. I believe many people will have the same idea. I hope that the employees of kaggle can make some actions, and we need a better environment", "1039448": "I assume you're talking about [this](https://www.kaggle.com/thebigd8ta/osic-ensemble-iv) notebook. It is a fork of another public ensemble notebook which is not even properly done and filled with errors. I believe it wouldn't be within top 500 in private leaderboard. However, It's really disappointing to see how people publish high scoring kernel on the very last day/week to ruin the ongoing competitions. ", "1039454": "It's best if kaggle doesn't allow publishing such kernels in the last week.", "1039800": "Don't be sad about something non-existent, that notebook is not gonna win any medals......👀👀", "1039508": "We understand this is frustrating. Kaggle can't control the behavior of thousands of people and no amount of automatic checking can prevent somebody who is truly determined to publish a late idea. If we attempt to punish/block the behavior on the site, we risk creating an incentive for external publication (i.e. on Github), which is an even less fair and democratic outcome. If it's here, at least everyone sees it and can run it.\n\nFor the record, our informal policy (https://www.kaggle.com/spoiler-alert) on last-minute notebooks is linked from a warning message on the notebooks page during the final week of competitions. It says this:\n```\nPublic sharing of high-scoring notebooks is highly disruptive and is strongly discouraged when a competition is nearing its deadline. We've placed warning messages to encourage participants to reconsider this behavior in the last week of a competition.\n\nKaggle created Notebooks for the purpose of educating and learning. However, when a high-scoring notebook is shared late in the competition, it leaves too little time for other competitors to understand those shared ideas. These scripts then become too easy for others to fork and submit, discrediting hard-earned time and work invested by others who become overshadowed in leaderboard placement.\n```\n\nAgain, we completely agree this is frustrating and imperfect. We don't have a foolproof solution to stop it and are always open to ideas on how to improve adherence to this messaging.", "1039646": "I agree with @wcysysu, the problem of the last week high score public notebook is that some new users just fork and submit one notebook and simply going to top position of LB and it's not good for PPL who are working hard at this competition.", "1041646": "Some Grandmasters said in there Kaggle days videos.\n- If Notebook perform best in Public Ranking does mean it won the Competition\n- Public Data may or may not be same as Private (Don't Worry)\n- Sometimes it can go more worse than that No Pattern between the\nPublic and Private Dataset\n\nAlso One last Quote >> It may help u\n> The best yardstick for our progress is not other people, but ourselves. Am I better than I was yesterday? This is the only question worth asking. As long as you go to bed at night a better practitioner than the one who woke up that morning, you have succeeded. Your worth should have nothing to do with how your progress stacks up relative to another \nGood Luck for your Journey🎉 \nGet to Work 🔥 🔥 ", "1039517": "@wcysysu Don't worry. I've tagged the mods, and I know that they would take this seriously.", "1039944": "", "1039828": "", "1039601": "" } }