Last year, the NCAA introduced a new way of ranking teams using something called N.E.T. (described here). Now that we have a grand sample size of 1 season with this metric, what are your opinions about the predictability of this metric on tournament success?
", "messages": [ { "id": "753026", "postDate": "02/21/2020 16:16:33", "content": "Last year, the NCAA introduced a new way of ranking teams using something called N.E.T. (described here). Now that we have a grand sample size of 1 season with this metric, what are your opinions about the predictability of this metric on tournament success?
", "rawMarkdown": "Last year, the NCAA introduced a new way of ranking teams using something called N.E.T. ([described here](https://www.ncaa.com/news/basketball-men/article/2018-11-26/net-explained-ncaa-adopts-new-college-basketball-ranking)). Now that we have a grand sample size of 1 season with this metric, what are your opinions about the predictability of this metric on tournament success?", "votes": null }, { "id": "753041", "postDate": "02/21/2020 16:49:18", "content": "This is from Wiki
\n\nReplacement with NET for Division I men's basketball\nThe NCAA announced on August 22, 2018 that the RPI would no longer be used in the Division I men's basketball selection process, and would be replaced by the aforementioned NET. This new metric takes the following into account:\n-Game results\n-Strength of schedule\n-Game location (home, away, or neutral court)\n-Scoring margin — While included in the NET, teams receive no extra credit for wins by more than 10 points. Additionally, overtime games are assigned a 1-point victory margin, regardless of the actual score.\n-Net offensive and defensive efficiency\n-Quality of wins and losses, using the existing quadrant system\n-Game date and order are not included in the NET—all games are treated equally, whether an early-season matchup or a conference tournament championship game. As noted above, the NET is initially being used only for the Division I men's basketball tournament
\n\nMy opinion:\nA good power rating system should have most of the qualities listed above. It is interesting they capped margin of victory at 10 points. I'm guessing they did so because it's a nice round number but it's too low and loses predictive power. I'm sure they don't want to reward teams for running up the score either so 10 points is a compromise of sorts. I've given a lot of thought to the idea of making an OT win a fixed number margin of victory - possibly even zero. Logically, it seems that at the end of regulation, tied teams played equally and anything that happens afterwards is immaterial for comparison. I am certain it's better than the RPI however. \nThe question is how does it impact seeding? It influences it but even in RPI years, it was only a guide and not strictly adhered to. Certainly it will help to change some seedings but it's hard to know how because the seeding happens by a bunch of people locked in a room making cases for an against teams. \nUnless we know the actual formula, all we can offer is an opinion on predictability. I'm certain - if we as a collective - really wanted to know the formula, we could 'crack' it or come very close.
", "rawMarkdown": "This is from Wiki\n\nReplacement with NET for Division I men's basketball\nThe NCAA announced on August 22, 2018 that the RPI would no longer be used in the Division I men's basketball selection process, and would be replaced by the aforementioned NET. This new metric takes the following into account:\n-Game results\n-Strength of schedule\n-Game location (home, away, or neutral court)\n-Scoring margin — While included in the NET, teams receive no extra credit for wins by more than 10 points. Additionally, overtime games are assigned a 1-point victory margin, regardless of the actual score.\n-Net offensive and defensive efficiency\n-Quality of wins and losses, using the existing quadrant system\n-Game date and order are not included in the NET—all games are treated equally, whether an early-season matchup or a conference tournament championship game. As noted above, the NET is initially being used only for the Division I men's basketball tournament\n\nMy opinion:\nA good power rating system should have most of the qualities listed above. It is interesting they capped margin of victory at 10 points. I'm guessing they did so because it's a nice round number but it's too low and loses predictive power. I'm sure they don't want to reward teams for running up the score either so 10 points is a compromise of sorts. I've given a lot of thought to the idea of making an OT win a fixed number margin of victory - possibly even zero. Logically, it seems that at the end of regulation, tied teams played equally and anything that happens afterwards is immaterial for comparison. I am certain it's better than the RPI however. \nThe question is how does it impact seeding? It influences it but even in RPI years, it was only a guide and not strictly adhered to. Certainly it will help to change some seedings but it's hard to know how because the seeding happens by a bunch of people locked in a room making cases for an against teams. \nUnless we know the actual formula, all we can offer is an opinion on predictability. I'm certain - if we as a collective - really wanted to know the formula, we could 'crack' it or come very close.", "votes": null } ], "comments": [ { "id": 753041, "author_name": "btesch", "author_url": "", "post_date": "02/21/2020 16:49:18", "content": "This is from Wiki
\n\nReplacement with NET for Division I men's basketball\nThe NCAA announced on August 22, 2018 that the RPI would no longer be used in the Division I men's basketball selection process, and would be replaced by the aforementioned NET. This new metric takes the following into account:\n-Game results\n-Strength of schedule\n-Game location (home, away, or neutral court)\n-Scoring margin — While included in the NET, teams receive no extra credit for wins by more than 10 points. Additionally, overtime games are assigned a 1-point victory margin, regardless of the actual score.\n-Net offensive and defensive efficiency\n-Quality of wins and losses, using the existing quadrant system\n-Game date and order are not included in the NET—all games are treated equally, whether an early-season matchup or a conference tournament championship game. As noted above, the NET is initially being used only for the Division I men's basketball tournament
\n\nMy opinion:\nA good power rating system should have most of the qualities listed above. It is interesting they capped margin of victory at 10 points. I'm guessing they did so because it's a nice round number but it's too low and loses predictive power. I'm sure they don't want to reward teams for running up the score either so 10 points is a compromise of sorts. I've given a lot of thought to the idea of making an OT win a fixed number margin of victory - possibly even zero. Logically, it seems that at the end of regulation, tied teams played equally and anything that happens afterwards is immaterial for comparison. I am certain it's better than the RPI however. \nThe question is how does it impact seeding? It influences it but even in RPI years, it was only a guide and not strictly adhered to. Certainly it will help to change some seedings but it's hard to know how because the seeding happens by a bunch of people locked in a room making cases for an against teams. \nUnless we know the actual formula, all we can offer is an opinion on predictability. I'm certain - if we as a collective - really wanted to know the formula, we could 'crack' it or come very close.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "753026": "Last year, the NCAA introduced a new way of ranking teams using something called N.E.T. ([described here](https://www.ncaa.com/news/basketball-men/article/2018-11-26/net-explained-ncaa-adopts-new-college-basketball-ranking)). Now that we have a grand sample size of 1 season with this metric, what are your opinions about the predictability of this metric on tournament success?", "753041": "This is from Wiki\n\nReplacement with NET for Division I men's basketball\nThe NCAA announced on August 22, 2018 that the RPI would no longer be used in the Division I men's basketball selection process, and would be replaced by the aforementioned NET. This new metric takes the following into account:\n-Game results\n-Strength of schedule\n-Game location (home, away, or neutral court)\n-Scoring margin — While included in the NET, teams receive no extra credit for wins by more than 10 points. Additionally, overtime games are assigned a 1-point victory margin, regardless of the actual score.\n-Net offensive and defensive efficiency\n-Quality of wins and losses, using the existing quadrant system\n-Game date and order are not included in the NET—all games are treated equally, whether an early-season matchup or a conference tournament championship game. As noted above, the NET is initially being used only for the Division I men's basketball tournament\n\nMy opinion:\nA good power rating system should have most of the qualities listed above. It is interesting they capped margin of victory at 10 points. I'm guessing they did so because it's a nice round number but it's too low and loses predictive power. I'm sure they don't want to reward teams for running up the score either so 10 points is a compromise of sorts. I've given a lot of thought to the idea of making an OT win a fixed number margin of victory - possibly even zero. Logically, it seems that at the end of regulation, tied teams played equally and anything that happens afterwards is immaterial for comparison. I am certain it's better than the RPI however. \nThe question is how does it impact seeding? It influences it but even in RPI years, it was only a guide and not strictly adhered to. Certainly it will help to change some seedings but it's hard to know how because the seeding happens by a bunch of people locked in a room making cases for an against teams. \nUnless we know the actual formula, all we can offer is an opinion on predictability. I'm certain - if we as a collective - really wanted to know the formula, we could 'crack' it or come very close." }, "source": "meta" }