{
  "id": 243497,
  "title": "How should we interpret the results?",
  "url": "/competitions/birdclef-2021/discussion/243497",
  "author_name": "",
  "post_date": "2021-06-02T20:20:09.434959100Z",
  "votes": 3,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I see that first place here is 0.69, whereas in Cornell Birds it was 0.68. Does that mean that the top models are roughly 69% accurate for both competitions? Can the two be viewed this way?</p>",
  "messages": [
    {
      "id": "1333526",
      "postDate": "06/02/2021 20:20:09",
      "content": "<p>I see that first place here is 0.69, whereas in Cornell Birds it was 0.68. Does that mean that the top models are roughly 69% accurate for both competitions? Can the two be viewed this way?</p>",
      "rawMarkdown": "I see that first place here is 0.69, whereas in Cornell Birds it was 0.68. Does that mean that the top models are roughly 69% accurate for both competitions? Can the two be viewed this way?",
      "votes": null
    },
    {
      "id": "1333636",
      "postDate": "06/02/2021 23:34:46",
      "content": "<p>Cornell had 264 birds, this competiton had 397 birds.</p>",
      "rawMarkdown": "Cornell had 264 birds, this competiton had 397 birds.",
      "votes": null
    },
    {
      "id": "1333942",
      "postDate": "06/03/2021 07:04:03",
      "content": "<p>The score depend heavily on the proportion of no call rows, as no calls are way easier to predict on average.  I think we made progress since last year because the private scores are similar, but private LB score for no calls is 0.48 this year vs 0.57 last year.</p>",
      "rawMarkdown": "The score depend heavily on the proportion of no call rows, as no calls are way easier to predict on average.  I think we made progress since last year because the private scores are similar, but private LB score for no calls is 0.48 this year vs 0.57 last year.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1333636,
      "author_name": "fffrrt",
      "author_url": "",
      "post_date": "06/02/2021 23:34:46",
      "content": "<p>Cornell had 264 birds, this competiton had 397 birds.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1333942,
      "author_name": "cpmpml",
      "author_url": "",
      "post_date": "06/03/2021 07:04:03",
      "content": "<p>The score depend heavily on the proportion of no call rows, as no calls are way easier to predict on average.  I think we made progress since last year because the private scores are similar, but private LB score for no calls is 0.48 this year vs 0.57 last year.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1333526": "I see that first place here is 0.69, whereas in Cornell Birds it was 0.68. Does that mean that the top models are roughly 69% accurate for both competitions? Can the two be viewed this way?",
    "1333636": "Cornell had 264 birds, this competiton had 397 birds.",
    "1333942": "The score depend heavily on the proportion of no call rows, as no calls are way easier to predict on average.  I think we made progress since last year because the private scores are similar, but private LB score for no calls is 0.48 this year vs 0.57 last year."
  },
  "source": "meta"
}