{
  "id": 275683,
  "title": "A lot of crap blender solutions just for increase personal score!",
  "url": "/competitions/rsna-miccai-brain-tumor-radiogenomic-classification/discussion/275683",
  "author_name": "",
  "post_date": "2021-10-01T08:24:18.033641400Z",
  "votes": 7,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hi everybody 🤘!</p>\n<p>I'm checking time to time, the best score, I guess that most of you, and I'm worry about a lot of proposals with blender scores, that just create a fake feeling of be up in the score table un the public leaderboard, but this is not the feeling of competitions and kaggle, and it worries me! I don't know about how is this allowed by <a href=\"https://www.kaggle.com/kaggleStuff\" target=\"_blank\">@kaggleStuff</a>, because honestly this is a huge problem in order to know how is the real classification, and also for sharing real valuable code, not crap code!</p>\n<p>I'm my humble opinion, Kaggle Stuff should control this kind of practices, because I tried to avoid to use them, but for one side, I see how my score if going down everyday, and most of the people are using this code for nothing, because in private score or even when the challenge finish they will go down. But for my day to day, is very difficult to know where I'm in the main score, I'm good? I'm not?</p>\n<p>I don't know if the feeling is shared by somebody else, or it's just me….Good of lucks kagglers!</p>",
  "messages": [
    {
      "id": "1530553",
      "postDate": "10/01/2021 08:24:18",
      "content": "<p>Hi everybody 🤘!</p>\n<p>I'm checking time to time, the best score, I guess that most of you, and I'm worry about a lot of proposals with blender scores, that just create a fake feeling of be up in the score table un the public leaderboard, but this is not the feeling of competitions and kaggle, and it worries me! I don't know about how is this allowed by <a href=\"https://www.kaggle.com/kaggleStuff\" target=\"_blank\">@kaggleStuff</a>, because honestly this is a huge problem in order to know how is the real classification, and also for sharing real valuable code, not crap code!</p>\n<p>I'm my humble opinion, Kaggle Stuff should control this kind of practices, because I tried to avoid to use them, but for one side, I see how my score if going down everyday, and most of the people are using this code for nothing, because in private score or even when the challenge finish they will go down. But for my day to day, is very difficult to know where I'm in the main score, I'm good? I'm not?</p>\n<p>I don't know if the feeling is shared by somebody else, or it's just me….Good of lucks kagglers!</p>",
      "rawMarkdown": "Hi everybody 🤘!\n\nI'm checking time to time, the best score, I guess that most of you, and I'm worry about a lot of proposals with blender scores, that just create a fake feeling of be up in the score table un the public leaderboard, but this is not the feeling of competitions and kaggle, and it worries me! I don't know about how is this allowed by @kaggleStuff, because honestly this is a huge problem in order to know how is the real classification, and also for sharing real valuable code, not crap code!\n\nI'm my humble opinion, Kaggle Stuff should control this kind of practices, because I tried to avoid to use them, but for one side, I see how my score if going down everyday, and most of the people are using this code for nothing, because in private score or even when the challenge finish they will go down. But for my day to day, is very difficult to know where I'm in the main score, I'm good? I'm not?\n\nI don't know if the feeling is shared by somebody else, or it's just me....Good of lucks kagglers!",
      "votes": null
    },
    {
      "id": "1530646",
      "postDate": "10/01/2021 10:20:26",
      "content": "<p>Completely agree with you. Personally I don't really care that much about the public LB position, but I often browse the code section to learn new ideas from the public notebooks. Now the code section is plagued with those useless blending notebooks. I guess most of us reached an impasse in this competition because the dataset is too small and the target is hard to predict based on the images alone, but this is really annoying…</p>",
      "rawMarkdown": "Completely agree with you. Personally I don't really care that much about the public LB position, but I often browse the code section to learn new ideas from the public notebooks. Now the code section is plagued with those useless blending notebooks. I guess most of us reached an impasse in this competition because the dataset is too small and the target is hard to predict based on the images alone, but this is really annoying...",
      "votes": null
    },
    {
      "id": "1531657",
      "postDate": "10/02/2021 08:22:43",
      "content": "<p>With only a few days to go the private lb will be the best indicator. I still believe most of the scores (non blender) are hugely because of the randomness from having a small sample size.</p>",
      "rawMarkdown": "With only a few days to go the private lb will be the best indicator. I still believe most of the scores (non blender) are hugely because of the randomness from having a small sample size.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1530646,
      "author_name": "nanguyen",
      "author_url": "",
      "post_date": "10/01/2021 10:20:26",
      "content": "<p>Completely agree with you. Personally I don't really care that much about the public LB position, but I often browse the code section to learn new ideas from the public notebooks. Now the code section is plagued with those useless blending notebooks. I guess most of us reached an impasse in this competition because the dataset is too small and the target is hard to predict based on the images alone, but this is really annoying…</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1531657,
      "author_name": "aryamansharma47",
      "author_url": "",
      "post_date": "10/02/2021 08:22:43",
      "content": "<p>With only a few days to go the private lb will be the best indicator. I still believe most of the scores (non blender) are hugely because of the randomness from having a small sample size.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1530553": "Hi everybody 🤘!\n\nI'm checking time to time, the best score, I guess that most of you, and I'm worry about a lot of proposals with blender scores, that just create a fake feeling of be up in the score table un the public leaderboard, but this is not the feeling of competitions and kaggle, and it worries me! I don't know about how is this allowed by @kaggleStuff, because honestly this is a huge problem in order to know how is the real classification, and also for sharing real valuable code, not crap code!\n\nI'm my humble opinion, Kaggle Stuff should control this kind of practices, because I tried to avoid to use them, but for one side, I see how my score if going down everyday, and most of the people are using this code for nothing, because in private score or even when the challenge finish they will go down. But for my day to day, is very difficult to know where I'm in the main score, I'm good? I'm not?\n\nI don't know if the feeling is shared by somebody else, or it's just me....Good of lucks kagglers!",
    "1530646": "Completely agree with you. Personally I don't really care that much about the public LB position, but I often browse the code section to learn new ideas from the public notebooks. Now the code section is plagued with those useless blending notebooks. I guess most of us reached an impasse in this competition because the dataset is too small and the target is hard to predict based on the images alone, but this is really annoying...",
    "1531657": "With only a few days to go the private lb will be the best indicator. I still believe most of the scores (non blender) are hugely because of the randomness from having a small sample size."
  },
  "source": "meta"
}