{
  "id": 290021,
  "title": "Error handling reminder",
  "url": "/competitions/g-research-crypto-forecasting/discussion/290021",
  "author_name": "Sohier Dane",
  "post_date": "2021-11-22T20:34:54.416000",
  "votes": 19,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Now that you've had a chance to get initial prototypes working, I want to encourage everyone to make sure your code will be robust enough to survive whatever edge cases the market generates. Consider the possibilities of:</p>\n<ul>\n<li>Extended periods of missing data for one cryptoasset.</li>\n<li>Extended periods of missing data for all cryptoassets.</li>\n<li>A coin fork or other surprise leading to an abrupt change in trading patterns.</li>\n</ul>\n<p>I'm sure there are many more events that could break pipelines that make too many assumptions. Please reply with any others you think your fellow competitors should be aware of!</p>",
  "messages": [
    {
      "id": 1592043,
      "postDate": "2021-11-22T20:34:54.417Z",
      "content": "<p>Now that you've had a chance to get initial prototypes working, I want to encourage everyone to make sure your code will be robust enough to survive whatever edge cases the market generates. Consider the possibilities of:</p>\n<ul>\n<li>Extended periods of missing data for one cryptoasset.</li>\n<li>Extended periods of missing data for all cryptoassets.</li>\n<li>A coin fork or other surprise leading to an abrupt change in trading patterns.</li>\n</ul>\n<p>I'm sure there are many more events that could break pipelines that make too many assumptions. Please reply with any others you think your fellow competitors should be aware of!</p>",
      "rawMarkdown": "Now that you've had a chance to get initial prototypes working, I want to encourage everyone to make sure your code will be robust enough to survive whatever edge cases the market generates. Consider the possibilities of:\n- Extended periods of missing data for one cryptoasset.\n- Extended periods of missing data for all cryptoassets.\n- A coin fork or other surprise leading to an abrupt change in trading patterns.\n\nI'm sure there are many more events that could break pipelines that make too many assumptions. Please reply with any others you think your fellow competitors should be aware of!",
      "votes": 18
    },
    {
      "id": 1593386,
      "postDate": "2021-11-23T20:56:49.490Z",
      "content": "<p>Thanks for the warning. Would you mind to clarify the target ? One month in the competition and it hasn't been made officially clear if we are dealing with a simple weighted mean of correlation or a weighted correlation.</p>",
      "rawMarkdown": "Thanks for the warning. Would you mind to clarify the target ? One month in the competition and it hasn't been made officially clear if we are dealing with a simple weighted mean of correlation or a weighted correlation.",
      "votes": 3,
      "replies": [
        {
          "id": 1595483,
          "postDate": "2021-11-25T18:07:37.077Z",
          "rawMarkdown": "",
          "votes": 3,
          "isDeleted": true
        },
        {
          "id": 1597949,
          "postDate": "2021-11-28T04:42:19.493Z",
          "content": "<p>Yeah it is very confusing i used pearson correlation between my predictions and the overlapped targets it showed a different score from the submission score</p>",
          "rawMarkdown": "Yeah it is very confusing i used pearson correlation between my predictions and the overlapped targets it showed a different score from the submission score",
          "votes": 2
        }
      ]
    },
    {
      "id": 1599048,
      "postDate": "2021-11-29T04:54:37.077Z",
      "content": "<p>Thanks for the warning Sohier</p>",
      "rawMarkdown": "Thanks for the warning Sohier",
      "votes": 1
    }
  ],
  "comments": [
    {
      "id": 1593386,
      "author_name": "Lucas Morin",
      "author_url": "",
      "post_date": "2021-11-23T20:56:49.490000",
      "content": "<p>Thanks for the warning. Would you mind to clarify the target ? One month in the competition and it hasn't been made officially clear if we are dealing with a simple weighted mean of correlation or a weighted correlation.</p>",
      "votes": 3,
      "replies": [
        {
          "id": 1595483,
          "author_name": "",
          "author_url": "",
          "post_date": "2021-11-25T18:07:37.077000",
          "content": "",
          "votes": 3,
          "replies": []
        },
        {
          "id": 1597949,
          "author_name": "Sayantan Mazumdar",
          "author_url": "",
          "post_date": "2021-11-28T04:42:19.493000",
          "content": "<p>Yeah it is very confusing i used pearson correlation between my predictions and the overlapped targets it showed a different score from the submission score</p>",
          "votes": 2,
          "replies": []
        }
      ]
    },
    {
      "id": 1599048,
      "author_name": "Chris Deotte",
      "author_url": "",
      "post_date": "2021-11-29T04:54:37.077000",
      "content": "<p>Thanks for the warning Sohier</p>",
      "votes": 1,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1592043": "Now that you've had a chance to get initial prototypes working, I want to encourage everyone to make sure your code will be robust enough to survive whatever edge cases the market generates. Consider the possibilities of:\n- Extended periods of missing data for one cryptoasset.\n- Extended periods of missing data for all cryptoassets.\n- A coin fork or other surprise leading to an abrupt change in trading patterns.\n\nI'm sure there are many more events that could break pipelines that make too many assumptions. Please reply with any others you think your fellow competitors should be aware of!",
    "1593386": "Thanks for the warning. Would you mind to clarify the target ? One month in the competition and it hasn't been made officially clear if we are dealing with a simple weighted mean of correlation or a weighted correlation.",
    "1599048": "Thanks for the warning Sohier"
  }
}