{
  "id": 131910,
  "title": "Use quality(q=.1) to fill tourney score",
  "url": "/competitions/google-cloud-ncaa-march-madness-2020-division-1-mens-tournament/discussion/131910",
  "author_name": "",
  "post_date": "2020-02-22T15:39:05.922828900Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>When we use tourney data, we get some na, because some team does not into the tourney before, and we need to fill that, I try fill with 0, but get bad score, and then, I try fill with quality(q=.1), and look better. What do you think?</p>",
  "messages": [
    {
      "id": "753718",
      "postDate": "02/22/2020 15:39:05",
      "content": "<p>When we use tourney data, we get some na, because some team does not into the tourney before, and we need to fill that, I try fill with 0, but get bad score, and then, I try fill with quality(q=.1), and look better. What do you think?</p>",
      "rawMarkdown": "When we use tourney data, we get some na, because some team does not into the tourney before, and we need to fill that, I try fill with 0, but get bad score, and then, I try fill with quality(q=.1), and look better. What do you think?",
      "votes": null
    },
    {
      "id": "753733",
      "postDate": "02/22/2020 15:52:20",
      "content": "<p>I'd be happy to offer feedback but I'm not clear on your problem. Can you post an example?</p>",
      "rawMarkdown": "I'd be happy to offer feedback but I'm not clear on your problem. Can you post an example?",
      "votes": null
    },
    {
      "id": "754236",
      "postDate": "02/23/2020 08:58:44",
      "content": "<p>Here it is, some team don't have tourney data, so we can't calculate tourney mean score for it, right? </p>\n\n<p>And we need to fill that, First I fill with 0, and logloss is very high, and I find use quality(q=.1) is a better choice.</p>\n\n<p>What do you think?</p>",
      "rawMarkdown": "Here it is, some team don't have tourney data, so we can't calculate tourney mean score for it, right? \n\n And we need to fill that, First I fill with 0, and logloss is very high, and I find use quality(q=.1) is a better choice.\n\nWhat do you think?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 753733,
      "author_name": "btesch",
      "author_url": "",
      "post_date": "02/22/2020 15:52:20",
      "content": "<p>I'd be happy to offer feedback but I'm not clear on your problem. Can you post an example?</p>",
      "votes": null,
      "replies": [
        {
          "id": 754236,
          "author_name": "holoong9291",
          "author_url": "",
          "post_date": "02/23/2020 08:58:44",
          "content": "<p>Here it is, some team don't have tourney data, so we can't calculate tourney mean score for it, right? </p>\n\n<p>And we need to fill that, First I fill with 0, and logloss is very high, and I find use quality(q=.1) is a better choice.</p>\n\n<p>What do you think?</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "753718": "When we use tourney data, we get some na, because some team does not into the tourney before, and we need to fill that, I try fill with 0, but get bad score, and then, I try fill with quality(q=.1), and look better. What do you think?",
    "753733": "I'd be happy to offer feedback but I'm not clear on your problem. Can you post an example?",
    "754236": "Here it is, some team don't have tourney data, so we can't calculate tourney mean score for it, right? \n\n And we need to fill that, First I fill with 0, and logloss is very high, and I find use quality(q=.1) is a better choice.\n\nWhat do you think?"
  },
  "source": "meta"
}