{
  "id": 314078,
  "title": "Rating and target",
  "url": "/competitions/football-match-probability-prediction/discussion/314078",
  "author_name": "",
  "post_date": "2022-03-20T20:23:32.118594400Z",
  "votes": 11,
  "comment_count": 2,
  "views": 0,
  "content": "<p>We received multiple questions about ratings and targets so we decided to open a discussion about it.</p>\n<ul>\n<li><strong>The targets</strong>: the target is calculated as follows. If the home team has scored more goals than the away team, the home team wins, the target is <code>home</code>.  On the other hand, if the away team has scored more goals, the target is <code>away</code>. Otherwise, the target is <code>draw</code>.  <strong>The missing information here is the final score</strong>. For instance, for 3-0, 3-1, or 3-2 the target is <code>home</code> but we could also say that 3-0 is a stronger victory for the home than 3-2 where a draw could have been more likely to occur. That is why we decided to add the final score in the train data with a new file <code>train_target_and_scores.csv</code> which also contains the initial target. Feel free to use it if you want.</li>\n<li><strong>The ratings</strong>: as mentioned in the data description, ratings are calculated by Octosport and correspond to the <strong>pre-match team's relative strength</strong>. It is provided for all 10 historical matches. But, <em>we do not provide the team's rating for the match you have to predict</em>. The reason being these ratings will already contain the winning probabilities information.</li>\n</ul>",
  "messages": [
    {
      "id": "1730019",
      "postDate": "03/20/2022 20:23:32",
      "content": "<p>We received multiple questions about ratings and targets so we decided to open a discussion about it.</p>\n<ul>\n<li><strong>The targets</strong>: the target is calculated as follows. If the home team has scored more goals than the away team, the home team wins, the target is <code>home</code>.  On the other hand, if the away team has scored more goals, the target is <code>away</code>. Otherwise, the target is <code>draw</code>.  <strong>The missing information here is the final score</strong>. For instance, for 3-0, 3-1, or 3-2 the target is <code>home</code> but we could also say that 3-0 is a stronger victory for the home than 3-2 where a draw could have been more likely to occur. That is why we decided to add the final score in the train data with a new file <code>train_target_and_scores.csv</code> which also contains the initial target. Feel free to use it if you want.</li>\n<li><strong>The ratings</strong>: as mentioned in the data description, ratings are calculated by Octosport and correspond to the <strong>pre-match team's relative strength</strong>. It is provided for all 10 historical matches. But, <em>we do not provide the team's rating for the match you have to predict</em>. The reason being these ratings will already contain the winning probabilities information.</li>\n</ul>",
      "rawMarkdown": "We received multiple questions about ratings and targets so we decided to open a discussion about it.\n\n- **The targets**: the target is calculated as follows. If the home team has scored more goals than the away team, the home team wins, the target is ```home```.  On the other hand, if the away team has scored more goals, the target is ```away```. Otherwise, the target is ```draw```.  **The missing information here is the final score**. For instance, for 3-0, 3-1, or 3-2 the target is ```home``` but we could also say that 3-0 is a stronger victory for the home than 3-2 where a draw could have been more likely to occur. That is why we decided to add the final score in the train data with a new file ```train_target_and_scores.csv``` which also contains the initial target. Feel free to use it if you want.\n- **The ratings**: as mentioned in the data description, ratings are calculated by Octosport and correspond to the **pre-match team's relative strength**. It is provided for all 10 historical matches. But, *we do not provide the team's rating for the match you have to predict*. The reason being these ratings will already contain the winning probabilities information.",
      "votes": null
    },
    {
      "id": "1744785",
      "postDate": "04/04/2022 10:09:03",
      "content": "<p>Is the calculation of ratings based on information in the dataset? If it is not, perhaps there is not point to try to extrapolate the ratings for the matches to be predicted.</p>",
      "rawMarkdown": "Is the calculation of ratings based on information in the dataset? If it is not, perhaps there is not point to try to extrapolate the ratings for the matches to be predicted.",
      "votes": null
    },
    {
      "id": "1744825",
      "postDate": "04/04/2022 10:58:06",
      "content": "<p>Not it is not. But you can expect to find the expecting rating of the two teams for the match to be predicted using their past ratings.</p>",
      "rawMarkdown": "Not it is not. But you can expect to find the expecting rating of the two teams for the match to be predicted using their past ratings.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1744785,
      "author_name": "szeyeung",
      "author_url": "",
      "post_date": "04/04/2022 10:09:03",
      "content": "<p>Is the calculation of ratings based on information in the dataset? If it is not, perhaps there is not point to try to extrapolate the ratings for the matches to be predicted.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1744825,
          "author_name": "octosportio",
          "author_url": "",
          "post_date": "04/04/2022 10:58:06",
          "content": "<p>Not it is not. But you can expect to find the expecting rating of the two teams for the match to be predicted using their past ratings.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1730019": "We received multiple questions about ratings and targets so we decided to open a discussion about it.\n\n- **The targets**: the target is calculated as follows. If the home team has scored more goals than the away team, the home team wins, the target is ```home```.  On the other hand, if the away team has scored more goals, the target is ```away```. Otherwise, the target is ```draw```.  **The missing information here is the final score**. For instance, for 3-0, 3-1, or 3-2 the target is ```home``` but we could also say that 3-0 is a stronger victory for the home than 3-2 where a draw could have been more likely to occur. That is why we decided to add the final score in the train data with a new file ```train_target_and_scores.csv``` which also contains the initial target. Feel free to use it if you want.\n- **The ratings**: as mentioned in the data description, ratings are calculated by Octosport and correspond to the **pre-match team's relative strength**. It is provided for all 10 historical matches. But, *we do not provide the team's rating for the match you have to predict*. The reason being these ratings will already contain the winning probabilities information.",
    "1744785": "Is the calculation of ratings based on information in the dataset? If it is not, perhaps there is not point to try to extrapolate the ratings for the matches to be predicted.",
    "1744825": "Not it is not. But you can expect to find the expecting rating of the two teams for the match to be predicted using their past ratings."
  },
  "source": "meta"
}