{
  "id": 77958,
  "title": "Using compression to iterate faster. Is it against the rules?",
  "url": "/competitions/quora-insincere-questions-classification/discussion/77958",
  "author_name": "",
  "post_date": "2019-01-18T04:54:20.047754500Z",
  "votes": 2,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Hi guys,</p>\n\n<p>I am using <a href=\"https://www.kaggle.com/paulorzp/compressing-a-binary-submission-within-a-string\">Compressing Binary Predictions Kernel</a> to submit my local predictions. This trick allows me to iterate faster (Of course I will go back to normal kernels at some point). </p>\n\n<p>One thing that bothers me is a possibility that I may be banned for using this possibility, so the question is: how do you think is using this leak against the rules of competition?</p>\n\n<p><a href=\"/juliaelliott\">@juliaelliott</a> ?</p>\n\n<p>Thanks in advance,\nPavel</p>",
  "messages": [
    {
      "id": "457781",
      "postDate": "01/18/2019 04:54:20",
      "content": "<p>Hi guys,</p>\n\n<p>I am using <a href=\"https://www.kaggle.com/paulorzp/compressing-a-binary-submission-within-a-string\">Compressing Binary Predictions Kernel</a> to submit my local predictions. This trick allows me to iterate faster (Of course I will go back to normal kernels at some point). </p>\n\n<p>One thing that bothers me is a possibility that I may be banned for using this possibility, so the question is: how do you think is using this leak against the rules of competition?</p>\n\n<p><a href=\"/juliaelliott\">@juliaelliott</a> ?</p>\n\n<p>Thanks in advance,\nPavel</p>",
      "rawMarkdown": "Hi guys,\n\nI am using [Compressing Binary Predictions Kernel][1] to submit my local predictions. This trick allows me to iterate faster (Of course I will go back to normal kernels at some point). \n\nOne thing that bothers me is a possibility that I may be banned for using this possibility, so the question is: how do you think is using this leak against the rules of competition?\n\n@juliaelliott ?\n\n  [1]: https://www.kaggle.com/paulorzp/compressing-a-binary-submission-within-a-string \"Compressing Binary Predictions Kernel\"\n\nThanks in advance,\nPavel",
      "votes": null
    },
    {
      "id": "459824",
      "postDate": "01/22/2019 12:40:28",
      "content": "<p>I think the organizers will not answer this question. My opinion you should guess.</p>",
      "rawMarkdown": "I think the organizers will not answer this question. My opinion you should guess.",
      "votes": null
    },
    {
      "id": "459882",
      "postDate": "01/22/2019 15:02:00",
      "content": "<p>Hi Pavel,\nThis method can bring you up the ladder on the public set, but the final evaluation will be based on rerunning best kernels on extended data set so eventually those tricks don't have such big impact on final score. \nSo probably you won't be banned but your score also may fall down in the evaluation phase.\ngood luck,\nA.</p>",
      "rawMarkdown": "Hi Pavel,\nThis method can bring you up the ladder on the public set, but the final evaluation will be based on rerunning best kernels on extended data set so eventually those tricks don't have such big impact on final score. \nSo probably you won't be banned but your score also may fall down in the evaluation phase.\ngood luck,\nA.",
      "votes": null
    },
    {
      "id": "460089",
      "postDate": "01/23/2019 01:27:56",
      "content": "<p>It seems to me that you're not actually training or running a model. Rather you're encoding your submission csv and submitting it. I'm pretty sure this is against the rules as it clearly enables you to use beyond the existing submission time limit as well as external data.</p>",
      "rawMarkdown": "It seems to me that you're not actually training or running a model. Rather you're encoding your submission csv and submitting it. I'm pretty sure this is against the rules as it clearly enables you to use beyond the existing submission time limit as well as external data.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 459824,
      "author_name": "paulorzp",
      "author_url": "",
      "post_date": "01/22/2019 12:40:28",
      "content": "<p>I think the organizers will not answer this question. My opinion you should guess.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 459882,
      "author_name": "akuropatwinski",
      "author_url": "",
      "post_date": "01/22/2019 15:02:00",
      "content": "<p>Hi Pavel,\nThis method can bring you up the ladder on the public set, but the final evaluation will be based on rerunning best kernels on extended data set so eventually those tricks don't have such big impact on final score. \nSo probably you won't be banned but your score also may fall down in the evaluation phase.\ngood luck,\nA.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 460089,
      "author_name": "msf908",
      "author_url": "",
      "post_date": "01/23/2019 01:27:56",
      "content": "<p>It seems to me that you're not actually training or running a model. Rather you're encoding your submission csv and submitting it. I'm pretty sure this is against the rules as it clearly enables you to use beyond the existing submission time limit as well as external data.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "457781": "Hi guys,\n\nI am using [Compressing Binary Predictions Kernel][1] to submit my local predictions. This trick allows me to iterate faster (Of course I will go back to normal kernels at some point). \n\nOne thing that bothers me is a possibility that I may be banned for using this possibility, so the question is: how do you think is using this leak against the rules of competition?\n\n@juliaelliott ?\n\n  [1]: https://www.kaggle.com/paulorzp/compressing-a-binary-submission-within-a-string \"Compressing Binary Predictions Kernel\"\n\nThanks in advance,\nPavel",
    "459824": "I think the organizers will not answer this question. My opinion you should guess.",
    "459882": "Hi Pavel,\nThis method can bring you up the ladder on the public set, but the final evaluation will be based on rerunning best kernels on extended data set so eventually those tricks don't have such big impact on final score. \nSo probably you won't be banned but your score also may fall down in the evaluation phase.\ngood luck,\nA.",
    "460089": "It seems to me that you're not actually training or running a model. Rather you're encoding your submission csv and submitting it. I'm pretty sure this is against the rules as it clearly enables you to use beyond the existing submission time limit as well as external data."
  },
  "source": "meta"
}