{
  "id": 505717,
  "title": "For Kaggle - Competition Pre-Launch Window Needed?",
  "url": "/competitions/home-credit-credit-risk-model-stability/discussion/505717",
  "author_name": "BJBA99",
  "post_date": "2024-05-18T18:05:29.610000",
  "votes": 1,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Just a thought - whether a <strong>competition pre-launch window of 2-3 weeks</strong> would help to avoid situations where issues with either competition metric and/or data are noticed? During pre-launch window, competition metric and data quality would get assessed and discussed by the Kagglers. This may give sufficient time to the Host to reconsider the metric, data corrections if any rather than those being considered or not considered (even when needed) during the competition? I can think of at least 3 competitions where issues* were noticed in metric or data quality. (*by issue I don't necessarily mean wrong metric but even a situation of metric hack can dilute the real purpose of the competition and leave the participants disheartened.)</p>",
  "messages": [
    {
      "id": 2822714,
      "postDate": "2024-05-18T18:05:29.610Z",
      "content": "<p>Just a thought - whether a <strong>competition pre-launch window of 2-3 weeks</strong> would help to avoid situations where issues with either competition metric and/or data are noticed? During pre-launch window, competition metric and data quality would get assessed and discussed by the Kagglers. This may give sufficient time to the Host to reconsider the metric, data corrections if any rather than those being considered or not considered (even when needed) during the competition? I can think of at least 3 competitions where issues* were noticed in metric or data quality. (*by issue I don't necessarily mean wrong metric but even a situation of metric hack can dilute the real purpose of the competition and leave the participants disheartened.)</p>",
      "rawMarkdown": "Just a thought - whether a **competition pre-launch window of 2-3 weeks** would help to avoid situations where issues with either competition metric and/or data are noticed? During pre-launch window, competition metric and data quality would get assessed and discussed by the Kagglers. This may give sufficient time to the Host to reconsider the metric, data corrections if any rather than those being considered or not considered (even when needed) during the competition? I can think of at least 3 competitions where issues* were noticed in metric or data quality. (*by issue I don't necessarily mean wrong metric but even a situation of metric hack can dilute the real purpose of the competition and leave the participants disheartened.)",
      "votes": 1
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2822714": "Just a thought - whether a **competition pre-launch window of 2-3 weeks** would help to avoid situations where issues with either competition metric and/or data are noticed? During pre-launch window, competition metric and data quality would get assessed and discussed by the Kagglers. This may give sufficient time to the Host to reconsider the metric, data corrections if any rather than those being considered or not considered (even when needed) during the competition? I can think of at least 3 competitions where issues* were noticed in metric or data quality. (*by issue I don't necessarily mean wrong metric but even a situation of metric hack can dilute the real purpose of the competition and leave the participants disheartened.)"
  }
}