{
  "id": 130735,
  "title": "NCAA Brackets",
  "url": "/competitions/google-cloud-ncaa-march-madness-2020-division-1-mens-tournament/discussion/130735",
  "author_name": "",
  "post_date": "2020-02-16T02:45:23.038186500Z",
  "votes": 2,
  "comment_count": 12,
  "views": 0,
  "content": "<p>This is my first competition for NCAA </p>\n\n<p>For each stage 1 and 2 in this competition ,host want us a model to predict brackets right? For instance in Stage 1 (season 2015-2019) correct bracket already there because it is already done. They want us how correct our model predict the 2015-2019 (training a model)</p>\n\n<p>Stage 1 - You should submit predicted probabilities for every possible matchup in the past 5 NCAA® tournaments (seasons 2015-2019).</p>\n\n<p>Based on that trained model from stage 1 we need to find the bracket of 2020 Stage 2?</p>\n\n<p>Thanks</p>",
  "messages": [
    {
      "id": "747146",
      "postDate": "02/16/2020 02:45:23",
      "content": "<p>This is my first competition for NCAA </p>\n\n<p>For each stage 1 and 2 in this competition ,host want us a model to predict brackets right? For instance in Stage 1 (season 2015-2019) correct bracket already there because it is already done. They want us how correct our model predict the 2015-2019 (training a model)</p>\n\n<p>Stage 1 - You should submit predicted probabilities for every possible matchup in the past 5 NCAA® tournaments (seasons 2015-2019).</p>\n\n<p>Based on that trained model from stage 1 we need to find the bracket of 2020 Stage 2?</p>\n\n<p>Thanks</p>",
      "rawMarkdown": "This is my first competition for NCAA \n\nFor each stage 1 and 2 in this competition ,host want us a model to predict brackets right? For instance in Stage 1 (season 2015-2019) correct bracket already there because it is already done. They want us how correct our model predict the 2015-2019 (training a model)\n\nStage 1 - You should submit predicted probabilities for every possible matchup in the past 5 NCAA® tournaments (seasons 2015-2019).\n\nBased on that trained model from stage 1 we need to find the bracket of 2020 Stage 2?\n\nThanks",
      "votes": null
    },
    {
      "id": "747603",
      "postDate": "02/16/2020 16:27:30",
      "content": "<p>If I understand what you're asking, you don't need to predict what the bracket for stage two will be. \nThe bracket will be officially announced by the NCAA in about four weeks. You will use your model to predict the outcomes of every possible match up that could occur (~ 2000) even though only 63 games actually happen for the official bracket. </p>\n\n<p>You don't even need to participate in stage one if you don't want to. </p>",
      "rawMarkdown": "If I understand what you're asking, you don't need to predict what the bracket for stage two will be. \nThe bracket will be officially announced by the NCAA in about four weeks. You will use your model to predict the outcomes of every possible match up that could occur (~ 2000) even though only 63 games actually happen for the official bracket. \n\nYou don't even need to participate in stage one if you don't want to.",
      "votes": null
    },
    {
      "id": "747866",
      "postDate": "02/17/2020 00:25:22",
      "content": "<p><a href=\"/btesch\">@btesch</a> thanks , question here 2020 season started since 04th Nov 2019. Possibly there could be some regular season match already played. What my understanding is there are a total of 68 men’s in the NCAA tournament. On Selection Sunday 32 teams gain automatic entry into the tournament due to winning their conference’s championships. The remaining teams rely on a selection committee to be granted entry into the tournament. 32 teams which gain automatic entry into the tournament due to winning their \"Regular Season\" (not sure but my understanding).However , my point is some matchup will already be played before Selection day. \"Regular Season\" games are simply defined to be all games played on DayNum=132 or earlier (DayNum=132 is Selection Sunday). My question is when all these regular season already gonna be played and some has already played then what is the point to predict all matchups? It is like some of the data are already seen indirectly by the competitors for stage 2</p>\n\n<p><a href=\"/addisonhoward\">@addisonhoward</a> , <a href=\"/jeffsonas\">@jeffsonas</a> can you please also shed some light to clear my confusion? </p>",
      "rawMarkdown": "btesch thanks , question here 2020 season started since 04th Nov 2019. Possibly there could be some regular season match already played. What my understanding is there are a total of 68 men’s in the NCAA tournament. On Selection Sunday 32 teams gain automatic entry into the tournament due to winning their conference’s championships. The remaining teams rely on a selection committee to be granted entry into the tournament. 32 teams which gain automatic entry into the tournament due to winning their \"Regular Season\" (not sure but my understanding).However , my point is some matchup will already be played before Selection day. \"Regular Season\" games are simply defined to be all games played on DayNum=132 or earlier (DayNum=132 is Selection Sunday). My question is when all these regular season already gonna be played and some has already played then what is the point to predict all matchups? It is like some of the data are already seen indirectly by the competitors for stage 2\n\n@addisonhoward , @jeffsonas can you please also shed some light to clear my confusion?",
      "votes": null
    },
    {
      "id": "747870",
      "postDate": "02/17/2020 00:42:48",
      "content": "<p>None of the games you need to predict are played until day 136. Technically there are two tournament games that happen before day 136 (I think day 134) but we don't have to predict those - or if we do, they aren't actually counted. </p>",
      "rawMarkdown": "None of the games you need to predict are played until day 136. Technically there are two tournament games that happen before day 136 (I think day 134) but we don't have to predict those - or if we do, they aren't actually counted.",
      "votes": null
    },
    {
      "id": "747900",
      "postDate": "02/17/2020 01:55:36",
      "content": "<p><a href=\"/btesch\">@btesch</a> i think day 134 or after. In both stages, the sample submission will tell you which games to predict right? Though bracket has to be announced on selection day by committee, what if given sample pair of teams is not going to be aligned with selection Sunday brackets?</p>",
      "rawMarkdown": "btesch i think day 134 or after. In both stages, the sample submission will tell you which games to predict right? Though bracket has to be announced on selection day by committee, what if given sample pair of teams is not going to be aligned with selection Sunday brackets?",
      "votes": null
    },
    {
      "id": "747901",
      "postDate": "02/17/2020 02:02:46",
      "content": "<p>On Selection Sunday (day 132), the 68-team bracket will be announced.  You will need to predict winning percentages for all possible pairs of teams from those 68 teams.  The four \"play-in\" games played on Tuesday and Wednesday (day 134/5) will indeed be four of these possible pairs of teams, but it doesn't matter whether you predicted these games well; they are not included in the scoring.  All games played on day 136 or later (a total of 63 games) will count toward your score.  Prior to the submission deadline you will need to submit and select up to two complete submission files (which will include 68*67/2 data rows).  You can submit as many submission files as you like, but can only pick up to two of them that will count toward the contest.  And of course you have to pick them before the submission deadline as well.</p>",
      "rawMarkdown": "On Selection Sunday (day 132), the 68-team bracket will be announced.  You will need to predict winning percentages for all possible pairs of teams from those 68 teams.  The four \"play-in\" games played on Tuesday and Wednesday (day 134/5) will indeed be four of these possible pairs of teams, but it doesn't matter whether you predicted these games well; they are not included in the scoring.  All games played on day 136 or later (a total of 63 games) will count toward your score.  Prior to the submission deadline you will need to submit and select up to two complete submission files (which will include 68*67/2 data rows).  You can submit as many submission files as you like, but can only pick up to two of them that will count toward the contest.  And of course you have to pick them before the submission deadline as well.",
      "votes": null
    },
    {
      "id": "747902",
      "postDate": "02/17/2020 02:04:33",
      "content": "<p>Once we see who the 68 teams are, we will include (as part of stage 2) a sample submission file which lists all possible pairs of those 68 teams and predicts a 50% winning percentage for each of them.  This way you can verify that you have the right 68 teams in your own submission files. </p>",
      "rawMarkdown": "Once we see who the 68 teams are, we will include (as part of stage 2) a sample submission file which lists all possible pairs of those 68 teams and predicts a 50% winning percentage for each of them.  This way you can verify that you have the right 68 teams in your own submission files.",
      "votes": null
    },
    {
      "id": "747905",
      "postDate": "02/17/2020 02:12:23",
      "content": "<p><a href=\"/jeffsonas\">@jeffsonas</a> thanks for replying , after selection day (68 teams announced) we will have only probably 4 to 5 days more to go for predicting from these announced team (bracket). </p>\n\n<p>In that case why in data section it is mentioned </p>\n\n<p>\"Refer to the Timeline page for specific dates.** In both stages, the sample submission will tell you which games to predict**</p>",
      "rawMarkdown": "jeffsonas thanks for replying , after selection day (68 teams announced) we will have only probably 4 to 5 days more to go for predicting from these announced team (bracket). \n\nIn that case why in data section it is mentioned \n\n\"Refer to the Timeline page for specific dates.** In both stages, the sample submission will tell you which games to predict**",
      "votes": null
    },
    {
      "id": "747919",
      "postDate": "02/17/2020 02:35:54",
      "content": "<p><a href=\"/jeffsonas\">@jeffsonas</a> gotcha , it means we have to filter out those team using Selection Sunday team list and predict their outcome before 19 March 3PM UTC right?</p>",
      "rawMarkdown": "jeffsonas gotcha , it means we have to filter out those team using Selection Sunday team list and predict their outcome before 19 March 3PM UTC right?",
      "votes": null
    },
    {
      "id": "747921",
      "postDate": "02/17/2020 02:41:23",
      "content": "<p>There is a file named MNCAATourneySeeds that lists the 68 teams (and their seeds) for each season.  You can either use this file to figure out the 68*67/2 combinations or you can just look at the sample submission file; either way it will tell you which matchups to predict.  The \"/2\" part of \"68*67/2\" is because we only predict data rows where the ID of the first team listed is less than the ID of the second team listed.  So if team 1101 is given a 60% chance to defeat team 1102 in the 2020 season tournament, then your data row will look like \"2020_1101_1102,0.60\" and it is not necessary to also provide a prediction like \"2020_1102_1101,0.40\" because that second prediction can be inferred from the first prediction.</p>",
      "rawMarkdown": "There is a file named MNCAATourneySeeds that lists the 68 teams (and their seeds) for each season.  You can either use this file to figure out the 68*67/2 combinations or you can just look at the sample submission file; either way it will tell you which matchups to predict.  The \"/2\" part of \"68*67/2\" is because we only predict data rows where the ID of the first team listed is less than the ID of the second team listed.  So if team 1101 is given a 60% chance to defeat team 1102 in the 2020 season tournament, then your data row will look like \"2020_1101_1102,0.60\" and it is not necessary to also provide a prediction like \"2020_1102_1101,0.40\" because that second prediction can be inferred from the first prediction.",
      "votes": null
    },
    {
      "id": "747922",
      "postDate": "02/17/2020 02:42:16",
      "content": "<p>But yes, it is only teams picked on the Selection Sunday team list that you need to make predictions for, and you will need to make predictions for all possible pairs of those 68 teams.</p>",
      "rawMarkdown": "But yes, it is only teams picked on the Selection Sunday team list that you need to make predictions for, and you will need to make predictions for all possible pairs of those 68 teams.",
      "votes": null
    },
    {
      "id": "747944",
      "postDate": "02/17/2020 03:10:51",
      "content": "<p><a href=\"/jeffsonas\">@jeffsonas</a> thanks a lot , what if sample submission pairs is not aligned with selection Sunday?</p>",
      "rawMarkdown": "jeffsonas thanks a lot , what if sample submission pairs is not aligned with selection Sunday?",
      "votes": null
    },
    {
      "id": "747958",
      "postDate": "02/17/2020 03:30:39",
      "content": "<p>I think there is error-checking in place when you try to submit a prediction file, and it gets rejected if the values in the first column do not match what is expected (such as if you picked the wrong set of 68 teams, or too many or too few).  I am not sure if they have to be sorted by the first column, but I don't think they do.</p>",
      "rawMarkdown": "I think there is error-checking in place when you try to submit a prediction file, and it gets rejected if the values in the first column do not match what is expected (such as if you picked the wrong set of 68 teams, or too many or too few).  I am not sure if they have to be sorted by the first column, but I don't think they do.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 747603,
      "author_name": "btesch",
      "author_url": "",
      "post_date": "02/16/2020 16:27:30",
      "content": "<p>If I understand what you're asking, you don't need to predict what the bracket for stage two will be. \nThe bracket will be officially announced by the NCAA in about four weeks. You will use your model to predict the outcomes of every possible match up that could occur (~ 2000) even though only 63 games actually happen for the official bracket. </p>\n\n<p>You don't even need to participate in stage one if you don't want to. </p>",
      "votes": null,
      "replies": [
        {
          "id": 747866,
          "author_name": "oraware",
          "author_url": "",
          "post_date": "02/17/2020 00:25:22",
          "content": "<p><a href=\"/btesch\">@btesch</a> thanks , question here 2020 season started since 04th Nov 2019. Possibly there could be some regular season match already played. What my understanding is there are a total of 68 men’s in the NCAA tournament. On Selection Sunday 32 teams gain automatic entry into the tournament due to winning their conference’s championships. The remaining teams rely on a selection committee to be granted entry into the tournament. 32 teams which gain automatic entry into the tournament due to winning their \"Regular Season\" (not sure but my understanding).However , my point is some matchup will already be played before Selection day. \"Regular Season\" games are simply defined to be all games played on DayNum=132 or earlier (DayNum=132 is Selection Sunday). My question is when all these regular season already gonna be played and some has already played then what is the point to predict all matchups? It is like some of the data are already seen indirectly by the competitors for stage 2</p>\n\n<p><a href=\"/addisonhoward\">@addisonhoward</a> , <a href=\"/jeffsonas\">@jeffsonas</a> can you please also shed some light to clear my confusion? </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 747870,
          "author_name": "btesch",
          "author_url": "",
          "post_date": "02/17/2020 00:42:48",
          "content": "<p>None of the games you need to predict are played until day 136. Technically there are two tournament games that happen before day 136 (I think day 134) but we don't have to predict those - or if we do, they aren't actually counted. </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 747900,
      "author_name": "oraware",
      "author_url": "",
      "post_date": "02/17/2020 01:55:36",
      "content": "<p><a href=\"/btesch\">@btesch</a> i think day 134 or after. In both stages, the sample submission will tell you which games to predict right? Though bracket has to be announced on selection day by committee, what if given sample pair of teams is not going to be aligned with selection Sunday brackets?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 747901,
      "author_name": "jeffsonas",
      "author_url": "",
      "post_date": "02/17/2020 02:02:46",
      "content": "<p>On Selection Sunday (day 132), the 68-team bracket will be announced.  You will need to predict winning percentages for all possible pairs of teams from those 68 teams.  The four \"play-in\" games played on Tuesday and Wednesday (day 134/5) will indeed be four of these possible pairs of teams, but it doesn't matter whether you predicted these games well; they are not included in the scoring.  All games played on day 136 or later (a total of 63 games) will count toward your score.  Prior to the submission deadline you will need to submit and select up to two complete submission files (which will include 68*67/2 data rows).  You can submit as many submission files as you like, but can only pick up to two of them that will count toward the contest.  And of course you have to pick them before the submission deadline as well.</p>",
      "votes": null,
      "replies": [
        {
          "id": 747905,
          "author_name": "oraware",
          "author_url": "",
          "post_date": "02/17/2020 02:12:23",
          "content": "<p><a href=\"/jeffsonas\">@jeffsonas</a> thanks for replying , after selection day (68 teams announced) we will have only probably 4 to 5 days more to go for predicting from these announced team (bracket). </p>\n\n<p>In that case why in data section it is mentioned </p>\n\n<p>\"Refer to the Timeline page for specific dates.** In both stages, the sample submission will tell you which games to predict**</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 747902,
      "author_name": "jeffsonas",
      "author_url": "",
      "post_date": "02/17/2020 02:04:33",
      "content": "<p>Once we see who the 68 teams are, we will include (as part of stage 2) a sample submission file which lists all possible pairs of those 68 teams and predicts a 50% winning percentage for each of them.  This way you can verify that you have the right 68 teams in your own submission files. </p>",
      "votes": null,
      "replies": [
        {
          "id": 747919,
          "author_name": "oraware",
          "author_url": "",
          "post_date": "02/17/2020 02:35:54",
          "content": "<p><a href=\"/jeffsonas\">@jeffsonas</a> gotcha , it means we have to filter out those team using Selection Sunday team list and predict their outcome before 19 March 3PM UTC right?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 747921,
          "author_name": "jeffsonas",
          "author_url": "",
          "post_date": "02/17/2020 02:41:23",
          "content": "<p>There is a file named MNCAATourneySeeds that lists the 68 teams (and their seeds) for each season.  You can either use this file to figure out the 68*67/2 combinations or you can just look at the sample submission file; either way it will tell you which matchups to predict.  The \"/2\" part of \"68*67/2\" is because we only predict data rows where the ID of the first team listed is less than the ID of the second team listed.  So if team 1101 is given a 60% chance to defeat team 1102 in the 2020 season tournament, then your data row will look like \"2020_1101_1102,0.60\" and it is not necessary to also provide a prediction like \"2020_1102_1101,0.40\" because that second prediction can be inferred from the first prediction.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 747922,
          "author_name": "jeffsonas",
          "author_url": "",
          "post_date": "02/17/2020 02:42:16",
          "content": "<p>But yes, it is only teams picked on the Selection Sunday team list that you need to make predictions for, and you will need to make predictions for all possible pairs of those 68 teams.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 747944,
      "author_name": "oraware",
      "author_url": "",
      "post_date": "02/17/2020 03:10:51",
      "content": "<p><a href=\"/jeffsonas\">@jeffsonas</a> thanks a lot , what if sample submission pairs is not aligned with selection Sunday?</p>",
      "votes": null,
      "replies": [
        {
          "id": 747958,
          "author_name": "jeffsonas",
          "author_url": "",
          "post_date": "02/17/2020 03:30:39",
          "content": "<p>I think there is error-checking in place when you try to submit a prediction file, and it gets rejected if the values in the first column do not match what is expected (such as if you picked the wrong set of 68 teams, or too many or too few).  I am not sure if they have to be sorted by the first column, but I don't think they do.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "747146": "This is my first competition for NCAA \n\nFor each stage 1 and 2 in this competition ,host want us a model to predict brackets right? For instance in Stage 1 (season 2015-2019) correct bracket already there because it is already done. They want us how correct our model predict the 2015-2019 (training a model)\n\nStage 1 - You should submit predicted probabilities for every possible matchup in the past 5 NCAA® tournaments (seasons 2015-2019).\n\nBased on that trained model from stage 1 we need to find the bracket of 2020 Stage 2?\n\nThanks",
    "747603": "If I understand what you're asking, you don't need to predict what the bracket for stage two will be. \nThe bracket will be officially announced by the NCAA in about four weeks. You will use your model to predict the outcomes of every possible match up that could occur (~ 2000) even though only 63 games actually happen for the official bracket. \n\nYou don't even need to participate in stage one if you don't want to.",
    "747866": "btesch thanks , question here 2020 season started since 04th Nov 2019. Possibly there could be some regular season match already played. What my understanding is there are a total of 68 men’s in the NCAA tournament. On Selection Sunday 32 teams gain automatic entry into the tournament due to winning their conference’s championships. The remaining teams rely on a selection committee to be granted entry into the tournament. 32 teams which gain automatic entry into the tournament due to winning their \"Regular Season\" (not sure but my understanding).However , my point is some matchup will already be played before Selection day. \"Regular Season\" games are simply defined to be all games played on DayNum=132 or earlier (DayNum=132 is Selection Sunday). My question is when all these regular season already gonna be played and some has already played then what is the point to predict all matchups? It is like some of the data are already seen indirectly by the competitors for stage 2\n\n@addisonhoward , @jeffsonas can you please also shed some light to clear my confusion?",
    "747870": "None of the games you need to predict are played until day 136. Technically there are two tournament games that happen before day 136 (I think day 134) but we don't have to predict those - or if we do, they aren't actually counted.",
    "747900": "btesch i think day 134 or after. In both stages, the sample submission will tell you which games to predict right? Though bracket has to be announced on selection day by committee, what if given sample pair of teams is not going to be aligned with selection Sunday brackets?",
    "747901": "On Selection Sunday (day 132), the 68-team bracket will be announced.  You will need to predict winning percentages for all possible pairs of teams from those 68 teams.  The four \"play-in\" games played on Tuesday and Wednesday (day 134/5) will indeed be four of these possible pairs of teams, but it doesn't matter whether you predicted these games well; they are not included in the scoring.  All games played on day 136 or later (a total of 63 games) will count toward your score.  Prior to the submission deadline you will need to submit and select up to two complete submission files (which will include 68*67/2 data rows).  You can submit as many submission files as you like, but can only pick up to two of them that will count toward the contest.  And of course you have to pick them before the submission deadline as well.",
    "747902": "Once we see who the 68 teams are, we will include (as part of stage 2) a sample submission file which lists all possible pairs of those 68 teams and predicts a 50% winning percentage for each of them.  This way you can verify that you have the right 68 teams in your own submission files.",
    "747905": "jeffsonas thanks for replying , after selection day (68 teams announced) we will have only probably 4 to 5 days more to go for predicting from these announced team (bracket). \n\nIn that case why in data section it is mentioned \n\n\"Refer to the Timeline page for specific dates.** In both stages, the sample submission will tell you which games to predict**",
    "747919": "jeffsonas gotcha , it means we have to filter out those team using Selection Sunday team list and predict their outcome before 19 March 3PM UTC right?",
    "747921": "There is a file named MNCAATourneySeeds that lists the 68 teams (and their seeds) for each season.  You can either use this file to figure out the 68*67/2 combinations or you can just look at the sample submission file; either way it will tell you which matchups to predict.  The \"/2\" part of \"68*67/2\" is because we only predict data rows where the ID of the first team listed is less than the ID of the second team listed.  So if team 1101 is given a 60% chance to defeat team 1102 in the 2020 season tournament, then your data row will look like \"2020_1101_1102,0.60\" and it is not necessary to also provide a prediction like \"2020_1102_1101,0.40\" because that second prediction can be inferred from the first prediction.",
    "747922": "But yes, it is only teams picked on the Selection Sunday team list that you need to make predictions for, and you will need to make predictions for all possible pairs of those 68 teams.",
    "747944": "jeffsonas thanks a lot , what if sample submission pairs is not aligned with selection Sunday?",
    "747958": "I think there is error-checking in place when you try to submit a prediction file, and it gets rejected if the values in the first column do not match what is expected (such as if you picked the wrong set of 68 teams, or too many or too few).  I am not sure if they have to be sorted by the first column, but I don't think they do."
  },
  "source": "meta"
}