Ostensibly Hoops

[Saint]

Why don't they just feed the tourney into AI and play it that way? It'll be about the same as playing it without all the extra cost and trouble of actually playing it.

[Saint]

I picked Florida. Are they any good?

[hedge]

Duke won't win.

Karolina uber alles...

[aTm]

Why don't they just feed the tourney into AI and play it that way? It'll be about the same as playing it without all the extra cost and trouble of actually playing it.

Thats already what this all is (the universe)

[eCat]

I have Florida winning it all

[hedge]

Mighty prouda ya...

[eCat]

big risk taker I am

[Saint]

I picked Florida for Gaytor

[sardis]

Is he dead, too?

[hedge]

[eCat]

so UK spends twice as much at UNC on basketball

I bet that won't stand in Chapel Hill

[Dave23]

Memphis is going to go up substantially if they want to keep PJ Haggerty…

Edit: and Dain Dainja for that matter…

[hedge]

so UK spends twice as much at UNC on basketball

Kids come to UNC for the love...

[eCat]

asian whores?

[hedge]

we need to remove the human element entirely.

every team should know from day one where they stand getting into the tournament based on a ranking system. One that values the entire season. So that if a player does get hurt in the last 5 games it doesn't derail their entire season.

I agree.

A computer algorithm does all this. I could do the whole thing with four or five data tables, a couple stored procedures, a single report view, and one daily ETL feed updating all division 1 wins and losses. Then everyone would know all the time exactly where they stand and what would have to happen to get into the tournament.

The NET ranking basically is an algorithm. Y'all are pissed Carolina got in just b/c you're haters, and I get that, but citing Quad 1 record is just pure emotionalism. Not all Quad 1 teams (or games) are nearly the same, to just point to a team's final Quad 1 record is short-sighted and dumb. From IC:

"The NET (NCAA Evaluation Tool) rankings in NCAA basketball consider game results, strength of schedule, game location, net offensive and defensive efficiency, and the quality of wins and losses to determine a team's overall strength and tournament resume. This is the NCAA's primary sorting tool for evaluating teams. The NET is designed to be most optimal in March, not in early December.

There it is. UNC’s NET is better than anyone else’s on the bubble. The entire NET thing is in place to measure and give defendable and objective criteria to split hairs between bubble teams.
If a team garners 3 or 4 Q1 W’s at home against NET ranked teams in the upper 30’s, that might not weigh much more - if any - than another team losing some Q1 games by single digits, away from home, against NET ranked teams 1-9.

If all of those Q1 losses were too give UNC such a black eye, why was the NET ranking 36? The best amongst the bubbles. How many teams in the tourney have a NET ranking worse than the Tar Heels #36?

[hedge]

I'll answer my own question. Thirty. 30 teams in the NCAAT have a worse NET ranking than do the Heels. 30 teams.
West Va. - who didn't make it - has a NET of 51 and a record of 19-13. They didn't even make the 20 W plateau. Lost their one and only game in their own conference tourney. The only teams in the tournament who didn't reach 20 W's for the season are the low, mid-majors who lucked up and won their conference auto bid like LOLstate did last year (looking at you St. Francis, Alabama St., ) So, STFU West Va.

So if indeed the NET is the beginning, middle and end all of splitting these hairs amongst the bubbles... I really think the committee looked at Carolina's NET over-all and not just the Q1 games - which were mostly ALL away games and against top tier NET teams. not lowly ranked NET teams - for the most part.

UNC's NET is better than Utah St., Creighton, Texas, Arkansas, New Mexico, Oklahoma, Colorado St. , Vandy, Memphis. Teams who made it comfortably into the tournament without any bubbles. And our match up: San Diego St., checks in at #52

I'm done with the discussion. The committee got it right. The Heel's deserve to be in. Period. End of story.

[hedge]

The quad system is so stupid. Not all teams within a certain quad are remotely equal. You could have one team that plays 10 quad 1 games against top 10 teams and another team that plays 10 quad 1 games that consist of 3 home games vs. teams ranked between 25-30, 3 neutral court games vs. teams ranked games between 40-50, and 4 away games vs. teams ranked 60-75. And a home game vs. the #1 team is worth the same as a home game vs. the #30 team, but a home game vs. the #31 team is worth significantly less than a home game vs. the #30 team. An away game vs. the #1 team is worth the same as an away game vs. the #75 team, but an away game vs. the #76 team is worth significantly less than away game vs. the #75 team.

[aTm]

NET barely takes wins and losses into consideration at all. Thats why UNC looks can still look good. Its an efficiency ranking. Scoring more points per 100 possessions is better, allowing fewer points per 100 possessions is better. The raw numbers are adjusted for the strength of the opponent you played against and where the game was played. The NCAA adds some bonus points for certain big wins but it doesnt impact much at all from a more pure efficiency metric like kenpom rating

As such, NET is therefore no more influential in determining who deserves to be in the NCAA Tournament than any other efficiency metric. These metrics are "predictive" in the sense that they tell you who can be expected to win a future game based on each team's past efficiency.

The biggest factor in whether a team deserves to make the NCAA Tournament is the results or "Resume." This is "who did you actually beat and who did you lose to, and what was the location of those games."

You can't really just use an efficiency ranking like NET alone, because an efficiency ranking relegates the most important stat, wins and losses, to a secondary status. An SEC team in a weird year like this year could go 0-18 and if they lost every single game by only one single point, their efficiency or NET rankings would actually look very good because they would seem like they are just about as good as Auburn, Alabama, Florida, Tennessee, A&M, Kentucky, etc etc. The reality though, would be that they didn't beat anybody.

The main purpose of the NET ranking is to judge the quality of those wins and losses and therefore have a greater understanding of the Resume. Knowing that Team A went 1-0 and Team B went 1-0 tells you nothing. But combined with a predictive rating like NET, you can judge the results and see that Team A beat the team expected to be the 5th best in the country on the road, while Team B beat the team expected to be #117 at home, for example.

Thats why they do all of the subsequent dividing up wins into Quads and quad records. It helps them judge resumes that consist of 30+ games for each team.

[hedge]

"NET barely takes wins and losses into consideration at all. Thats why UNC looks good. Its an efficiency ranking. Scoring more points per 100 possessions is better, allowing fewer points per 100 possessions is better."

But doesn't that take the so-called human element out of it? It's natural for fans to just look at raw wins and losses and not care about anything else, but wouldn't an algorithm that takes into account all the things that the NET does, while not giving disproportionate weight to the raw win/loss numbers, be a better indicator of the quality of any given team?

[aTm]

The quad rankings are just used to summarize each team results. They can and do look at who a team beat and lost to individually. UNC’s resume was very weak because they didnt beat any trams that NET considered good.