{
  "id": 130912,
  "title": "Don't be fooled by 0.0 log loss on LB",
  "url": "/competitions/google-cloud-ncaa-march-madness-2020-division-1-womens-tournament/discussion/130912",
  "author_name": "Shangqiu Li",
  "post_date": "2020-02-17T04:48:01.916000",
  "votes": 3,
  "comment_count": 2,
  "views": 0,
  "content": "<p>The reason why they got 0.0 log loss is because that their model overfitted. FYI, the data used to calculate stage 1 LB is in the past competition data. Don't get fooled by those 0.0 log loss LB. Trust your CV! The important part is stage 2 when we actually use the 2020 competition data to calculate.</p>",
  "messages": [
    {
      "id": 747990,
      "postDate": "2020-02-17T04:48:01.917Z",
      "content": "<p>The reason why they got 0.0 log loss is because that their model overfitted. FYI, the data used to calculate stage 1 LB is in the past competition data. Don't get fooled by those 0.0 log loss LB. Trust your CV! The important part is stage 2 when we actually use the 2020 competition data to calculate.</p>",
      "rawMarkdown": "The reason why they got 0.0 log loss is because that their model overfitted. FYI, the data used to calculate stage 1 LB is in the past competition data. Don't get fooled by those 0.0 log loss LB. Trust your CV! The important part is stage 2 when we actually use the 2020 competition data to calculate.",
      "votes": 3
    },
    {
      "id": 748595,
      "postDate": "2020-02-17T18:00:46.940Z",
      "rawMarkdown": "",
      "isDeleted": true,
      "replies": [
        {
          "id": 748597,
          "postDate": "2020-02-17T18:06:16.563Z",
          "content": "<p>I mean \"competition\" by past NCAA basketball competition not kaggle past competition.</p>",
          "rawMarkdown": "I mean \"competition\" by past NCAA basketball competition not kaggle past competition."
        }
      ]
    }
  ],
  "comments": [
    {
      "id": 748595,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-02-17T18:00:46.940000",
      "content": "",
      "votes": 0,
      "replies": [
        {
          "id": 748597,
          "author_name": "Shangqiu Li",
          "author_url": "",
          "post_date": "2020-02-17T18:06:16.563000",
          "content": "<p>I mean \"competition\" by past NCAA basketball competition not kaggle past competition.</p>",
          "votes": 0,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "747990": "The reason why they got 0.0 log loss is because that their model overfitted. FYI, the data used to calculate stage 1 LB is in the past competition data. Don't get fooled by those 0.0 log loss LB. Trust your CV! The important part is stage 2 when we actually use the 2020 competition data to calculate.",
    "748595": ""
  }
}