{
  "id": 52352,
  "title": "Original Test Data vs Halved Test Data",
  "url": "/competitions/talkingdata-adtracking-fraud-detection/discussion/52352",
  "author_name": "",
  "post_date": "2018-03-19T13:03:47.401286300Z",
  "votes": 4,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Hi we all know by now that a lot of test data has been nuked in the data directory compared with the original test data presumably to make the data more manageable.</p>\n\n<p>Can Kaggle confirm that solutions that use the original test  data will be regarded as invalid? ;)</p>",
  "messages": [
    {
      "id": "298367",
      "postDate": "03/19/2018 13:03:47",
      "content": "<p>Hi we all know by now that a lot of test data has been nuked in the data directory compared with the original test data presumably to make the data more manageable.</p>\n\n<p>Can Kaggle confirm that solutions that use the original test  data will be regarded as invalid? ;)</p>",
      "rawMarkdown": "Hi we all know by now that a lot of test data has been nuked in the data directory compared with the original test data presumably to make the data more manageable.\n\nCan Kaggle confirm that solutions that use the original test  data will be regarded as invalid? ;)",
      "votes": null
    },
    {
      "id": "298430",
      "postDate": "03/19/2018 15:01:36",
      "content": "<p>I think the best way is that Kaggle reposts the original test set and let us use that.</p>",
      "rawMarkdown": "I think the best way is that Kaggle reposts the original test set and let us use that.",
      "votes": null
    },
    {
      "id": "298444",
      "postDate": "03/19/2018 15:22:03",
      "content": "<p>I just want a level playing field - this should be made clear early on in the competition.  We still have a couple of months to go.</p>",
      "rawMarkdown": "I just want a level playing field - this should be made clear early on in the competition.  We still have a couple of months to go.",
      "votes": null
    },
    {
      "id": "298448",
      "postDate": "03/19/2018 15:31:59",
      "content": "<p>I totally agree, but Kaggle has known this issue for a long time(some other user posted on discussion board) and say nothing useful about this. Thus I guess they dont care if you use old test.</p>",
      "rawMarkdown": "I totally agree, but Kaggle has known this issue for a long time(some other user posted on discussion board) and say nothing useful about this. Thus I guess they dont care if you use old test.",
      "votes": null
    },
    {
      "id": "298469",
      "postDate": "03/19/2018 15:53:26",
      "content": "<p>The issue is that the whole set is too large to upload for evaluation.</p>\n\n<p>Explained here:\n<a href=\"https://www.kaggle.com/c/talkingdata-adtracking-fraud-detection/discussion/51877#295772\">https://www.kaggle.com/c/talkingdata-adtracking-fraud-detection/discussion/51877#295772</a></p>",
      "rawMarkdown": "The issue is that the whole set is too large to upload for evaluation.\n\nExplained here:\nhttps://www.kaggle.com/c/talkingdata-adtracking-fraud-detection/discussion/51877#295772",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 298430,
      "author_name": "wythhh",
      "author_url": "",
      "post_date": "03/19/2018 15:01:36",
      "content": "<p>I think the best way is that Kaggle reposts the original test set and let us use that.</p>",
      "votes": null,
      "replies": [
        {
          "id": 298469,
          "author_name": "cjwhiteuk",
          "author_url": "",
          "post_date": "03/19/2018 15:53:26",
          "content": "<p>The issue is that the whole set is too large to upload for evaluation.</p>\n\n<p>Explained here:\n<a href=\"https://www.kaggle.com/c/talkingdata-adtracking-fraud-detection/discussion/51877#295772\">https://www.kaggle.com/c/talkingdata-adtracking-fraud-detection/discussion/51877#295772</a></p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 298444,
      "author_name": "scirpus",
      "author_url": "",
      "post_date": "03/19/2018 15:22:03",
      "content": "<p>I just want a level playing field - this should be made clear early on in the competition.  We still have a couple of months to go.</p>",
      "votes": null,
      "replies": [
        {
          "id": 298448,
          "author_name": "wythhh",
          "author_url": "",
          "post_date": "03/19/2018 15:31:59",
          "content": "<p>I totally agree, but Kaggle has known this issue for a long time(some other user posted on discussion board) and say nothing useful about this. Thus I guess they dont care if you use old test.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "298367": "Hi we all know by now that a lot of test data has been nuked in the data directory compared with the original test data presumably to make the data more manageable.\n\nCan Kaggle confirm that solutions that use the original test  data will be regarded as invalid? ;)",
    "298430": "I think the best way is that Kaggle reposts the original test set and let us use that.",
    "298444": "I just want a level playing field - this should be made clear early on in the competition.  We still have a couple of months to go.",
    "298448": "I totally agree, but Kaggle has known this issue for a long time(some other user posted on discussion board) and say nothing useful about this. Thus I guess they dont care if you use old test.",
    "298469": "The issue is that the whole set is too large to upload for evaluation.\n\nExplained here:\nhttps://www.kaggle.com/c/talkingdata-adtracking-fraud-detection/discussion/51877#295772"
  },
  "source": "meta"
}