Is it possible to have a few more digits on the leaderboard? A lot of us are on similar numbers, and seeing on the leaderboard (and on our own submission) which one was actually better is becoming almost impossible.
\nMy suggestion would be to increase the resolution from two to three or four digits.
", "messages": [ { "id": 3375936, "postDate": "2025-12-13T12:24:19.607Z", "content": "Is it possible to have a few more digits on the leaderboard? A lot of us are on similar numbers, and seeing on the leaderboard (and on our own submission) which one was actually better is becoming almost impossible.
\nMy suggestion would be to increase the resolution from two to three or four digits.
", "rawMarkdown": "Is it possible to have a few more digits on the leaderboard? A lot of us are on similar numbers, and seeing on the leaderboard (and on our own submission) which one was actually better is becoming almost impossible.\n\nMy suggestion would be to increase the resolution from two to three or four digits.", "votes": 8 }, { "id": 3376131, "postDate": "2025-12-13T17:41:09.950Z", "content": "Don't forget - you can display your submission by Public Score - pretty sure that handles all the digits that kaggle knows about your submission.
", "rawMarkdown": "Don't forget - you can display your submission by Public Score - pretty sure that handles all the digits that kaggle knows about your submission.", "votes": 3, "replies": [ { "id": 3376169, "postDate": "2025-12-13T19:46:56.130Z", "content": "This basically solves the use case I needed… Thanks again @pcjimmmy
", "rawMarkdown": "This basically solves the use case I needed... Thanks again @pcjimmmy ", "votes": 1 }, { "id": 3376269, "postDate": "2025-12-14T03:35:04.670Z", "content": "2 years, and this is the first time I saw that. Thanks alot @pcjimmmy
", "rawMarkdown": "2 years, and this is the first time I saw that. Thanks alot @pcjimmmy ", "votes": 1 } ] }, { "id": 3376036, "postDate": "2025-12-13T14:29:49.980Z", "content": "More digits can allow one to probing dataset's distribution. Consider train and public test contain about 2/3 of the whole dataset (and the way score is calculated), this kind of probing should be prohibited.
", "rawMarkdown": "More digits can allow one to probing dataset's distribution. Consider train and public test contain about 2/3 of the whole dataset (and the way score is calculated), this kind of probing should be prohibited.", "votes": 2 }, { "id": 3375939, "postDate": "2025-12-13T12:33:02.203Z", "content": "I am not really sure but I think that is the point of keeping the decimals to two only.
\nIf I have two submissions getting the same result (up to two decimals), I am forced to think and choose the better one myself. If I am simply picking the better one using the 3rd or 4th decimal, it may not be the best for the problem statement given such a small amount of data and I might have got lucky.
", "rawMarkdown": "I am not really sure but I think that is the point of keeping the decimals to two only.\n\nIf I have two submissions getting the same result (up to two decimals), I am forced to think and choose the better one myself. If I am simply picking the better one using the 3rd or 4th decimal, it may not be the best for the problem statement given such a small amount of data and I might have got lucky. ", "votes": 2, "replies": [ { "id": 3376050, "postDate": "2025-12-13T15:16:28.527Z", "content": "@komilparmar when the competition will be over next month, they may release the private leaderboard in 5-6 decimals. I have seen this happening in other competitions
", "rawMarkdown": "@komilparmar when the competition will be over next month, they may release the private leaderboard in 5-6 decimals. I have seen this happening in other competitions ", "votes": 1 } ] }, { "id": 3375953, "postDate": "2025-12-13T12:47:11.223Z", "content": "Do you think it will really generalize beyond two digits? I don't expect huge shakeup or anything, but I do expect variance in that second digit.
", "rawMarkdown": "Do you think it will really generalize beyond two digits? I don't expect huge shakeup or anything, but I do expect variance in that second digit.", "votes": -1 }, { "id": 3380594, "postDate": "2025-12-22T17:54:41.137Z", "content": "That would be really helpful here, gets confusing sometimes
", "rawMarkdown": "That would be really helpful here, gets confusing sometimes " }, { "id": 3376426, "postDate": "2025-12-14T12:56:03.080Z", "content": "More digits is helpful!
", "rawMarkdown": "More digits is helpful!", "replies": [ { "id": 3376510, "postDate": "2025-12-14T16:06:17.313Z", "content": "How does it help?
", "rawMarkdown": "How does it help?" } ] }, { "id": 3376081, "postDate": "2025-12-13T16:07:24.003Z", "rawMarkdown": "", "votes": -3, "isDeleted": true } ], "comments": [ { "id": 3376131, "author_name": "PC Jimmmy", "author_url": "", "post_date": "2025-12-13T17:41:09.950000", "content": "Don't forget - you can display your submission by Public Score - pretty sure that handles all the digits that kaggle knows about your submission.
", "votes": 3, "replies": [ { "id": 3376169, "author_name": "Tim Krige", "author_url": "", "post_date": "2025-12-13T19:46:56.130000", "content": "This basically solves the use case I needed… Thanks again @pcjimmmy
", "votes": 1, "replies": [] }, { "id": 3376269, "author_name": "Komil Parmar", "author_url": "", "post_date": "2025-12-14T03:35:04.670000", "content": "2 years, and this is the first time I saw that. Thanks alot @pcjimmmy
", "votes": 1, "replies": [] } ] }, { "id": 3376036, "author_name": "Sayaka Miki", "author_url": "", "post_date": "2025-12-13T14:29:49.980000", "content": "More digits can allow one to probing dataset's distribution. Consider train and public test contain about 2/3 of the whole dataset (and the way score is calculated), this kind of probing should be prohibited.
", "votes": 2, "replies": [] }, { "id": 3375939, "author_name": "Komil Parmar", "author_url": "", "post_date": "2025-12-13T12:33:02.203000", "content": "I am not really sure but I think that is the point of keeping the decimals to two only.
\nIf I have two submissions getting the same result (up to two decimals), I am forced to think and choose the better one myself. If I am simply picking the better one using the 3rd or 4th decimal, it may not be the best for the problem statement given such a small amount of data and I might have got lucky.
", "votes": 2, "replies": [ { "id": 3376050, "author_name": "Ravi Ramakrishnan", "author_url": "", "post_date": "2025-12-13T15:16:28.527000", "content": "@komilparmar when the competition will be over next month, they may release the private leaderboard in 5-6 decimals. I have seen this happening in other competitions
", "votes": 1, "replies": [] } ] }, { "id": 3375953, "author_name": "chemdatafarmer", "author_url": "", "post_date": "2025-12-13T12:47:11.223000", "content": "Do you think it will really generalize beyond two digits? I don't expect huge shakeup or anything, but I do expect variance in that second digit.
", "votes": -1, "replies": [] }, { "id": 3380594, "author_name": "Ahmed Walid", "author_url": "", "post_date": "2025-12-22T17:54:41.137000", "content": "That would be really helpful here, gets confusing sometimes
", "votes": 0, "replies": [] }, { "id": 3376426, "author_name": "ForcewithMe", "author_url": "", "post_date": "2025-12-14T12:56:03.080000", "content": "More digits is helpful!
", "votes": 0, "replies": [ { "id": 3376510, "author_name": "chemdatafarmer", "author_url": "", "post_date": "2025-12-14T16:06:17.313000", "content": "How does it help?
", "votes": 0, "replies": [] } ] }, { "id": 3376081, "author_name": "", "author_url": "", "post_date": "2025-12-13T16:07:24.003000", "content": "", "votes": -3, "replies": [] } ], "raw_markdown_by_id": { "3375936": "Is it possible to have a few more digits on the leaderboard? A lot of us are on similar numbers, and seeing on the leaderboard (and on our own submission) which one was actually better is becoming almost impossible.\n\nMy suggestion would be to increase the resolution from two to three or four digits.", "3376131": "Don't forget - you can display your submission by Public Score - pretty sure that handles all the digits that kaggle knows about your submission.", "3376036": "More digits can allow one to probing dataset's distribution. Consider train and public test contain about 2/3 of the whole dataset (and the way score is calculated), this kind of probing should be prohibited.", "3375939": "I am not really sure but I think that is the point of keeping the decimals to two only.\n\nIf I have two submissions getting the same result (up to two decimals), I am forced to think and choose the better one myself. If I am simply picking the better one using the 3rd or 4th decimal, it may not be the best for the problem statement given such a small amount of data and I might have got lucky. ", "3375953": "Do you think it will really generalize beyond two digits? I don't expect huge shakeup or anything, but I do expect variance in that second digit.", "3380594": "That would be really helpful here, gets confusing sometimes ", "3376426": "More digits is helpful!", "3376081": "" } }