{
  "id": 191018,
  "title": "Do we know how large private test groups are?",
  "url": "/competitions/riiid-test-answer-prediction/discussion/191018",
  "author_name": "",
  "post_date": "2020-10-14T07:42:56.718411Z",
  "votes": 6,
  "comment_count": 3,
  "views": 0,
  "content": "<p>We know that there are some 2.5 million predictions to be made on the private test set. In the public test set, we get groups of some 20 rows each, what is this sizing for the private test set? I presume there will be considerably larger groups?</p>",
  "messages": [
    {
      "id": "1049195",
      "postDate": "10/14/2020 07:42:56",
      "content": "<p>We know that there are some 2.5 million predictions to be made on the private test set. In the public test set, we get groups of some 20 rows each, what is this sizing for the private test set? I presume there will be considerably larger groups?</p>",
      "rawMarkdown": "We know that there are some 2.5 million predictions to be made on the private test set. In the public test set, we get groups of some 20 rows each, what is this sizing for the private test set? I presume there will be considerably larger groups?",
      "votes": null
    },
    {
      "id": "1049202",
      "postDate": "10/14/2020 07:47:46",
      "content": "<blockquote>\n  <p>The API provides user interactions groups in the order in which they occurred. Each group will contain interactions from many different users, but no more than one task_container_id of questions from any single user. Each group has between 1 and 1000 users.</p>\n</blockquote>\n<p>We cannot say that max we can have is 1k rows, as there's a possible of streak of questions here…</p>",
      "rawMarkdown": ">The API provides user interactions groups in the order in which they occurred. Each group will contain interactions from many different users, but no more than one task_container_id of questions from any single user. Each group has between 1 and 1000 users.\n\nWe cannot say that max we can have is 1k rows, as there's a possible of streak of questions here...",
      "votes": null
    },
    {
      "id": "1049215",
      "postDate": "10/14/2020 07:57:58",
      "content": "<p>Reading data documentation is an underrated skill :)<br>\nWell I guess the best one can do then is to optimize for both small and large sets… peanuts…</p>",
      "rawMarkdown": "Reading data documentation is an underrated skill :)\nWell I guess the best one can do then is to optimize for both small and large sets... peanuts...",
      "votes": null
    },
    {
      "id": "1049301",
      "postDate": "10/14/2020 10:04:37",
      "content": "<p>I am assuming max group size to be not more than ~ 4000.  Max 1 user_id and 1 task_container_id implies ~ 2-4 average num sessions per user per group</p>",
      "rawMarkdown": "I am assuming max group size to be not more than ~ 4000.  Max 1 user_id and 1 task_container_id implies ~ 2-4 average num sessions per user per group",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1049202,
      "author_name": "adityaecdrid",
      "author_url": "",
      "post_date": "10/14/2020 07:47:46",
      "content": "<blockquote>\n  <p>The API provides user interactions groups in the order in which they occurred. Each group will contain interactions from many different users, but no more than one task_container_id of questions from any single user. Each group has between 1 and 1000 users.</p>\n</blockquote>\n<p>We cannot say that max we can have is 1k rows, as there's a possible of streak of questions here…</p>",
      "votes": null,
      "replies": [
        {
          "id": 1049215,
          "author_name": "spacelx",
          "author_url": "",
          "post_date": "10/14/2020 07:57:58",
          "content": "<p>Reading data documentation is an underrated skill :)<br>\nWell I guess the best one can do then is to optimize for both small and large sets… peanuts…</p>",
          "votes": null,
          "replies": [
            {
              "id": 1049301,
              "author_name": "abhimanyud",
              "author_url": "",
              "post_date": "10/14/2020 10:04:37",
              "content": "<p>I am assuming max group size to be not more than ~ 4000.  Max 1 user_id and 1 task_container_id implies ~ 2-4 average num sessions per user per group</p>",
              "votes": null,
              "replies": []
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1049195": "We know that there are some 2.5 million predictions to be made on the private test set. In the public test set, we get groups of some 20 rows each, what is this sizing for the private test set? I presume there will be considerably larger groups?",
    "1049202": ">The API provides user interactions groups in the order in which they occurred. Each group will contain interactions from many different users, but no more than one task_container_id of questions from any single user. Each group has between 1 and 1000 users.\n\nWe cannot say that max we can have is 1k rows, as there's a possible of streak of questions here...",
    "1049215": "Reading data documentation is an underrated skill :)\nWell I guess the best one can do then is to optimize for both small and large sets... peanuts...",
    "1049301": "I am assuming max group size to be not more than ~ 4000.  Max 1 user_id and 1 task_container_id implies ~ 2-4 average num sessions per user per group"
  },
  "source": "meta"
}