{
  "id": 47632,
  "title": "Just one word difference between 1st/2nd/3rd place?",
  "url": "/competitions/tensorflow-speech-recognition-challenge/discussion/47632",
  "author_name": "",
  "post_date": "2018-01-17T05:55:53.886367600Z",
  "votes": 6,
  "comment_count": 2,
  "views": 0,
  "content": "<p>It looks like minimum difference between LB scores among all ranks (not just 1st,2nd,3rd) are either .00011. I guess this is one word difference. </p>\n\n<p>This means that 1st/2nd/3rd places are just one word apart each other. \nThis was really close competition!</p>\n\n<p>And I'm curious how Kaggle handles tie cases. \nFirst one to submit the answer wins? If so, delaying a good submission until deadline (to deter team merge, etc) may not be good strategy in really close competition :)</p>",
  "messages": [
    {
      "id": "269695",
      "postDate": "01/17/2018 05:55:53",
      "content": "<p>It looks like minimum difference between LB scores among all ranks (not just 1st,2nd,3rd) are either .00011. I guess this is one word difference. </p>\n\n<p>This means that 1st/2nd/3rd places are just one word apart each other. \nThis was really close competition!</p>\n\n<p>And I'm curious how Kaggle handles tie cases. \nFirst one to submit the answer wins? If so, delaying a good submission until deadline (to deter team merge, etc) may not be good strategy in really close competition :)</p>",
      "rawMarkdown": "It looks like minimum difference between LB scores among all ranks (not just 1st,2nd,3rd) are either .00011. I guess this is one word difference. \n\nThis means that 1st/2nd/3rd places are just one word apart each other. \nThis was really close competition!\n \nAnd I'm curious how Kaggle handles tie cases. \nFirst one to submit the answer wins? If so, delaying a good submission until deadline (to deter team merge, etc) may not be good strategy in really close competition :)",
      "votes": null
    },
    {
      "id": "269698",
      "postDate": "01/17/2018 06:07:37",
      "content": "<p>Hi Sukjae - Yes, if scores are identical, the first one submitted gets the higher placing. </p>",
      "rawMarkdown": "Hi Sukjae - Yes, if scores are identical, the first one submitted gets the higher placing.",
      "votes": null
    },
    {
      "id": "269829",
      "postDate": "01/17/2018 11:17:52",
      "content": "<p>Hi looking at the tweet @OpenAI\nDeep linear networks can learn to perform nonlinear computation due to imprecision in floating point arithmetic\nThe word is imprecision.  Taking it into account and rounding the results to 0.001 there are 3 winners. </p>",
      "rawMarkdown": "Hi looking at the tweet @OpenAI\nDeep linear networks can learn to perform nonlinear computation due to imprecision in floating point arithmetic\nThe word is imprecision.  Taking it into account and rounding the results to 0.001 there are 3 winners.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 269698,
      "author_name": "inversion",
      "author_url": "",
      "post_date": "01/17/2018 06:07:37",
      "content": "<p>Hi Sukjae - Yes, if scores are identical, the first one submitted gets the higher placing. </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 269829,
      "author_name": "andre1233",
      "author_url": "",
      "post_date": "01/17/2018 11:17:52",
      "content": "<p>Hi looking at the tweet @OpenAI\nDeep linear networks can learn to perform nonlinear computation due to imprecision in floating point arithmetic\nThe word is imprecision.  Taking it into account and rounding the results to 0.001 there are 3 winners. </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "269695": "It looks like minimum difference between LB scores among all ranks (not just 1st,2nd,3rd) are either .00011. I guess this is one word difference. \n\nThis means that 1st/2nd/3rd places are just one word apart each other. \nThis was really close competition!\n \nAnd I'm curious how Kaggle handles tie cases. \nFirst one to submit the answer wins? If so, delaying a good submission until deadline (to deter team merge, etc) may not be good strategy in really close competition :)",
    "269698": "Hi Sukjae - Yes, if scores are identical, the first one submitted gets the higher placing.",
    "269829": "Hi looking at the tweet @OpenAI\nDeep linear networks can learn to perform nonlinear computation due to imprecision in floating point arithmetic\nThe word is imprecision.  Taking it into account and rounding the results to 0.001 there are 3 winners."
  },
  "source": "meta"
}