It doesn't look like any of the provided data sets provide the overall seeds of the teams in the tourneys, from 1-68(64 plus 4 play in teams). Has anyone found a place that offers this historical data? It seems like that would allow for a much better model.
", "messages": [ { "id": 755804, "postDate": "2020-02-25T06:16:35.343Z", "content": "Wow, I thought this was a new thing - one or two years ago. Apparently they have been releasing this information for quite some time. Looks like a good dataset for next year - unless someone wants to clean it up for this year...
\n\n2019: https://www.ncaa.com/news/basketball-men/2019-03-18/march-madness-2019-every-teams-seed-no-1-duke-no-68-nc-central\n2018: https://www.ncaa.com/news/basketball-men/article/2018-03-13/heres-every-march-madness-team-ranked-and-1-thing-know-about\n2017: https://www.ncaa.com/news/basketball-men/article/2017-03-12/march-madness-bracket-every-seed-ranked-68-1\n2016: https://www.cbssports.com/college-basketball/news/2016-ncaa-bracket-tournament-committees-official-1-68-seed-list/\n2015: https://www.cbssports.com/college-basketball/news/2015-ncaa-tournament-selection-committee-releases-1-68-ranking/\n2014: https://www.cbssports.com/college-basketball/news/official-ncaa-1-68-seeding-order/\n2013: https://www.sbnation.com/2013/3/17/4117000/ncaa-bracket-2013-seed-list-selection-committee\n2012: https://www.sbnation.com/2012-ncaa-tournament/2012/3/11/2863068/ncaa-bracket-2012-overall-seeds
", "rawMarkdown": "Wow, I thought this was a new thing - one or two years ago. Apparently they have been releasing this information for quite some time. Looks like a good dataset for next year - unless someone wants to clean it up for this year...\n\n2019: https://www.ncaa.com/news/basketball-men/2019-03-18/march-madness-2019-every-teams-seed-no-1-duke-no-68-nc-central\n2018: https://www.ncaa.com/news/basketball-men/article/2018-03-13/heres-every-march-madness-team-ranked-and-1-thing-know-about\n2017: https://www.ncaa.com/news/basketball-men/article/2017-03-12/march-madness-bracket-every-seed-ranked-68-1\n2016: https://www.cbssports.com/college-basketball/news/2016-ncaa-bracket-tournament-committees-official-1-68-seed-list/\n2015: https://www.cbssports.com/college-basketball/news/2015-ncaa-tournament-selection-committee-releases-1-68-ranking/\n2014: https://www.cbssports.com/college-basketball/news/official-ncaa-1-68-seeding-order/\n2013: https://www.sbnation.com/2013/3/17/4117000/ncaa-bracket-2013-seed-list-selection-committee\n2012: https://www.sbnation.com/2012-ncaa-tournament/2012/3/11/2863068/ncaa-bracket-2012-overall-seeds\n", "votes": 3 }, { "id": 755690, "postDate": "2020-02-25T03:06:50.930Z", "content": "It doesn't look like any of the provided data sets provide the overall seeds of the teams in the tourneys, from 1-68(64 plus 4 play in teams). Has anyone found a place that offers this historical data? It seems like that would allow for a much better model.
", "rawMarkdown": "It doesn't look like any of the provided data sets provide the overall seeds of the teams in the tourneys, from 1-68(64 plus 4 play in teams). Has anyone found a place that offers this historical data? It seems like that would allow for a much better model.", "votes": 3 }, { "id": 759243, "postDate": "2020-02-28T19:20:47.607Z", "content": "It is worth noting that these overall rankings become nothing more/less than another set of ordinal rankings, similar to the Massey Ordinals. I would probably place higher importance on Massey Ordinals from more scientific sources like Pomeroy, Sagarin, etc. The NCAA overall 1-68 rankings are subject to some constraints, like they don't want to have teams from the same conference playing each other too early in the tournament, and they (I think) may be influenced by the geographical location of the tournament games when placing teams in the regionals, etc. So I would expect (totally off the cuff) that these are not markedly better than just using the seed numbers that range from 1-16.
", "rawMarkdown": "It is worth noting that these overall rankings become nothing more/less than another set of ordinal rankings, similar to the Massey Ordinals. I would probably place higher importance on Massey Ordinals from more scientific sources like Pomeroy, Sagarin, etc. The NCAA overall 1-68 rankings are subject to some constraints, like they don't want to have teams from the same conference playing each other too early in the tournament, and they (I think) may be influenced by the geographical location of the tournament games when placing teams in the regionals, etc. So I would expect (totally off the cuff) that these are not markedly better than just using the seed numbers that range from 1-16.", "votes": 2, "replies": [ { "id": 760236, "postDate": "2020-03-01T02:33:08.750Z", "content": "That's true. I'm not sure how much the NCAA overall rankings will help, but I figure it can't be any less accurate than the 1-16 rankings.
", "rawMarkdown": "That's true. I'm not sure how much the NCAA overall rankings will help, but I figure it can't be any less accurate than the 1-16 rankings." } ] }, { "id": 756118, "postDate": "2020-02-25T12:38:04.200Z", "content": "I have created a dataset from @jeffsonas sources below at https://www.kaggle.com/christoffer/march-madness-overall-seeds-men
\n\n(I have done some sanity checking, but I don't rule out I might have swapped some teams. 🙂 )
", "rawMarkdown": "I have created a dataset from @jeffsonas sources below at https://www.kaggle.com/christoffer/march-madness-overall-seeds-men\n\n(I have done some sanity checking, but I don't rule out I might have swapped some teams. 🙂 )", "votes": 2, "replies": [ { "id": 760233, "postDate": "2020-03-01T02:30:14.207Z", "content": "Thanks for doing that!
", "rawMarkdown": "Thanks for doing that!" } ] }, { "id": 755738, "postDate": "2020-02-25T04:30:10.567Z", "content": "I know that last year they released the overall 1-68 ranks - https://www.ncaa.com/news/basketball-men/2019-03-18/march-madness-2019-every-teams-seed-no-1-duke-no-68-nc-central - but I don't know that they have been doing this for very long.
", "rawMarkdown": "I know that last year they released the overall 1-68 ranks - https://www.ncaa.com/news/basketball-men/2019-03-18/march-madness-2019-every-teams-seed-no-1-duke-no-68-nc-central - but I don't know that they have been doing this for very long." }, { "id": 755735, "postDate": "2020-02-25T04:26:18.213Z", "rawMarkdown": "", "isDeleted": true, "replies": [ { "id": 755765, "postDate": "2020-02-25T05:28:35.737Z", "rawMarkdown": "", "isDeleted": true } ] } ], "comments": [ { "id": 755804, "author_name": "Jeff Sonas", "author_url": "", "post_date": "2020-02-25T06:16:35.343000", "content": "Wow, I thought this was a new thing - one or two years ago. Apparently they have been releasing this information for quite some time. Looks like a good dataset for next year - unless someone wants to clean it up for this year...
\n\n2019: https://www.ncaa.com/news/basketball-men/2019-03-18/march-madness-2019-every-teams-seed-no-1-duke-no-68-nc-central\n2018: https://www.ncaa.com/news/basketball-men/article/2018-03-13/heres-every-march-madness-team-ranked-and-1-thing-know-about\n2017: https://www.ncaa.com/news/basketball-men/article/2017-03-12/march-madness-bracket-every-seed-ranked-68-1\n2016: https://www.cbssports.com/college-basketball/news/2016-ncaa-bracket-tournament-committees-official-1-68-seed-list/\n2015: https://www.cbssports.com/college-basketball/news/2015-ncaa-tournament-selection-committee-releases-1-68-ranking/\n2014: https://www.cbssports.com/college-basketball/news/official-ncaa-1-68-seeding-order/\n2013: https://www.sbnation.com/2013/3/17/4117000/ncaa-bracket-2013-seed-list-selection-committee\n2012: https://www.sbnation.com/2012-ncaa-tournament/2012/3/11/2863068/ncaa-bracket-2012-overall-seeds
", "votes": 3, "replies": [] }, { "id": 759243, "author_name": "Jeff Sonas", "author_url": "", "post_date": "2020-02-28T19:20:47.607000", "content": "It is worth noting that these overall rankings become nothing more/less than another set of ordinal rankings, similar to the Massey Ordinals. I would probably place higher importance on Massey Ordinals from more scientific sources like Pomeroy, Sagarin, etc. The NCAA overall 1-68 rankings are subject to some constraints, like they don't want to have teams from the same conference playing each other too early in the tournament, and they (I think) may be influenced by the geographical location of the tournament games when placing teams in the regionals, etc. So I would expect (totally off the cuff) that these are not markedly better than just using the seed numbers that range from 1-16.
", "votes": 2, "replies": [ { "id": 760236, "author_name": "Aaron Huskerson", "author_url": "", "post_date": "2020-03-01T02:33:08.750000", "content": "That's true. I'm not sure how much the NCAA overall rankings will help, but I figure it can't be any less accurate than the 1-16 rankings.
", "votes": 0, "replies": [] } ] }, { "id": 756118, "author_name": "Christoffer Karlsson", "author_url": "", "post_date": "2020-02-25T12:38:04.200000", "content": "I have created a dataset from @jeffsonas sources below at https://www.kaggle.com/christoffer/march-madness-overall-seeds-men
\n\n(I have done some sanity checking, but I don't rule out I might have swapped some teams. 🙂 )
", "votes": 2, "replies": [ { "id": 760233, "author_name": "Aaron Huskerson", "author_url": "", "post_date": "2020-03-01T02:30:14.207000", "content": "Thanks for doing that!
", "votes": 0, "replies": [] } ] }, { "id": 755738, "author_name": "Jeff Sonas", "author_url": "", "post_date": "2020-02-25T04:30:10.567000", "content": "I know that last year they released the overall 1-68 ranks - https://www.ncaa.com/news/basketball-men/2019-03-18/march-madness-2019-every-teams-seed-no-1-duke-no-68-nc-central - but I don't know that they have been doing this for very long.
", "votes": 0, "replies": [] }, { "id": 755735, "author_name": "", "author_url": "", "post_date": "2020-02-25T04:26:18.213000", "content": "", "votes": 0, "replies": [ { "id": 755765, "author_name": "", "author_url": "", "post_date": "2020-02-25T05:28:35.737000", "content": "", "votes": 0, "replies": [] } ] } ], "raw_markdown_by_id": { "755804": "Wow, I thought this was a new thing - one or two years ago. Apparently they have been releasing this information for quite some time. Looks like a good dataset for next year - unless someone wants to clean it up for this year...\n\n2019: https://www.ncaa.com/news/basketball-men/2019-03-18/march-madness-2019-every-teams-seed-no-1-duke-no-68-nc-central\n2018: https://www.ncaa.com/news/basketball-men/article/2018-03-13/heres-every-march-madness-team-ranked-and-1-thing-know-about\n2017: https://www.ncaa.com/news/basketball-men/article/2017-03-12/march-madness-bracket-every-seed-ranked-68-1\n2016: https://www.cbssports.com/college-basketball/news/2016-ncaa-bracket-tournament-committees-official-1-68-seed-list/\n2015: https://www.cbssports.com/college-basketball/news/2015-ncaa-tournament-selection-committee-releases-1-68-ranking/\n2014: https://www.cbssports.com/college-basketball/news/official-ncaa-1-68-seeding-order/\n2013: https://www.sbnation.com/2013/3/17/4117000/ncaa-bracket-2013-seed-list-selection-committee\n2012: https://www.sbnation.com/2012-ncaa-tournament/2012/3/11/2863068/ncaa-bracket-2012-overall-seeds\n", "755690": "It doesn't look like any of the provided data sets provide the overall seeds of the teams in the tourneys, from 1-68(64 plus 4 play in teams). Has anyone found a place that offers this historical data? It seems like that would allow for a much better model.", "759243": "It is worth noting that these overall rankings become nothing more/less than another set of ordinal rankings, similar to the Massey Ordinals. I would probably place higher importance on Massey Ordinals from more scientific sources like Pomeroy, Sagarin, etc. The NCAA overall 1-68 rankings are subject to some constraints, like they don't want to have teams from the same conference playing each other too early in the tournament, and they (I think) may be influenced by the geographical location of the tournament games when placing teams in the regionals, etc. So I would expect (totally off the cuff) that these are not markedly better than just using the seed numbers that range from 1-16.", "756118": "I have created a dataset from @jeffsonas sources below at https://www.kaggle.com/christoffer/march-madness-overall-seeds-men\n\n(I have done some sanity checking, but I don't rule out I might have swapped some teams. 🙂 )", "755738": "I know that last year they released the overall 1-68 ranks - https://www.ncaa.com/news/basketball-men/2019-03-18/march-madness-2019-every-teams-seed-no-1-duke-no-68-nc-central - but I don't know that they have been doing this for very long.", "755735": "" } }