{
  "id": 284925,
  "title": "Expecting exactly 2015112 rows in test phase?",
  "url": "/competitions/g-research-crypto-forecasting/discussion/284925",
  "author_name": "Ultron",
  "post_date": "2021-11-02T22:20:14.051000",
  "votes": 8,
  "comment_count": 0,
  "views": 0,
  "content": "<p><code>supplemental_train.csv - After the submission period is over this file's data will be replaced with crypto asset prices from the submission period. The current copy, which is just filled approximately the right amount of data from train.csv is provided as a placeholder.</code></p>\n<p>As mentioned in the data description, the supplemental train is just a placeholder to contain the real-time data for the next three months once the final submission deadline is over. Breaking down the <code>2015112</code> number as follows:</p>\n<ul>\n<li>We have 14 Crypto assets, hence it's expected to see 2015112/14 ~ 143937 data points per asset in the test set.</li>\n<li>Further breaking down the <code>143937</code> number: 60 minutes, 24 hours, 30 days, and 3 months =&gt; 133920 minutes. Still, falling short by ~10K minutes i.e. 1-week time window? Just an extra-placeholder as a buffer or it has some relation with the re-run(test) phase? So, the test set is expected to contain data points of every minute in that 3-months phase. So how'll the re-run of submissions take place? </li>\n</ul>",
  "messages": [
    {
      "id": 1568603,
      "postDate": "2021-11-02T22:20:14.053Z",
      "content": "<p><code>supplemental_train.csv - After the submission period is over this file's data will be replaced with crypto asset prices from the submission period. The current copy, which is just filled approximately the right amount of data from train.csv is provided as a placeholder.</code></p>\n<p>As mentioned in the data description, the supplemental train is just a placeholder to contain the real-time data for the next three months once the final submission deadline is over. Breaking down the <code>2015112</code> number as follows:</p>\n<ul>\n<li>We have 14 Crypto assets, hence it's expected to see 2015112/14 ~ 143937 data points per asset in the test set.</li>\n<li>Further breaking down the <code>143937</code> number: 60 minutes, 24 hours, 30 days, and 3 months =&gt; 133920 minutes. Still, falling short by ~10K minutes i.e. 1-week time window? Just an extra-placeholder as a buffer or it has some relation with the re-run(test) phase? So, the test set is expected to contain data points of every minute in that 3-months phase. So how'll the re-run of submissions take place? </li>\n</ul>",
      "rawMarkdown": "`supplemental_train.csv - After the submission period is over this file's data will be replaced with crypto asset prices from the submission period. The current copy, which is just filled approximately the right amount of data from train.csv is provided as a placeholder.`\n\nAs mentioned in the data description, the supplemental train is just a placeholder to contain the real-time data for the next three months once the final submission deadline is over. Breaking down the `2015112` number as follows:\n- We have 14 Crypto assets, hence it's expected to see 2015112/14 ~ 143937 data points per asset in the test set.\n- Further breaking down the `143937` number: 60 minutes, 24 hours, 30 days, and 3 months => 133920 minutes. Still, falling short by ~10K minutes i.e. 1-week time window? Just an extra-placeholder as a buffer or it has some relation with the re-run(test) phase? So, the test set is expected to contain data points of every minute in that 3-months phase. So how'll the re-run of submissions take place? ",
      "votes": 8
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1568603": "`supplemental_train.csv - After the submission period is over this file's data will be replaced with crypto asset prices from the submission period. The current copy, which is just filled approximately the right amount of data from train.csv is provided as a placeholder.`\n\nAs mentioned in the data description, the supplemental train is just a placeholder to contain the real-time data for the next three months once the final submission deadline is over. Breaking down the `2015112` number as follows:\n- We have 14 Crypto assets, hence it's expected to see 2015112/14 ~ 143937 data points per asset in the test set.\n- Further breaking down the `143937` number: 60 minutes, 24 hours, 30 days, and 3 months => 133920 minutes. Still, falling short by ~10K minutes i.e. 1-week time window? Just an extra-placeholder as a buffer or it has some relation with the re-run(test) phase? So, the test set is expected to contain data points of every minute in that 3-months phase. So how'll the re-run of submissions take place? "
  }
}