Sorry to post something irrelevant.
\n\nI did some simple calculation about Kaggle points. Based on the equation (assume solo player and no decay occurs yet), ranking of 242 of 242 will earn 100000/sqrt(1)*(242)^(-0.75)*log(1+log(242)) = 863 pts...
\n\nThis is similar to rank 320 out of 3000 in a big competition. Is this equation fair enough?
", "messages": [ { "id": 191329, "postDate": "2017-06-10T00:06:38.223Z", "content": "Sorry to post something irrelevant.
\n\nI did some simple calculation about Kaggle points. Based on the equation (assume solo player and no decay occurs yet), ranking of 242 of 242 will earn 100000/sqrt(1)*(242)^(-0.75)*log(1+log(242)) = 863 pts...
\n\nThis is similar to rank 320 out of 3000 in a big competition. Is this equation fair enough?
", "rawMarkdown": "Sorry to post something irrelevant.\n \nI did some simple calculation about [Kaggle points][1]. Based on the equation (assume solo player and no decay occurs yet), ranking of 242 of 242 will earn 100000/sqrt(1)*(242)^(-0.75)*log(1+log(242)) = 863 pts... \n\nThis is similar to rank 320 out of 3000 in a big competition. Is this equation fair enough?\n\n\n [1]: https://www.kaggle.com/progression", "votes": 7 }, { "id": 191575, "postDate": "2017-06-10T22:24:45.643Z", "content": "You should raise the issue in the product feedback forum: https://www.kaggle.com/product-feedback
", "rawMarkdown": "You should raise the issue in the product feedback forum: https://www.kaggle.com/product-feedback", "votes": 1, "replies": [ { "id": 191579, "postDate": "2017-06-10T22:44:52.857Z", "content": "@Radu, I just made a post there. Thanks for provide the link.
", "rawMarkdown": "@Radu, I just made a post there. Thanks for provide the link." } ] }, { "id": 191526, "postDate": "2017-06-10T17:25:55.577Z", "content": "The point system is terribly flawed. It encourages sub-mediocre participations.
\n\nThere are many obvious solutions, like awarding points only if you finish in the medals.
", "rawMarkdown": "The point system is terribly flawed. It encourages sub-mediocre participations. \n\nThere are many obvious solutions, like awarding points only if you finish in the medals.", "votes": 1, "replies": [ { "id": 191530, "postDate": "2017-06-10T18:00:44.160Z", "content": "Yes. I feel \"awarding points only if you finish in the medals\" can be a good solution.
", "rawMarkdown": "Yes. I feel \"awarding points only if you finish in the medals\" can be a good solution." }, { "id": 191674, "postDate": "2017-06-11T10:25:27.347Z", "content": "Well, Kaagle wrote a bit in detail about how they ended up with current formula and you can see a lot of discussion happened there too! While I am not in total support of the current point system since it's too biased towards small competitions, giving points to only medal winners would mean many people who worked hard and were in top 15% etc will not get any points in big/tough competitions. The medals are a recognition already which only top 10% get, so the others can get something else atleast.
\n\nIt's been very long since the points system was last designed. The community has grown much much bigger as well as the prize money and the competing teams. So, perhaps it's time to think of a new system, with input from the community this time!
", "rawMarkdown": "Well, Kaagle [wrote a bit in detail][1] about how they ended up with current formula and you can see a lot of discussion happened there too! While I am not in total support of the current point system since it's too biased towards small competitions, giving points to only medal winners would mean many people who worked hard and were in top 15% etc will not get any points in big/tough competitions. The medals are a recognition already which only top 10% get, so the others can get something else atleast.\n\nIt's been very long since the points system was last designed. The community has grown much much bigger as well as the prize money and the competing teams. So, perhaps it's time to think of a new system, with input from the community this time!\n\n [1]: http://blog.kaggle.com/2015/05/13/improved-kaggle-rankings/", "votes": 1 } ] }, { "id": 191355, "postDate": "2017-06-10T02:46:28.277Z", "content": "Wow, I too did the math just now! Looks like you are right!.
\n\n242/242 here gives 862.835 points but 320/3000 in a big competition gives 860.434 points!
", "rawMarkdown": "Wow, I too did the math just now! Looks like you are right!.\n\n242/242 here gives 862.835 points but 320/3000 in a big competition gives 860.434 points!\n", "votes": 1, "replies": [ { "id": 191362, "postDate": "2017-06-10T03:39:11.173Z", "content": "Yes... So I think I need to join every single competition to earn points, even rank the last :)
", "rawMarkdown": "Yes... So I think I need to join every single competition to earn points, even rank the last :)" }, { "id": 191366, "postDate": "2017-06-10T03:42:55.170Z", "content": "==
", "rawMarkdown": "==" }, { "id": 191624, "postDate": "2017-06-11T05:35:24.120Z", "content": "@Feater, are you asking me or @Avinash?
", "rawMarkdown": "@Feater, are you asking me or @Avinash?" } ] }, { "id": 191357, "postDate": "2017-06-10T02:50:52.257Z", "content": "HI MechCoder,
\n\nFinal rank (XXth place) is more important than rate of porition (TOP XX %).
\nSmall competiton is chance to get many points.
I agree!
", "rawMarkdown": "I agree! " } ] }, { "id": 191576, "postDate": "2017-06-10T22:30:37.023Z", "rawMarkdown": "", "votes": 1, "isDeleted": true, "replies": [ { "id": 191580, "postDate": "2017-06-10T22:46:05.420Z", "content": "Haha. Thanks, Charles. I am trying to learn some computer vision techniques here.
", "rawMarkdown": "Haha. Thanks, Charles. I am trying to learn some computer vision techniques here.", "votes": 1 } ] } ], "comments": [ { "id": 191575, "author_name": "Radu Stoicescu", "author_url": "", "post_date": "2017-06-10T22:24:45.643000", "content": "You should raise the issue in the product feedback forum: https://www.kaggle.com/product-feedback
", "votes": 1, "replies": [ { "id": 191579, "author_name": "Shujian Liu", "author_url": "", "post_date": "2017-06-10T22:44:52.857000", "content": "@Radu, I just made a post there. Thanks for provide the link.
", "votes": 0, "replies": [] } ] }, { "id": 191526, "author_name": "Radu Stoicescu", "author_url": "", "post_date": "2017-06-10T17:25:55.577000", "content": "The point system is terribly flawed. It encourages sub-mediocre participations.
\n\nThere are many obvious solutions, like awarding points only if you finish in the medals.
", "votes": 1, "replies": [ { "id": 191530, "author_name": "Shujian Liu", "author_url": "", "post_date": "2017-06-10T18:00:44.160000", "content": "Yes. I feel \"awarding points only if you finish in the medals\" can be a good solution.
", "votes": 0, "replies": [] }, { "id": 191674, "author_name": "Avinash Hindupur", "author_url": "", "post_date": "2017-06-11T10:25:27.347000", "content": "Well, Kaagle wrote a bit in detail about how they ended up with current formula and you can see a lot of discussion happened there too! While I am not in total support of the current point system since it's too biased towards small competitions, giving points to only medal winners would mean many people who worked hard and were in top 15% etc will not get any points in big/tough competitions. The medals are a recognition already which only top 10% get, so the others can get something else atleast.
\n\nIt's been very long since the points system was last designed. The community has grown much much bigger as well as the prize money and the competing teams. So, perhaps it's time to think of a new system, with input from the community this time!
", "votes": 1, "replies": [] } ] }, { "id": 191355, "author_name": "Avinash Hindupur", "author_url": "", "post_date": "2017-06-10T02:46:28.277000", "content": "Wow, I too did the math just now! Looks like you are right!.
\n\n242/242 here gives 862.835 points but 320/3000 in a big competition gives 860.434 points!
", "votes": 1, "replies": [ { "id": 191362, "author_name": "Shujian Liu", "author_url": "", "post_date": "2017-06-10T03:39:11.173000", "content": "Yes... So I think I need to join every single competition to earn points, even rank the last :)
", "votes": 0, "replies": [] }, { "id": 191366, "author_name": "sfx7rail", "author_url": "", "post_date": "2017-06-10T03:42:55.170000", "content": "==
", "votes": 0, "replies": [] }, { "id": 191624, "author_name": "Shujian Liu", "author_url": "", "post_date": "2017-06-11T05:35:24.120000", "content": "@Feater, are you asking me or @Avinash?
", "votes": 0, "replies": [] } ] }, { "id": 191357, "author_name": "toshi_k", "author_url": "", "post_date": "2017-06-10T02:50:52.257000", "content": "HI MechCoder,
\n\nFinal rank (XXth place) is more important than rate of porition (TOP XX %).
\nSmall competiton is chance to get many points.
I agree!
", "votes": 0, "replies": [] } ] }, { "id": 191576, "author_name": "", "author_url": "", "post_date": "2017-06-10T22:30:37.023000", "content": "", "votes": 1, "replies": [ { "id": 191580, "author_name": "Shujian Liu", "author_url": "", "post_date": "2017-06-10T22:46:05.420000", "content": "Haha. Thanks, Charles. I am trying to learn some computer vision techniques here.
", "votes": 1, "replies": [] } ] } ], "raw_markdown_by_id": { "191329": "Sorry to post something irrelevant.\n \nI did some simple calculation about [Kaggle points][1]. Based on the equation (assume solo player and no decay occurs yet), ranking of 242 of 242 will earn 100000/sqrt(1)*(242)^(-0.75)*log(1+log(242)) = 863 pts... \n\nThis is similar to rank 320 out of 3000 in a big competition. Is this equation fair enough?\n\n\n [1]: https://www.kaggle.com/progression", "191575": "You should raise the issue in the product feedback forum: https://www.kaggle.com/product-feedback", "191526": "The point system is terribly flawed. It encourages sub-mediocre participations. \n\nThere are many obvious solutions, like awarding points only if you finish in the medals.", "191355": "Wow, I too did the math just now! Looks like you are right!.\n\n242/242 here gives 862.835 points but 320/3000 in a big competition gives 860.434 points!\n", "191357": "HI MechCoder,\n\nFinal rank (XXth place) is more important than rate of porition (TOP XX %).