UPDATED: Tournaments have been canceled. See this post for the latest updates.
\n\nHey All,
\n\nThe latest announcement from the NCAA has announced that they will continue with the NCAA tournaments, but with no fans in attendance (limited fans like family and essential personnel are allowed).
\n\nWe have no intention of canceling this Kaggle competition as long as the games are still scheduled to occur. As there will be no one in attendance, make appropriate adjustments to your model (as you see fit) should any weights accounts for a \"home-team advantage\" or some related geography feature.
\n\nThe NBA regular season has been suspended, as a player from the Utah Jazz has tested positive for the coronavirus. We don't think think this will have an impact on the NCAA tournament, but we're monitoring the news to see if anything changes.
\n\nIn the rare instance that the full tournament be postponed, suspended, or canceled, we (Kaggle) will consider alternatives to keep the Kaggle community involved in some form or fashion, with a worst-case scenario being that we cancel the competition.
\n\nAs we fully anticipate the tournament to move forward, let the feature engineering begin in lieu of recent news!
\n\nHappy Modeling!
\n\nKaggle
", "messages": [ { "id": 769646, "postDate": "2020-03-12T05:35:43.117Z", "content": "UPDATED: Tournaments have been canceled. See this post for the latest updates.
\n\nHey All,
\n\nThe latest announcement from the NCAA has announced that they will continue with the NCAA tournaments, but with no fans in attendance (limited fans like family and essential personnel are allowed).
\n\nWe have no intention of canceling this Kaggle competition as long as the games are still scheduled to occur. As there will be no one in attendance, make appropriate adjustments to your model (as you see fit) should any weights accounts for a \"home-team advantage\" or some related geography feature.
\n\nThe NBA regular season has been suspended, as a player from the Utah Jazz has tested positive for the coronavirus. We don't think think this will have an impact on the NCAA tournament, but we're monitoring the news to see if anything changes.
\n\nIn the rare instance that the full tournament be postponed, suspended, or canceled, we (Kaggle) will consider alternatives to keep the Kaggle community involved in some form or fashion, with a worst-case scenario being that we cancel the competition.
\n\nAs we fully anticipate the tournament to move forward, let the feature engineering begin in lieu of recent news!
\n\nHappy Modeling!
\n\nKaggle
", "rawMarkdown": "UPDATED: Tournaments have been canceled. See [this post](https://www.kaggle.com/c/google-cloud-ncaa-march-madness-2020-division-1-mens-tournament/discussion/135258) for the latest updates.\n\nHey All,\n\nThe [latest announcement from the NCAA](https://www.ncaa.com/news/basketball-men/article/2020-03-11/ncaa-announces-division-i-mens-and-womens-basketball) has announced that they will continue with the NCAA tournaments, but with no fans in attendance (limited fans like family and essential personnel are allowed). \n\nWe have no intention of canceling this Kaggle competition as long as the games are still scheduled to occur. As there will be no one in attendance, make appropriate adjustments to your model (as you see fit) should any weights accounts for a \"home-team advantage\" or some related geography feature.\n\nThe [NBA regular season has been suspended](https://www.espn.com/nba/story/_/id/28887560/nba-suspends-season-further-notice-player-tests-positive-coronavirus), as a player from the Utah Jazz has tested positive for the coronavirus. We don't think think this will have an impact on the NCAA tournament, but we're monitoring the news to see if anything changes. \n\nIn the rare instance that the full tournament be postponed, suspended, or canceled, we (Kaggle) will consider alternatives to keep the Kaggle community involved in some form or fashion, with a worst-case scenario being that we cancel the competition.\n\nAs we fully anticipate the tournament to move forward, let the feature engineering begin in lieu of recent news! \n\nHappy Modeling!\n\nKaggle", "votes": 16 }, { "id": 770319, "postDate": "2020-03-12T20:21:06.043Z", "content": "Cancelled https://twitter.com/NCAA/status/1238197201928667136?s=20 probably worth pinning the thread at the top
", "rawMarkdown": "Cancelled https://twitter.com/NCAA/status/1238197201928667136?s=20 probably worth pinning the thread at the top", "votes": 1 }, { "id": 770258, "postDate": "2020-03-12T18:39:43.670Z", "content": "Duke and KU have both announced that they are suspending all athletics for the Spring. I don't see any way the march madness happens at this point.
", "rawMarkdown": "Duke and KU have both announced that they are suspending all athletics for the Spring. I don't see any way the march madness happens at this point. ", "votes": 1 }, { "id": 770240, "postDate": "2020-03-12T18:12:50.600Z", "content": "I'm wondering what the alternatives could be?
\n\nUsing only data prior to 2014, predicting all of the 2014-2019 results? I think it could work as long as it was a kernels comp. \nMaybe a simulated tournament with simulated historicals?\nCould a simulated 2020 tourny work with actual historicals?
\n\nMaybe I'm looking at this from too narrow of a scope.
\n\nGiven the various pro league announcements, it seems this won't happen - at least on schedule.
\n\nEDIT - *and now we get to find out. *
", "rawMarkdown": "I'm wondering what the alternatives could be?\n\nUsing only data prior to 2014, predicting all of the 2014-2019 results? I think it could work as long as it was a kernels comp. \nMaybe a simulated tournament with simulated historicals?\nCould a simulated 2020 tourny work with actual historicals?\n\nMaybe I'm looking at this from too narrow of a scope. \n\nGiven the various pro league announcements, it seems this won't happen - at least on schedule. \n\n\nEDIT - *and now we get to find out. *\n", "votes": 1, "replies": [ { "id": 770826, "postDate": "2020-03-13T13:08:42.373Z", "content": "We can create useless overfit models for previous competitions, even if we weren't allowed to use post-2014 data.
", "rawMarkdown": "We can create useless overfit models for previous competitions, even if we weren't allowed to use post-2014 data.", "votes": 1 }, { "id": 771019, "postDate": "2020-03-13T17:00:24.667Z", "content": "We could debate the value of the effort but I suspect no one would get a model above 77%-78% accurate. It's not an easy undertaking. \nOr to take it one step further, try to predict the tourny winner or who will survive and until what round using only pre 2014 data.
", "rawMarkdown": "We could debate the value of the effort but I suspect no one would get a model above 77%-78% accurate. It's not an easy undertaking. \nOr to take it one step further, try to predict the tourny winner or who will survive and until what round using only pre 2014 data. \n" }, { "id": 852501, "postDate": "2020-05-18T13:40:53.320Z", "content": "Right.
", "rawMarkdown": "Right." } ] }, { "id": 770253, "postDate": "2020-03-12T18:33:13.133Z", "content": "If I'm the first competitor to officially predict that the tournament will be cancelled, do I win the prize money? My predicted probability of this outcome is 1.0. Where do I submit?
\n\nhttps://twitter.com/ESPNRittenberg/status/1238166962422927360?s=20
", "rawMarkdown": "If I'm the first competitor to officially predict that the tournament will be cancelled, do I win the prize money? My predicted probability of this outcome is 1.0. Where do I submit?\n\nhttps://twitter.com/ESPNRittenberg/status/1238166962422927360?s=20", "votes": 2 }, { "id": 770829, "postDate": "2020-03-13T13:11:38.703Z", "rawMarkdown": "", "isDeleted": true, "replies": [ { "id": 770845, "postDate": "2020-03-13T13:25:31.777Z", "rawMarkdown": "", "isDeleted": true } ] }, { "id": 770320, "postDate": "2020-03-12T20:21:30.737Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 770319, "author_name": "Blake Atkinson", "author_url": "", "post_date": "2020-03-12T20:21:06.043000", "content": "Cancelled https://twitter.com/NCAA/status/1238197201928667136?s=20 probably worth pinning the thread at the top
", "votes": 1, "replies": [] }, { "id": 770258, "author_name": "Marginal Returns", "author_url": "", "post_date": "2020-03-12T18:39:43.670000", "content": "Duke and KU have both announced that they are suspending all athletics for the Spring. I don't see any way the march madness happens at this point.
", "votes": 1, "replies": [] }, { "id": 770240, "author_name": "Ben Tesch", "author_url": "", "post_date": "2020-03-12T18:12:50.600000", "content": "I'm wondering what the alternatives could be?
\n\nUsing only data prior to 2014, predicting all of the 2014-2019 results? I think it could work as long as it was a kernels comp. \nMaybe a simulated tournament with simulated historicals?\nCould a simulated 2020 tourny work with actual historicals?
\n\nMaybe I'm looking at this from too narrow of a scope.
\n\nGiven the various pro league announcements, it seems this won't happen - at least on schedule.
\n\nEDIT - *and now we get to find out. *
", "votes": 1, "replies": [ { "id": 770826, "author_name": "", "author_url": "", "post_date": "2020-03-13T13:08:42.373000", "content": "We can create useless overfit models for previous competitions, even if we weren't allowed to use post-2014 data.
", "votes": 1, "replies": [] }, { "id": 771019, "author_name": "Ben Tesch", "author_url": "", "post_date": "2020-03-13T17:00:24.667000", "content": "We could debate the value of the effort but I suspect no one would get a model above 77%-78% accurate. It's not an easy undertaking. \nOr to take it one step further, try to predict the tourny winner or who will survive and until what round using only pre 2014 data.
", "votes": 0, "replies": [] }, { "id": 852501, "author_name": "Mukesh Choudhary", "author_url": "", "post_date": "2020-05-18T13:40:53.320000", "content": "Right.
", "votes": 0, "replies": [] } ] }, { "id": 770253, "author_name": "Josh Stone", "author_url": "", "post_date": "2020-03-12T18:33:13.133000", "content": "If I'm the first competitor to officially predict that the tournament will be cancelled, do I win the prize money? My predicted probability of this outcome is 1.0. Where do I submit?
\n\nhttps://twitter.com/ESPNRittenberg/status/1238166962422927360?s=20
", "votes": 2, "replies": [] }, { "id": 770829, "author_name": "", "author_url": "", "post_date": "2020-03-13T13:11:38.703000", "content": "", "votes": 0, "replies": [ { "id": 770845, "author_name": "", "author_url": "", "post_date": "2020-03-13T13:25:31.777000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 770320, "author_name": "", "author_url": "", "post_date": "2020-03-12T20:21:30.737000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "769646": "UPDATED: Tournaments have been canceled. See [this post](https://www.kaggle.com/c/google-cloud-ncaa-march-madness-2020-division-1-mens-tournament/discussion/135258) for the latest updates.\n\nHey All,\n\nThe [latest announcement from the NCAA](https://www.ncaa.com/news/basketball-men/article/2020-03-11/ncaa-announces-division-i-mens-and-womens-basketball) has announced that they will continue with the NCAA tournaments, but with no fans in attendance (limited fans like family and essential personnel are allowed). \n\nWe have no intention of canceling this Kaggle competition as long as the games are still scheduled to occur. As there will be no one in attendance, make appropriate adjustments to your model (as you see fit) should any weights accounts for a \"home-team advantage\" or some related geography feature.\n\nThe [NBA regular season has been suspended](https://www.espn.com/nba/story/_/id/28887560/nba-suspends-season-further-notice-player-tests-positive-coronavirus), as a player from the Utah Jazz has tested positive for the coronavirus. We don't think think this will have an impact on the NCAA tournament, but we're monitoring the news to see if anything changes. \n\nIn the rare instance that the full tournament be postponed, suspended, or canceled, we (Kaggle) will consider alternatives to keep the Kaggle community involved in some form or fashion, with a worst-case scenario being that we cancel the competition.\n\nAs we fully anticipate the tournament to move forward, let the feature engineering begin in lieu of recent news! \n\nHappy Modeling!\n\nKaggle", "770319": "Cancelled https://twitter.com/NCAA/status/1238197201928667136?s=20 probably worth pinning the thread at the top", "770258": "Duke and KU have both announced that they are suspending all athletics for the Spring. I don't see any way the march madness happens at this point. ", "770240": "I'm wondering what the alternatives could be?\n\nUsing only data prior to 2014, predicting all of the 2014-2019 results? I think it could work as long as it was a kernels comp. \nMaybe a simulated tournament with simulated historicals?\nCould a simulated 2020 tourny work with actual historicals?\n\nMaybe I'm looking at this from too narrow of a scope. \n\nGiven the various pro league announcements, it seems this won't happen - at least on schedule. \n\n\nEDIT - *and now we get to find out. *\n", "770253": "If I'm the first competitor to officially predict that the tournament will be cancelled, do I win the prize money? My predicted probability of this outcome is 1.0. Where do I submit?\n\nhttps://twitter.com/ESPNRittenberg/status/1238166962422927360?s=20", "770829": "", "770320": "" } }