{
  "id": 580239,
  "title": "Is label just BTC minute price percent change?",
  "url": "/competitions/drw-crypto-market-prediction/discussion/580239",
  "author_name": "",
  "post_date": "2025-05-23T09:41:36.262656400Z",
  "votes": 21,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Hi, I have applied cumulative sum operation on the target <code>label</code> (data sorted by timestamp) and the results resemble BTC normalized closed price between 2023-03-01 and 2024-03-01.</p>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F378049%2F460beff6009a5606760cf3dfdac78b30%2Fbtc_price.png?generation=1747992948070532&amp;alt=media\" alt=\"\"> </p>\n<p>How come it is not explicitly stated that our target is BTC price changes (or some crypto derivative)? </p>\n<p>My initial guess that some people will try to derive test dataset predictions from volumes and historical BTC prices.</p>\n<p>I have used following dataset to get historical BTC prices: <a href=\"https://www.kaggle.com/datasets/imranbukhari/comprehensive-btcusd-1m-data\" target=\"_blank\">https://www.kaggle.com/datasets/imranbukhari/comprehensive-btcusd-1m-data</a></p>\n<p>Chart was generated using this code: <a href=\"https://www.kaggle.com/code/uzdavinys/drw-crypto-comprehensive-eda\" target=\"_blank\">https://www.kaggle.com/code/uzdavinys/drw-crypto-comprehensive-eda</a></p>",
  "messages": [
    {
      "id": "3207830",
      "postDate": "05/23/2025 09:41:36",
      "content": "<p>Hi, I have applied cumulative sum operation on the target <code>label</code> (data sorted by timestamp) and the results resemble BTC normalized closed price between 2023-03-01 and 2024-03-01.</p>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F378049%2F460beff6009a5606760cf3dfdac78b30%2Fbtc_price.png?generation=1747992948070532&amp;alt=media\" alt=\"\"> </p>\n<p>How come it is not explicitly stated that our target is BTC price changes (or some crypto derivative)? </p>\n<p>My initial guess that some people will try to derive test dataset predictions from volumes and historical BTC prices.</p>\n<p>I have used following dataset to get historical BTC prices: <a href=\"https://www.kaggle.com/datasets/imranbukhari/comprehensive-btcusd-1m-data\" target=\"_blank\">https://www.kaggle.com/datasets/imranbukhari/comprehensive-btcusd-1m-data</a></p>\n<p>Chart was generated using this code: <a href=\"https://www.kaggle.com/code/uzdavinys/drw-crypto-comprehensive-eda\" target=\"_blank\">https://www.kaggle.com/code/uzdavinys/drw-crypto-comprehensive-eda</a></p>",
      "rawMarkdown": "Hi, I have applied cumulative sum operation on the target `label` (data sorted by timestamp) and the results resemble BTC normalized closed price between 2023-03-01 and 2024-03-01.\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F378049%2F460beff6009a5606760cf3dfdac78b30%2Fbtc_price.png?generation=1747992948070532&alt=media) \n\nHow come it is not explicitly stated that our target is BTC price changes (or some crypto derivative)? \n\nMy initial guess that some people will try to derive test dataset predictions from volumes and historical BTC prices.\n\nI have used following dataset to get historical BTC prices: https://www.kaggle.com/datasets/imranbukhari/comprehensive-btcusd-1m-data\n\nChart was generated using this code: https://www.kaggle.com/code/uzdavinys/drw-crypto-comprehensive-eda",
      "votes": null
    },
    {
      "id": "3207877",
      "postDate": "05/23/2025 11:04:45",
      "content": "<p>Sharp observation. But they will reshuffle the test dataset(as stated in the dataset intro), so I guess disclosing the underlying asset does not help much？</p>",
      "rawMarkdown": "Sharp observation. But they will reshuffle the test dataset(as stated in the dataset intro), so I guess disclosing the underlying asset does not help much？",
      "votes": null
    },
    {
      "id": "3207956",
      "postDate": "05/23/2025 13:22:16",
      "content": "<p>sharp observation to a sharp observation</p>",
      "rawMarkdown": "sharp observation to a sharp observation",
      "votes": null
    },
    {
      "id": "3210595",
      "postDate": "05/27/2025 12:30:32",
      "content": "<p>I agree with you.<br>\nComparing the values it seems to me that the label roughly represents (1 min price movement) / $1.9</p>",
      "rawMarkdown": "I agree with you.\nComparing the values it seems to me that the label roughly represents (1 min price movement) / $1.9",
      "votes": null
    },
    {
      "id": "3210725",
      "postDate": "05/27/2025 16:13:27",
      "content": "<p>Every crypto is very related to Bitcoin, moreover, the data set for testing will be changed</p>",
      "rawMarkdown": "Every crypto is very related to Bitcoin, moreover, the data set for testing will be changed",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3207877,
      "author_name": "tony271ynot",
      "author_url": "",
      "post_date": "05/23/2025 11:04:45",
      "content": "<p>Sharp observation. But they will reshuffle the test dataset(as stated in the dataset intro), so I guess disclosing the underlying asset does not help much？</p>",
      "votes": null,
      "replies": [
        {
          "id": 3207956,
          "author_name": "kshitijs1703",
          "author_url": "",
          "post_date": "05/23/2025 13:22:16",
          "content": "<p>sharp observation to a sharp observation</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 3210595,
      "author_name": "hyoko19",
      "author_url": "",
      "post_date": "05/27/2025 12:30:32",
      "content": "<p>I agree with you.<br>\nComparing the values it seems to me that the label roughly represents (1 min price movement) / $1.9</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 3210725,
      "author_name": "guillaumemichelot",
      "author_url": "",
      "post_date": "05/27/2025 16:13:27",
      "content": "<p>Every crypto is very related to Bitcoin, moreover, the data set for testing will be changed</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3207830": "Hi, I have applied cumulative sum operation on the target `label` (data sorted by timestamp) and the results resemble BTC normalized closed price between 2023-03-01 and 2024-03-01.\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F378049%2F460beff6009a5606760cf3dfdac78b30%2Fbtc_price.png?generation=1747992948070532&alt=media) \n\nHow come it is not explicitly stated that our target is BTC price changes (or some crypto derivative)? \n\nMy initial guess that some people will try to derive test dataset predictions from volumes and historical BTC prices.\n\nI have used following dataset to get historical BTC prices: https://www.kaggle.com/datasets/imranbukhari/comprehensive-btcusd-1m-data\n\nChart was generated using this code: https://www.kaggle.com/code/uzdavinys/drw-crypto-comprehensive-eda",
    "3207877": "Sharp observation. But they will reshuffle the test dataset(as stated in the dataset intro), so I guess disclosing the underlying asset does not help much？",
    "3207956": "sharp observation to a sharp observation",
    "3210595": "I agree with you.\nComparing the values it seems to me that the label roughly represents (1 min price movement) / $1.9",
    "3210725": "Every crypto is very related to Bitcoin, moreover, the data set for testing will be changed"
  },
  "source": "meta"
}