{
  "id": 117717,
  "title": "How do you explain the significantly low scores on stage 2?",
  "url": "/competitions/rsna-intracranial-hemorrhage-detection/discussion/117717",
  "author_name": "Hadar",
  "post_date": "2019-11-17T10:12:51.735000",
  "votes": 5,
  "comment_count": 0,
  "views": 0,
  "content": "<p>The scores on stage 2 test set are about 0.01 lower than those of stage 1. We've got a score that we could not achieve with any of our test sets, and from what I read in the discussions it is the same for other competitors. One could easily say \"it is the statistics\", but yet, statistics brings noise, while the order in the leaderboard was somehow preserved, even though not completely. Gaining a 0.01 improvement while working hard on 0.001 improvements requires some thought for me.</p>",
  "messages": [
    {
      "id": 674953,
      "postDate": "2019-11-17T10:12:51.737Z",
      "content": "<p>The scores on stage 2 test set are about 0.01 lower than those of stage 1. We've got a score that we could not achieve with any of our test sets, and from what I read in the discussions it is the same for other competitors. One could easily say \"it is the statistics\", but yet, statistics brings noise, while the order in the leaderboard was somehow preserved, even though not completely. Gaining a 0.01 improvement while working hard on 0.001 improvements requires some thought for me.</p>",
      "rawMarkdown": "The scores on stage 2 test set are about 0.01 lower than those of stage 1. We've got a score that we could not achieve with any of our test sets, and from what I read in the discussions it is the same for other competitors. One could easily say \"it is the statistics\", but yet, statistics brings noise, while the order in the leaderboard was somehow preserved, even though not completely. Gaining a 0.01 improvement while working hard on 0.001 improvements requires some thought for me.",
      "votes": 5
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "674953": "The scores on stage 2 test set are about 0.01 lower than those of stage 1. We've got a score that we could not achieve with any of our test sets, and from what I read in the discussions it is the same for other competitors. One could easily say \"it is the statistics\", but yet, statistics brings noise, while the order in the leaderboard was somehow preserved, even though not completely. Gaining a 0.01 improvement while working hard on 0.001 improvements requires some thought for me."
  }
}