{
  "fetched_at": "2026-10-19T11:31:18Z",
  "source_url": "https://www.kaggle.com/competitions/duke-x-smt-march-madness-ml-challenge-2024/data",
  "pages": [
    {
      "name": "data-description",
      "content": "We sourced this data from the Kaggle March Madness Machine Learning Mania 2024 Kaggle page. You are not required to use all or any of this data to make your predictions, and you may use any public data sources as long as they are cited. They are merely offered as guides. For more information on this data, visit:\n\nhttps://www.kaggle.com/c/march-machine-learning-mania-2024/data",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Competition Rules",
      "content": "# Eligibility\n•\tAll team members must be enrolled in a university program at time of submission.\n•\tTeams may include up to three people.\n\n# Submission \n•\tOne submission per team\n•\tRequired Materials (three files)\n1.\tReadMe describing methodology\n•\tAny outside data sources must be cited.\n2.\tAnalysis code\n3.\tResults CSV: Submission_MM_With_Teams.csv \n•\tColumns: GameID, Team1WinPercentage\n•\tRows must provide win percentages for all possible matchups (at least 2,016). If a game occurs but is not included, it will be assigned a win percentage of 0 for both teams.\n\nEvaluation\n•\tOnly the Results CSV will be evaluated.\n•\tMetric: Brier Score (https://en.wikipedia.org/wiki/Brier_score)\n\n \n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "Our goal with this competition is for college students to be able to test their forecasting and data science skills while gaining valuable experience in a sports analytics context. We also want to improve on the existing bracket challenge model to incentivize accurate predictions rather than more of a game theory approach.\n\nYou will be forecasting the outcomes of the men's 2024 collegiate basketball tournament by submitting a CSV with the win percentages of every possible matchup in the tournament. Because there are 2016 possible matchups (after removing teams that lose in the first four, your submission CSV should have at least 2016 rows (you may include all First Four teams if you wish, but you will not be graded on First Four matchups). \nEach row should have a Team 1 column, a Team 2 column, and a Win Percentage column denoting the expected win percentage for Team 1 if these two teams played.\n\nBecause the first matchup of the tournament is March 21st, you have until then to make your submission. We will provide data to help you, but you may use any outside sources you wish for training and inference. Selection Sunday is March 17th, so of course you will not be able to generate your inferences until that date, but you should be able to train you model as early as you wish.\n\nTwo important notes: \n\n- This competition is open only to College Students.\n\n- The TOP 3 TEAMS WILL ALL RECEIVE SMT MERCH AND SWAG FOR WINNING!\n\nGood luck to everyone and Go Duke!",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "# Eligibility\n•\tAll team members must be enrolled in a university program at time of submission.\n•\tTeams may include up to three people.\n\n# Submission \n•\tOne submission per team\n•\tRequired Materials (three files)\n1.\tReadMe describing methodology\n•\tAny outside data sources must be cited.\n2.\tAnalysis code\n3.\tResults CSV: Submission_MM_With_Teams.csv \n•\tColumns: GameID, Team1WinPercentage\n•\tRows must provide win percentages for all possible matchups (at least 2,016). If a game occurs but is not included, it will be assigned a win percentage of 0 for both teams.\n\nEvaluation\n•\tOnly the Results CSV will be evaluated.\n•\tMetric: Brier Score (https://en.wikipedia.org/wiki/Brier_score)\n\n \n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "abstract",
      "content": "Duke Sports Analytics Club and SMT (SportsMEDIA Technology) are hosting a 2024 March Madness modeling competition, in which participants will use machine learning to forecast men’s tournament outcomes. \n\nThis competition incentivizes match outcome predictions rather than the more typical game theory approach. Models should forecast win percentages of every possible matchup in the Tournament of 64—a total of 2016. (This does not include the First Four teams. Those matchups can be included, but they will not be considered as part of evaluation.) \n\nRegistration is limited to US university students, with teams of up to three members. Participants will be provided with a training data set and external sources may be used as part of model development. Although Selection Sunday is not until March 17, you may begin model training immediately.\nNow available in the data section is **final_deduplicated.csv** with columns \n•\t*GameID*\n•\t*Team1*\n•\t*Team2*\n\nTeams must edit and submit their own versions of this file, containing ONLY *GameID* and adding a column for *Team1WinPercentage*.\n\nTeams are permitted one entry, which must be submitted by the start of the tournament on March 21. Winners will be announced shortly after the tournament ends, with prizes awarded to the top three teams.\n\n\nSend any questions to ethan.goldberg@duke.edu\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "We will give you a CSV titled Submission_MM_With_Teams.csv that you can edit to make your submission. Each row will have columns for ID(gameID), Team1, Team2, and Team1WinPercentage. We will initially release this CSV as a junk file with 0s as the true values for all the games and fake names for all of the teams. Once the team are released on March 17th we will update this file to contain every possible matchup for the tournament, still with the WinPercentages all initialized to junk. You must submit your own version of this file but with ONLY GameID and Team1WinPercentage included with the win percentage based on your model.\n\nThe evaluation metric will be the Brier Score. The submission with a Brier Score closest to 0 will be our champion. For more information on how that is calculated and why that is the best method for our purposes, check here:\n\nhttps://en.wikipedia.org/wiki/Brier_score\n\nIn general, all you need to know is to make the most accurate predictions possible and that you will be rewarded for that.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Dataset Description",
      "content": "We sourced this data from the Kaggle March Madness Machine Learning Mania 2024 Kaggle page. You are not required to use all or any of this data to make your predictions, and you may use any public data sources as long as they are cited. They are merely offered as guides. For more information on this data, visit:\n\nhttps://www.kaggle.com/c/march-machine-learning-mania-2024/data",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}