{
  "id": 82117,
  "title": "Ridiculous way to get Bronze",
  "url": "/competitions/elo-merchant-category-recommendation/discussion/82117",
  "author_name": "Ale",
  "post_date": "2019-02-27T09:39:38.050000",
  "votes": 8,
  "comment_count": 9,
  "views": 0,
  "content": "<p>First of all I'd like to congratulate all winners! the competition was quite tough,  especially outliers part and big gap between cv and public lb</p>\n\n<p>After the end of competition I decided to check how perform published kernels and results I got is quite unexpected.</p>\n\n<p>If we take submissions from 2 kernels with top scores (which were posted more than <strong>3 weeks ago</strong>):</p>\n\n<ol>\n<li><a href=\"https://www.kaggle.com/ashishpatel26/lb-3-687-truncated-mean\">3.688 Public LB By Ashish Patel</a>   </li>\n<li><a href=\"https://www.kaggle.com/saurabh502/why-no-simple-blend-plb-3-687\">3.688 Public LB by Saurabh Kumar</a>   </li>\n</ol>\n\n<p>And simply take a mean of it we get relatively high private score: 3.612 </p>\n\n<p><img src=\"https://i.ibb.co/qdGVfYx/blend.png\" alt=\"enter image description here\"></p>\n\n<p>To me it is quite astonishing that You could get a bronze and top 7% score without any effort</p>",
  "messages": [
    {
      "id": 479621,
      "postDate": "2019-02-27T09:39:38.050Z",
      "content": "<p>First of all I'd like to congratulate all winners! the competition was quite tough,  especially outliers part and big gap between cv and public lb</p>\n\n<p>After the end of competition I decided to check how perform published kernels and results I got is quite unexpected.</p>\n\n<p>If we take submissions from 2 kernels with top scores (which were posted more than <strong>3 weeks ago</strong>):</p>\n\n<ol>\n<li><a href=\"https://www.kaggle.com/ashishpatel26/lb-3-687-truncated-mean\">3.688 Public LB By Ashish Patel</a>   </li>\n<li><a href=\"https://www.kaggle.com/saurabh502/why-no-simple-blend-plb-3-687\">3.688 Public LB by Saurabh Kumar</a>   </li>\n</ol>\n\n<p>And simply take a mean of it we get relatively high private score: 3.612 </p>\n\n<p><img src=\"https://i.ibb.co/qdGVfYx/blend.png\" alt=\"enter image description here\"></p>\n\n<p>To me it is quite astonishing that You could get a bronze and top 7% score without any effort</p>",
      "rawMarkdown": "First of all I'd like to congratulate all winners! the competition was quite tough,  especially outliers part and big gap between cv and public lb\n\nAfter the end of competition I decided to check how perform published kernels and results I got is quite unexpected.\n\nIf we take submissions from 2 kernels with top scores (which were posted more than **3 weeks ago**):\n\n1. [3.688 Public LB By Ashish Patel][1]   \n2. [3.688 Public LB by Saurabh Kumar][2]   \n\n\nAnd simply take a mean of it we get relatively high private score: 3.612 \n\n![enter image description here][3]\n\n\n  [1]: https://www.kaggle.com/ashishpatel26/lb-3-687-truncated-mean\n  [2]: https://www.kaggle.com/saurabh502/why-no-simple-blend-plb-3-687\n  [3]: https://i.ibb.co/qdGVfYx/blend.png\n\n\nTo me it is quite astonishing that You could get a bronze and top 7% score without any effort",
      "votes": 8
    },
    {
      "id": 479700,
      "postDate": "2019-02-27T10:53:56.800Z",
      "content": "<p>&gt;You could get a bronze and top 7% score without any effort</p>\n\n<p>Welcome to Kaggle :D</p>",
      "rawMarkdown": "&gt;You could get a bronze and top 7% score without any effort\n\nWelcome to Kaggle :D",
      "votes": 5,
      "replies": [
        {
          "id": 479785,
          "postDate": "2019-02-27T12:26:03.090Z",
          "content": "<p><code>You could get a bronze and top 7% score without any effort</code>\n<img src=\"https://i.kym-cdn.com/entries/icons/mobile/000/027/763/07B89120-B48D-45FB-AF1D-49AF6CD16790.jpg\"></p>",
          "rawMarkdown": "```You could get a bronze and top 7% score without any effort```\n<img src=\"https://i.kym-cdn.com/entries/icons/mobile/000/027/763/07B89120-B48D-45FB-AF1D-49AF6CD16790.jpg\">",
          "votes": 5
        }
      ]
    },
    {
      "id": 479639,
      "postDate": "2019-02-27T10:01:39.293Z",
      "content": "<p>what is the probability of knowing that it will work. and selecting only that for you final submission :P</p>",
      "rawMarkdown": "what is the probability of knowing that it will work. and selecting only that for you final submission :P",
      "votes": 5,
      "replies": [
        {
          "id": 480160,
          "postDate": "2019-02-27T21:07:14.750Z",
          "content": "<p>The 3.685 on public leaderboard wasn't too shabby though if someone had thought of trying it. If you only want to spend an hour on a competition, then choosing the top 2-3 public kernels in the last week and performing an average blend might not be a bad strategy, if it's medals you're after and not actually learning something.</p>",
          "rawMarkdown": "The 3.685 on public leaderboard wasn't too shabby though if someone had thought of trying it. If you only want to spend an hour on a competition, then choosing the top 2-3 public kernels in the last week and performing an average blend might not be a bad strategy, if it's medals you're after and not actually learning something.",
          "votes": 3
        },
        {
          "id": 480162,
          "postDate": "2019-02-27T21:14:44.300Z",
          "content": "<p>Very true, keep collecting the medals and waste your precious time in copying and pasting. </p>",
          "rawMarkdown": "Very true, keep collecting the medals and waste your precious time in copying and pasting. ",
          "votes": 3
        }
      ]
    },
    {
      "id": 479918,
      "postDate": "2019-02-27T15:13:06.803Z",
      "content": "<p>Thats why I prefer being a Kernel Expert than a competition expert.</p>",
      "rawMarkdown": "Thats why I prefer being a Kernel Expert than a competition expert."
    },
    {
      "id": 480472,
      "postDate": "2019-02-28T08:23:45.093Z",
      "content": "<p>But then, how do you get any better ?</p>",
      "rawMarkdown": "But then, how do you get any better ?"
    },
    {
      "id": 480364,
      "postDate": "2019-02-28T05:16:22.407Z",
      "content": "<p>Say whaaaat ?!?!?!?!</p>",
      "rawMarkdown": "Say whaaaat ?!?!?!?!"
    },
    {
      "id": 479641,
      "postDate": "2019-02-27T10:02:08.193Z",
      "rawMarkdown": "",
      "isDeleted": true
    }
  ],
  "comments": [
    {
      "id": 479700,
      "author_name": "",
      "author_url": "",
      "post_date": "2019-02-27T10:53:56.800000",
      "content": "<p>&gt;You could get a bronze and top 7% score without any effort</p>\n\n<p>Welcome to Kaggle :D</p>",
      "votes": 5,
      "replies": [
        {
          "id": 479785,
          "author_name": "Sanyam Bhutani",
          "author_url": "",
          "post_date": "2019-02-27T12:26:03.090000",
          "content": "<p><code>You could get a bronze and top 7% score without any effort</code>\n<img src=\"https://i.kym-cdn.com/entries/icons/mobile/000/027/763/07B89120-B48D-45FB-AF1D-49AF6CD16790.jpg\"></p>",
          "votes": 5,
          "replies": []
        }
      ]
    },
    {
      "id": 479639,
      "author_name": "Manoj",
      "author_url": "",
      "post_date": "2019-02-27T10:01:39.293000",
      "content": "<p>what is the probability of knowing that it will work. and selecting only that for you final submission :P</p>",
      "votes": 5,
      "replies": [
        {
          "id": 480160,
          "author_name": "Freddie Karlbom",
          "author_url": "",
          "post_date": "2019-02-27T21:07:14.750000",
          "content": "<p>The 3.685 on public leaderboard wasn't too shabby though if someone had thought of trying it. If you only want to spend an hour on a competition, then choosing the top 2-3 public kernels in the last week and performing an average blend might not be a bad strategy, if it's medals you're after and not actually learning something.</p>",
          "votes": 3,
          "replies": []
        },
        {
          "id": 480162,
          "author_name": "Manoj",
          "author_url": "",
          "post_date": "2019-02-27T21:14:44.300000",
          "content": "<p>Very true, keep collecting the medals and waste your precious time in copying and pasting. </p>",
          "votes": 3,
          "replies": []
        }
      ]
    },
    {
      "id": 479918,
      "author_name": "Ashish Gupta",
      "author_url": "",
      "post_date": "2019-02-27T15:13:06.803000",
      "content": "<p>Thats why I prefer being a Kernel Expert than a competition expert.</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 480472,
      "author_name": "Christian",
      "author_url": "",
      "post_date": "2019-02-28T08:23:45.093000",
      "content": "<p>But then, how do you get any better ?</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 480364,
      "author_name": "Pritesh Shrivastava",
      "author_url": "",
      "post_date": "2019-02-28T05:16:22.407000",
      "content": "<p>Say whaaaat ?!?!?!?!</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 479641,
      "author_name": "",
      "author_url": "",
      "post_date": "2019-02-27T10:02:08.193000",
      "content": "",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "479621": "First of all I'd like to congratulate all winners! the competition was quite tough,  especially outliers part and big gap between cv and public lb\n\nAfter the end of competition I decided to check how perform published kernels and results I got is quite unexpected.\n\nIf we take submissions from 2 kernels with top scores (which were posted more than **3 weeks ago**):\n\n1. [3.688 Public LB By Ashish Patel][1]   \n2. [3.688 Public LB by Saurabh Kumar][2]   \n\n\nAnd simply take a mean of it we get relatively high private score: 3.612 \n\n![enter image description here][3]\n\n\n  [1]: https://www.kaggle.com/ashishpatel26/lb-3-687-truncated-mean\n  [2]: https://www.kaggle.com/saurabh502/why-no-simple-blend-plb-3-687\n  [3]: https://i.ibb.co/qdGVfYx/blend.png\n\n\nTo me it is quite astonishing that You could get a bronze and top 7% score without any effort",
    "479700": "&gt;You could get a bronze and top 7% score without any effort\n\nWelcome to Kaggle :D",
    "479639": "what is the probability of knowing that it will work. and selecting only that for you final submission :P",
    "479918": "Thats why I prefer being a Kernel Expert than a competition expert.",
    "480472": "But then, how do you get any better ?",
    "480364": "Say whaaaat ?!?!?!?!",
    "479641": ""
  }
}