{
  "id": 544576,
  "title": "Who did best on RMS?",
  "url": "/competitions/ariel-data-challenge-2024/discussion/544576",
  "author_name": "Jeroen Cottaar",
  "post_date": "2024-11-05T18:34:33.758000",
  "votes": 3,
  "comment_count": 1,
  "views": 0,
  "content": "<p>So we've seen how our scores panned out overall, but a big part of this competition was sigma prediction. I'm curious how our solutions compare if purely ranked on RMS. If you want to compare this, please make a submission for your best model, but set sigma to a fixed value of 25e-6 for all observations. That way the score is proportional to pure RMS. When you post results, please post both your submission score and 25e-6 score on the private leaderboard.</p>\n<p>My score (private LB/private LB with sigma=25e-6): 0.745 / 0.672</p>",
  "messages": [
    {
      "id": 3037480,
      "postDate": "2024-11-05T18:34:33.760Z",
      "content": "<p>So we've seen how our scores panned out overall, but a big part of this competition was sigma prediction. I'm curious how our solutions compare if purely ranked on RMS. If you want to compare this, please make a submission for your best model, but set sigma to a fixed value of 25e-6 for all observations. That way the score is proportional to pure RMS. When you post results, please post both your submission score and 25e-6 score on the private leaderboard.</p>\n<p>My score (private LB/private LB with sigma=25e-6): 0.745 / 0.672</p>",
      "rawMarkdown": "So we've seen how our scores panned out overall, but a big part of this competition was sigma prediction. I'm curious how our solutions compare if purely ranked on RMS. If you want to compare this, please make a submission for your best model, but set sigma to a fixed value of 25e-6 for all observations. That way the score is proportional to pure RMS. When you post results, please post both your submission score and 25e-6 score on the private leaderboard.\n\nMy score (private LB/private LB with sigma=25e-6): 0.745 / 0.672",
      "votes": 3
    },
    {
      "id": 3038176,
      "postDate": "2024-11-06T17:26:22.193Z",
      "content": "<p>private LB / private LB with sigma=25e-6: 0.662 / 0.125</p>",
      "rawMarkdown": "private LB / private LB with sigma=25e-6: 0.662 / 0.125"
    }
  ],
  "comments": [
    {
      "id": 3038176,
      "author_name": "Oleh Kivernyk",
      "author_url": "",
      "post_date": "2024-11-06T17:26:22.193000",
      "content": "<p>private LB / private LB with sigma=25e-6: 0.662 / 0.125</p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3037480": "So we've seen how our scores panned out overall, but a big part of this competition was sigma prediction. I'm curious how our solutions compare if purely ranked on RMS. If you want to compare this, please make a submission for your best model, but set sigma to a fixed value of 25e-6 for all observations. That way the score is proportional to pure RMS. When you post results, please post both your submission score and 25e-6 score on the private leaderboard.\n\nMy score (private LB/private LB with sigma=25e-6): 0.745 / 0.672",
    "3038176": "private LB / private LB with sigma=25e-6: 0.662 / 0.125"
  }
}