I think everyone knows this competition is a joke, that has been more than proven.
\nWhat has not been discussed enough, is Kaggle implications.
\nI understand this is a business, and a business should make money, but by letting inexperienced (stubborn, even when proven wrong) people run a competition on your platform, you lose something way more valuable than whatever money you got.
\nYou have lost reputation, and you wont get that one back.
\n", "messages": [ { "id": 2822229, "postDate": "2024-05-18T13:54:17.307Z", "content": "I think everyone knows this competition is a joke, that has been more than proven.
\nWhat has not been discussed enough, is Kaggle implications.
\nI understand this is a business, and a business should make money, but by letting inexperienced (stubborn, even when proven wrong) people run a competition on your platform, you lose something way more valuable than whatever money you got.
\nYou have lost reputation, and you wont get that one back.
\n", "rawMarkdown": "I think everyone knows this competition is a joke, that has been more than proven.\n\nWhat has not been discussed enough, is Kaggle implications.\n\nI understand this is a business, and a business should make money, but by letting inexperienced (stubborn, even when proven wrong) people run a competition on your platform, you lose something way more valuable than whatever money you got.\n\nYou have lost reputation, and you wont get that one back.\n\n", "votes": 32 }, { "id": 2822242, "postDate": "2024-05-18T14:00:39.760Z", "content": "The same goes for \"Home Credit\". People don't even want to hear that name anymore.
", "rawMarkdown": "The same goes for \"Home Credit\". People don't even want to hear that name anymore.", "votes": 5, "replies": [ { "id": 2823023, "postDate": "2024-05-19T01:05:18.280Z", "content": "To be honest, in the early stages of the competition, I learned a lot of methods for handling large data and error handling, and later on, everyone knows about them.
", "rawMarkdown": "To be honest, in the early stages of the competition, I learned a lot of methods for handling large data and error handling, and later on, everyone knows about them.", "votes": 3, "replies": [ { "id": 2823025, "postDate": "2024-05-19T01:09:07.277Z", "content": "I had considered entering the competition in the early stages, but his advice changed things. Now it's down to ICR Episode 2.
", "rawMarkdown": "I had considered entering the competition in the early stages, but [his advice](https://www.kaggle.com/competitions/home-credit-credit-risk-model-stability/discussion/476449) changed things. Now it's down to ICR Episode 2.", "votes": 1 }, { "id": 2823225, "postDate": "2024-05-19T05:01:09.973Z", "content": "@mitsuyasuhoshino usually metrics engineered with constants in them don't generalize well. This is a live example for that
", "rawMarkdown": "@mitsuyasuhoshino usually metrics engineered with constants in them don't generalize well. This is a live example for that", "votes": 3 } ] } ] }, { "id": 2827088, "postDate": "2024-05-21T08:19:46.117Z", "content": "You know the situation's bad when people start busting out the Warren Buffet YT vids đź’€
", "rawMarkdown": "You know the situation's bad when people start busting out the Warren Buffet YT vids 💀", "votes": 1 }, { "id": 2822481, "postDate": "2024-05-18T16:03:20.007Z", "content": "Don’t understand what you expect Kaggle to do? The host is the one paying you can’t reasonably expect them to tell the hosts to change their metric..
", "rawMarkdown": "Don’t understand what you expect Kaggle to do? The host is the one paying you can’t reasonably expect them to tell the hosts to change their metric..", "replies": [ { "id": 2822499, "postDate": "2024-05-18T16:24:58.503Z", "content": "That's the funny thing, it's really not my problem, it's Kaggle's.
\nIt doesn't really change the end result for them, reputation is lost.
Kaggle has the responsibility to reject a proposal if it makes no scientific sense, and this one doesn't and it was super clear since day 1.
", "rawMarkdown": "That's the funny thing, it's really not my problem, it's Kaggle's.\nIt doesn't really change the end result for them, reputation is lost.\n\nKaggle has the responsibility to reject a proposal if it makes no scientific sense, and this one doesn't and it was super clear since day 1.", "votes": 3, "replies": [ { "id": 2822989, "postDate": "2024-05-19T00:15:48.943Z", "content": "I don't believe projects are that closely scrutinized by Kaggle as you might think. And I also don't think they are worried about losing reputation b/c to my knowledge they are \"the only game in town\" of this caliber… nothing comes even close in terms of data science playgrounds that connect people from all over the world.
\nThis is not the only sketchy thing that I saw Kaggle allowing. For example, there are people out there publishing \"Explained\" notebooks based on high scoring notebooks without writing a single line of code… they simply copy/paste stuff in LLMs and copy/paste the response back into a notebook… 2 months later they're \"Masters\" or \"GrandMasters\" b/c of all the upvotes they got from other individuals. There are other issues as well… too long of a story.
\nI come from a different field other than data science, and once I joined I was amazed of the stuff that goes on on this platform. However, I take it for what it is… a playground, nothing more. Very little stuff that goes on in here would be permissible in real life… at least in the civilized world. Anyway… in short, I don't think they'll lose any reputation over this. People like Warren Buffet, and until recently Charlie Munger, play in the real world and cannot afford to fool around… Kaggle is not the real world… it's a virtual reality at best.
", "rawMarkdown": "I don't believe projects are that closely scrutinized by Kaggle as you might think. And I also don't think they are worried about losing reputation b/c to my knowledge they are \"the only game in town\" of this caliber... nothing comes even close in terms of data science playgrounds that connect people from all over the world.\n\nThis is not the only sketchy thing that I saw Kaggle allowing. For example, there are people out there publishing \"Explained\" notebooks based on high scoring notebooks without writing a single line of code... they simply copy/paste stuff in LLMs and copy/paste the response back into a notebook... 2 months later they're \"Masters\" or \"GrandMasters\" b/c of all the upvotes they got from other individuals. There are other issues as well... too long of a story.\n\nI come from a different field other than data science, and once I joined I was amazed of the stuff that goes on on this platform. However, I take it for what it is... a playground, nothing more. Very little stuff that goes on in here would be permissible in real life... at least in the civilized world. Anyway... in short, I don't think they'll lose any reputation over this. People like Warren Buffet, and until recently Charlie Munger, play in the real world and cannot afford to fool around... Kaggle is not the real world... it's a virtual reality at best. ", "votes": 5 }, { "id": 2823390, "postDate": "2024-05-19T06:49:32.190Z", "content": "This is a very good point about \"Masters\" or \"Grand Masters\" for \"Explained\" notebooks. Gold for a high-scoring fork and silver for a useless ensemble in the first weeks after the start - I'm tired of this disappointment.
\nI also noticed a terrible situation in the Discussions section. You see, Kaggle launched a space where people can share their \"achievements\". And what is going on here? Someone become Grand Master just for nothing in a few months. They write posts where boast rank rise literally by every 100 points and receive a gold medal for describing this \"achievement,\" while this dubious gold itself is \"achievement\", which will be presented in the next silly post. This is ridiculous and insane! And all this polite comments with fake appreciation, twisted goodwill and zero value. What happened? Is this a real competition in the current world?
\nSame situation in the Datasets section. I think, datasets aren't estimated by usefulness, upvotes go to people who can upvote back. Thus, a group of individuals rises up by a circle vote. Moreover, in recent times, so many competitions have been paused and exposed to correction. It looks like become too difficult to take care of us for Kaggle, I mean competition culture. And it seems to be okay, because we are just a bunch of volunteers here, nothing more.
", "rawMarkdown": "This is a very good point about \"Masters\" or \"Grand Masters\" for \"Explained\" notebooks. Gold for a high-scoring fork and silver for a useless ensemble in the first weeks after the start - I'm tired of this disappointment. \n\nI also noticed a terrible situation in the Discussions section. You see, Kaggle launched a space where people can share their \"achievements\". And what is going on here? Someone become Grand Master just for nothing in a few months. They write posts where boast rank rise literally by every 100 points and receive a gold medal for describing this \"achievement,\" while this dubious gold itself is \"achievement\", which will be presented in the next silly post. This is ridiculous and insane! And all this polite comments with fake appreciation, twisted goodwill and zero value. What happened? Is this a real competition in the current world?\n\nSame situation in the Datasets section. I think, datasets aren't estimated by usefulness, upvotes go to people who can upvote back. Thus, a group of individuals rises up by a circle vote. Moreover, in recent times, so many competitions have been paused and exposed to correction. It looks like become too difficult to take care of us for Kaggle, I mean competition culture. And it seems to be okay, because we are just a bunch of volunteers here, nothing more.", "votes": 13 }, { "id": 2825535, "postDate": "2024-05-20T13:08:29.850Z", "content": "I agree 100% with what you said @yekenot
\nThis is precisely why Kaggle should not be treated as more than what it actually is… a simple playground, nothing more. Maybe in the beginning it had more credence, but today I'm not sure. Once you join you realize how things work around here. In fairness, it is not entirely their fault… these days people will try to abuse anything, they'll try find the weak spot and take advantage of it.
I see your advice @eduardastefanescu to accept things as it is now and reconcile, but I can’t resist and would add some.
\nMany people came here to learn, not for medals and tiers at first. But to reflect the learning process, there needs to be a healthy progression system. Consider Kaggle as a large database consisting of tons of brilliant expertise and tons of gambling outcomes. What point is it if valuable stuff is estimated to be the same as someone's attempt to self-promote? We may look at Kaggle as a playground, yes, but we must not forget those who make this game truly interesting and educational. There are researchers here who dedicate a lot of effort to extract some value and share it with us, just for scientific interest. I feel sorry for them. They struggle to overcome the disbalance produced by gamblers who exploit vulnerabilities.
\nWhen I dig into some stuff from Kaggle, I realize how much trash needs to be sorted and reevaluated because the knowledge base becomes corrupted because real contributions are underrated while jokes or fancy titles attract respect. I don't even understand whose point of view I can trust now. I should check if this is a true science leader or if it's a leader of gamblers with high tiers who become them by begging for votes, spamming anywhere with the same phrases, or \"celebrating\" another 100 views of their notebook. Or a non-detected cheater, or a metric hacker, or someone lucky who rose to the top of leaderboard because of imperfect competition conditions. This is very frustrating, sorry.
\nMaybe I too dramatize, maybe say idealistic things, but I really like this platform and want it to be organic and fair. I would say Kaggle should and can be a perfect cognitive environment for experiments, not a service with a dopamine button. A place where compete and evolve innovative ideas and technologies by themselves, not someone's ambitions and ego.
", "rawMarkdown": "I see your advice @eduardastefanescu to accept things as it is now and reconcile, but I can’t resist and would add some.\n\nMany people came here to learn, not for medals and tiers at first. But to reflect the learning process, there needs to be a healthy progression system. Consider Kaggle as a large database consisting of tons of brilliant expertise and tons of gambling outcomes. What point is it if valuable stuff is estimated to be the same as someone's attempt to self-promote? We may look at Kaggle as a playground, yes, but we must not forget those who make this game truly interesting and educational. There are researchers here who dedicate a lot of effort to extract some value and share it with us, just for scientific interest. I feel sorry for them. They struggle to overcome the disbalance produced by gamblers who exploit vulnerabilities.\n\nWhen I dig into some stuff from Kaggle, I realize how much trash needs to be sorted and reevaluated because the knowledge base becomes corrupted because real contributions are underrated while jokes or fancy titles attract respect. I don't even understand whose point of view I can trust now. I should check if this is a true science leader or if it's a leader of gamblers with high tiers who become them by begging for votes, spamming anywhere with the same phrases, or \"celebrating\" another 100 views of their notebook. Or a non-detected cheater, or a metric hacker, or someone lucky who rose to the top of leaderboard because of imperfect competition conditions. This is very frustrating, sorry.\n\nMaybe I too dramatize, maybe say idealistic things, but I really like this platform and want it to be organic and fair. I would say Kaggle should and can be a perfect cognitive environment for experiments, not a service with a dopamine button. A place where compete and evolve innovative ideas and technologies by themselves, not someone's ambitions and ego.", "votes": 3 }, { "id": 2826164, "postDate": "2024-05-20T18:37:00.863Z", "content": "@yekenot … I hate to be the cynic here, but you should first think of Kaggle as a business, and to a lesser extent as a provider for opportunities to learn. I'm not saying they don't provide opportunities to learn, because they do. However, they are a business first, and as a business their main goal is to make as much money as possible. I'll give you a simple example out of a multitude of examples… there is a clear reason why there is no interest from Kaggle to crack down on sharing of high scoring notebooks just 10days before the end of a competition… because such sharing generates insane traffic on the platform. Everybody suddenly rushes in to fork and re-run the notebook in order to achieve a higher score. This is just one example… I'm sure you can find multiple similar examples in other segments of Kaggle such as discussions, datasets, models, etc, etc. Kaggle is a revenue generating business first, and a provider for \"opportunities to learn\" second.
", "rawMarkdown": "@yekenot ... I hate to be the cynic here, but you should first think of Kaggle as a business, and to a lesser extent as a provider for opportunities to learn. I'm not saying they don't provide opportunities to learn, because they do. However, they are a business first, and as a business their main goal is to make as much money as possible. I'll give you a simple example out of a multitude of examples... there is a clear reason why there is no interest from Kaggle to crack down on sharing of high scoring notebooks just 10days before the end of a competition... because such sharing generates insane traffic on the platform. Everybody suddenly rushes in to fork and re-run the notebook in order to achieve a higher score. This is just one example... I'm sure you can find multiple similar examples in other segments of Kaggle such as discussions, datasets, models, etc, etc. Kaggle is a revenue generating business first, and a provider for \"opportunities to learn\" second.", "votes": 1 }, { "id": 2826233, "postDate": "2024-05-20T19:39:42.273Z", "content": "It's hard to disagree with you, man. I forgot just a while in which world we are living in and who actually rules. It's a dark side that I do not want to see, focusing only on a positive utopian vision. I believe most people who made breakthrough discoveries think about money in last order, concentrating attention on research. They were humble and just did what they loved to do, like who first solved mathematical problem from the seven millennium problems, for example. Needless to say, the true treasure does not need advertising. When it comes, it captures all attention, while useless stuff needs promotion, money, and points for self-affirmation, because without money or rating, it's nothing. But yeah, Kaggle is, at some point, just a business in a highly competitive world, and it's going to need money at least to survive.
", "rawMarkdown": "It's hard to disagree with you, man. I forgot just a while in which world we are living in and who actually rules. It's a dark side that I do not want to see, focusing only on a positive utopian vision. I believe most people who made breakthrough discoveries think about money in last order, concentrating attention on research. They were humble and just did what they loved to do, like who first solved mathematical problem from the seven millennium problems, for example. Needless to say, the true treasure does not need advertising. When it comes, it captures all attention, while useless stuff needs promotion, money, and points for self-affirmation, because without money or rating, it's nothing. But yeah, Kaggle is, at some point, just a business in a highly competitive world, and it's going to need money at least to survive.", "votes": 3 }, { "id": 2827845, "postDate": "2024-05-21T17:22:35.800Z", "content": "Your words matters. @eduardastefanescu
", "rawMarkdown": "Your words matters. @eduardastefanescu " } ] } ] }, { "id": 2822465, "postDate": "2024-05-18T15:50:40.273Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 2822242, "author_name": "mh", "author_url": "", "post_date": "2024-05-18T14:00:39.760000", "content": "The same goes for \"Home Credit\". People don't even want to hear that name anymore.
", "votes": 5, "replies": [ { "id": 2823023, "author_name": "yunsuxiaozi", "author_url": "", "post_date": "2024-05-19T01:05:18.280000", "content": "To be honest, in the early stages of the competition, I learned a lot of methods for handling large data and error handling, and later on, everyone knows about them.
", "votes": 3, "replies": [ { "id": 2823025, "author_name": "mh", "author_url": "", "post_date": "2024-05-19T01:09:07.277000", "content": "I had considered entering the competition in the early stages, but his advice changed things. Now it's down to ICR Episode 2.
", "votes": 1, "replies": [] }, { "id": 2823225, "author_name": "Ravi Ramakrishnan", "author_url": "", "post_date": "2024-05-19T05:01:09.973000", "content": "@mitsuyasuhoshino usually metrics engineered with constants in them don't generalize well. This is a live example for that
", "votes": 3, "replies": [] } ] } ] }, { "id": 2827088, "author_name": "Taizhuo Tang", "author_url": "", "post_date": "2024-05-21T08:19:46.117000", "content": "You know the situation's bad when people start busting out the Warren Buffet YT vids đź’€
", "votes": 1, "replies": [] }, { "id": 2822481, "author_name": "JM", "author_url": "", "post_date": "2024-05-18T16:03:20.007000", "content": "Don’t understand what you expect Kaggle to do? The host is the one paying you can’t reasonably expect them to tell the hosts to change their metric..
", "votes": 0, "replies": [ { "id": 2822499, "author_name": "NxGTR", "author_url": "", "post_date": "2024-05-18T16:24:58.503000", "content": "That's the funny thing, it's really not my problem, it's Kaggle's.
\nIt doesn't really change the end result for them, reputation is lost.
Kaggle has the responsibility to reject a proposal if it makes no scientific sense, and this one doesn't and it was super clear since day 1.
", "votes": 3, "replies": [ { "id": 2822989, "author_name": "Eduard Stefanescu", "author_url": "", "post_date": "2024-05-19T00:15:48.943000", "content": "I don't believe projects are that closely scrutinized by Kaggle as you might think. And I also don't think they are worried about losing reputation b/c to my knowledge they are \"the only game in town\" of this caliber… nothing comes even close in terms of data science playgrounds that connect people from all over the world.
\nThis is not the only sketchy thing that I saw Kaggle allowing. For example, there are people out there publishing \"Explained\" notebooks based on high scoring notebooks without writing a single line of code… they simply copy/paste stuff in LLMs and copy/paste the response back into a notebook… 2 months later they're \"Masters\" or \"GrandMasters\" b/c of all the upvotes they got from other individuals. There are other issues as well… too long of a story.
\nI come from a different field other than data science, and once I joined I was amazed of the stuff that goes on on this platform. However, I take it for what it is… a playground, nothing more. Very little stuff that goes on in here would be permissible in real life… at least in the civilized world. Anyway… in short, I don't think they'll lose any reputation over this. People like Warren Buffet, and until recently Charlie Munger, play in the real world and cannot afford to fool around… Kaggle is not the real world… it's a virtual reality at best.
", "votes": 5, "replies": [] }, { "id": 2823390, "author_name": "Vladimir Demidov", "author_url": "", "post_date": "2024-05-19T06:49:32.190000", "content": "This is a very good point about \"Masters\" or \"Grand Masters\" for \"Explained\" notebooks. Gold for a high-scoring fork and silver for a useless ensemble in the first weeks after the start - I'm tired of this disappointment.
\nI also noticed a terrible situation in the Discussions section. You see, Kaggle launched a space where people can share their \"achievements\". And what is going on here? Someone become Grand Master just for nothing in a few months. They write posts where boast rank rise literally by every 100 points and receive a gold medal for describing this \"achievement,\" while this dubious gold itself is \"achievement\", which will be presented in the next silly post. This is ridiculous and insane! And all this polite comments with fake appreciation, twisted goodwill and zero value. What happened? Is this a real competition in the current world?
\nSame situation in the Datasets section. I think, datasets aren't estimated by usefulness, upvotes go to people who can upvote back. Thus, a group of individuals rises up by a circle vote. Moreover, in recent times, so many competitions have been paused and exposed to correction. It looks like become too difficult to take care of us for Kaggle, I mean competition culture. And it seems to be okay, because we are just a bunch of volunteers here, nothing more.
", "votes": 13, "replies": [] }, { "id": 2825535, "author_name": "Eduard Stefanescu", "author_url": "", "post_date": "2024-05-20T13:08:29.850000", "content": "I agree 100% with what you said @yekenot
\nThis is precisely why Kaggle should not be treated as more than what it actually is… a simple playground, nothing more. Maybe in the beginning it had more credence, but today I'm not sure. Once you join you realize how things work around here. In fairness, it is not entirely their fault… these days people will try to abuse anything, they'll try find the weak spot and take advantage of it.
I see your advice @eduardastefanescu to accept things as it is now and reconcile, but I can’t resist and would add some.
\nMany people came here to learn, not for medals and tiers at first. But to reflect the learning process, there needs to be a healthy progression system. Consider Kaggle as a large database consisting of tons of brilliant expertise and tons of gambling outcomes. What point is it if valuable stuff is estimated to be the same as someone's attempt to self-promote? We may look at Kaggle as a playground, yes, but we must not forget those who make this game truly interesting and educational. There are researchers here who dedicate a lot of effort to extract some value and share it with us, just for scientific interest. I feel sorry for them. They struggle to overcome the disbalance produced by gamblers who exploit vulnerabilities.
\nWhen I dig into some stuff from Kaggle, I realize how much trash needs to be sorted and reevaluated because the knowledge base becomes corrupted because real contributions are underrated while jokes or fancy titles attract respect. I don't even understand whose point of view I can trust now. I should check if this is a true science leader or if it's a leader of gamblers with high tiers who become them by begging for votes, spamming anywhere with the same phrases, or \"celebrating\" another 100 views of their notebook. Or a non-detected cheater, or a metric hacker, or someone lucky who rose to the top of leaderboard because of imperfect competition conditions. This is very frustrating, sorry.
\nMaybe I too dramatize, maybe say idealistic things, but I really like this platform and want it to be organic and fair. I would say Kaggle should and can be a perfect cognitive environment for experiments, not a service with a dopamine button. A place where compete and evolve innovative ideas and technologies by themselves, not someone's ambitions and ego.
", "votes": 3, "replies": [] }, { "id": 2826164, "author_name": "Eduard Stefanescu", "author_url": "", "post_date": "2024-05-20T18:37:00.863000", "content": "@yekenot … I hate to be the cynic here, but you should first think of Kaggle as a business, and to a lesser extent as a provider for opportunities to learn. I'm not saying they don't provide opportunities to learn, because they do. However, they are a business first, and as a business their main goal is to make as much money as possible. I'll give you a simple example out of a multitude of examples… there is a clear reason why there is no interest from Kaggle to crack down on sharing of high scoring notebooks just 10days before the end of a competition… because such sharing generates insane traffic on the platform. Everybody suddenly rushes in to fork and re-run the notebook in order to achieve a higher score. This is just one example… I'm sure you can find multiple similar examples in other segments of Kaggle such as discussions, datasets, models, etc, etc. Kaggle is a revenue generating business first, and a provider for \"opportunities to learn\" second.
", "votes": 1, "replies": [] }, { "id": 2826233, "author_name": "Vladimir Demidov", "author_url": "", "post_date": "2024-05-20T19:39:42.273000", "content": "It's hard to disagree with you, man. I forgot just a while in which world we are living in and who actually rules. It's a dark side that I do not want to see, focusing only on a positive utopian vision. I believe most people who made breakthrough discoveries think about money in last order, concentrating attention on research. They were humble and just did what they loved to do, like who first solved mathematical problem from the seven millennium problems, for example. Needless to say, the true treasure does not need advertising. When it comes, it captures all attention, while useless stuff needs promotion, money, and points for self-affirmation, because without money or rating, it's nothing. But yeah, Kaggle is, at some point, just a business in a highly competitive world, and it's going to need money at least to survive.
", "votes": 3, "replies": [] }, { "id": 2827845, "author_name": "Muhammed Tausif", "author_url": "", "post_date": "2024-05-21T17:22:35.800000", "content": "Your words matters. @eduardastefanescu
", "votes": 0, "replies": [] } ] } ] }, { "id": 2822465, "author_name": "", "author_url": "", "post_date": "2024-05-18T15:50:40.273000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "2822229": "I think everyone knows this competition is a joke, that has been more than proven.\n\nWhat has not been discussed enough, is Kaggle implications.\n\nI understand this is a business, and a business should make money, but by letting inexperienced (stubborn, even when proven wrong) people run a competition on your platform, you lose something way more valuable than whatever money you got.\n\nYou have lost reputation, and you wont get that one back.\n\n", "2822242": "The same goes for \"Home Credit\". People don't even want to hear that name anymore.", "2827088": "You know the situation's bad when people start busting out the Warren Buffet YT vids 💀", "2822481": "Don’t understand what you expect Kaggle to do? The host is the one paying you can’t reasonably expect them to tell the hosts to change their metric..", "2822465": "" } }