{
  "id": 114577,
  "title": "Submission",
  "url": "/competitions/pku-autonomous-driving/discussion/114577",
  "author_name": "",
  "post_date": "2019-10-27T16:27:30.711507Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Do your predictions (cars detected) need to be in a certain order for the public score to be accurate? Like does the closest car need to be first and the furthest last in the string? I don't know because in the train.csv the order seems to be random and sometimes it's ordered... <br>\nEDIT: 'Confidence scores are used to sort submission objects.' In Evaluation Overview</p>",
  "messages": [
    {
      "id": "659421",
      "postDate": "10/27/2019 16:27:30",
      "content": "<p>Do your predictions (cars detected) need to be in a certain order for the public score to be accurate? Like does the closest car need to be first and the furthest last in the string? I don't know because in the train.csv the order seems to be random and sometimes it's ordered... <br>\nEDIT: 'Confidence scores are used to sort submission objects.' In Evaluation Overview</p>",
      "rawMarkdown": "Do your predictions (cars detected) need to be in a certain order for the public score to be accurate? Like does the closest car need to be first and the furthest last in the string? I don't know because in the train.csv the order seems to be random and sometimes it's ordered...  \nEDIT: 'Confidence scores are used to sort submission objects.' In Evaluation Overview",
      "votes": null
    },
    {
      "id": "659536",
      "postDate": "10/27/2019 21:30:53",
      "content": "<p>Mix and check?\nI don't believe, that it matters</p>",
      "rawMarkdown": "Mix and check?\nI don't believe, that it matters",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 659536,
      "author_name": "ilu000",
      "author_url": "",
      "post_date": "10/27/2019 21:30:53",
      "content": "<p>Mix and check?\nI don't believe, that it matters</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "659421": "Do your predictions (cars detected) need to be in a certain order for the public score to be accurate? Like does the closest car need to be first and the furthest last in the string? I don't know because in the train.csv the order seems to be random and sometimes it's ordered...  \nEDIT: 'Confidence scores are used to sort submission objects.' In Evaluation Overview",
    "659536": "Mix and check?\nI don't believe, that it matters"
  },
  "source": "meta"
}