NET Rankings

[stever20]

So one of the things I was really curious about was the impact of opponent games....

Big East had 2 games where their opponents played each other last night.
Creighton had played Clemson
Seton Hall had played Nebraska

Nebraska beat Clemson.

Seton Hall improved from 85 to 74 as a result
Creighton dropped from 36 to 41 as a result

seems pretty reasonable....

As a whole the league moved up from 70.2 up to 68.3 today.

[adoraz]

I posted this over at Redmen, any thoughts?

As you may know, the NCAA released their first NET ratings a few days ago. They have received a lot of criticism, however the system could work in our favor.

There is something pretty interesting with the (partially) secret formula- it appears to only marginally take into account level of opponent. Instead, it focuses a lot more on whether you win/lose and by what margin- regardless of what team you played.

Obviously, given our weak OOC.... this could actually be beneficial.

Take, for example, last night:
https://www.ncaa.com/rankings/basketbal ... t-rankings

SJU is now #18, up from #32 yesterday. I was surprised to see that, because we beat a horrible opponent (#340 UMES) at home. With RPI the opposite happened and we dropped from #4 to #15 because our opponent was so bad. I noticed yesterday too that other undefeated teams with weak schedules (such as NC State/Pitt) were extremely overvalued in NET compared to other rankings. To me and a lot of others, the system seems flawed in this regard.

The best part about this is we already played the tougher half of our OOC schedule. If we can eat up cupcakes during December without dropping in the rankings (like what will happen with RPI), that'd be pretty amazing.

Now, to be honest it wouldn't surprise me if the NCAA adjusted the numbers to more heavily weigh in level of opponent. They are keeping some parts of the formula a secret, so they could change it gradually and we wouldn't notice. At least for now though, things may be looking up.

[CPJays]

I posted this over at Redmen, any thoughts?

As you may know, the NCAA released their first NET ratings a few days ago. They have received a lot of criticism, however the system could work in our favor.

There is something pretty interesting with the (partially) secret formula- it appears to only marginally take into account level of opponent. Instead, it focuses a lot more on whether you win/lose and by what margin- regardless of what team you played.

Obviously, given our weak OOC.... this could actually be beneficial.

Take, for example, last night:
https://www.ncaa.com/rankings/basketbal ... t-rankings

SJU is now #18, up from #32 yesterday. I was surprised to see that, because we beat a horrible opponent (#340 UMES) at home. With RPI the opposite happened and we dropped from #4 to #15 because our opponent was so bad. I noticed yesterday too that other undefeated teams with weak schedules (such as NC State/Pitt) were extremely overvalued in NET compared to other rankings. To me and a lot of others, the system seems flawed in this regard.

The best part about this is we already played the tougher half of our OOC schedule. If we can eat up cupcakes during December without dropping in the rankings (like what will happen with RPI), that'd be pretty amazing.

Now, to be honest it wouldn't surprise me if the NCAA adjusted the numbers to more heavily weigh in level of opponent. They are keeping some parts of the formula a secret, so they could change it gradually and we wouldn't notice. At least for now though, things may be looking up.

Looking at it daily is probably going to be an unhealthy practice for me.

Creighton jumped up 10 spots in a day they didn't play and one of their previous opponents lost.

[stever20]

I posted this over at Redmen, any thoughts?

As you may know, the NCAA released their first NET ratings a few days ago. They have received a lot of criticism, however the system could work in our favor.

There is something pretty interesting with the (partially) secret formula- it appears to only marginally take into account level of opponent. Instead, it focuses a lot more on whether you win/lose and by what margin- regardless of what team you played.

Obviously, given our weak OOC.... this could actually be beneficial.

Take, for example, last night:
https://www.ncaa.com/rankings/basketbal ... t-rankings

SJU is now #18, up from #32 yesterday. I was surprised to see that, because we beat a horrible opponent (#340 UMES) at home. With RPI the opposite happened and we dropped from #4 to #15 because our opponent was so bad. I noticed yesterday too that other undefeated teams with weak schedules (such as NC State/Pitt) were extremely overvalued in NET compared to other rankings. To me and a lot of others, the system seems flawed in this regard.

The best part about this is we already played the tougher half of our OOC schedule. If we can eat up cupcakes during December without dropping in the rankings (like what will happen with RPI), that'd be pretty amazing.

Now, to be honest it wouldn't surprise me if the NCAA adjusted the numbers to more heavily weigh in level of opponent. They are keeping some parts of the formula a secret, so they could change it gradually and we wouldn't notice. At least for now though, things may be looking up.

I think one thing that happened to help you out was Temple got a big win at Missouri. Just looking with that- VCU- up from 51 to 43. Cal- up from 240 to 230. So that would make some difference.... Also, teams that were just ahead of you lost. Pitt/NC State come to mind.

Saw that back on Monday- the top team with no road/neutral wins was all the way down at #18. So this system definitely values those wins. Given Temple was neutral site for both VCU/Cal- that would be big...

[adoraz]

I posted this over at Redmen, any thoughts?

As you may know, the NCAA released their first NET ratings a few days ago. They have received a lot of criticism, however the system could work in our favor.

There is something pretty interesting with the (partially) secret formula- it appears to only marginally take into account level of opponent. Instead, it focuses a lot more on whether you win/lose and by what margin- regardless of what team you played.

Obviously, given our weak OOC.... this could actually be beneficial.

Take, for example, last night:
https://www.ncaa.com/rankings/basketbal ... t-rankings

SJU is now #18, up from #32 yesterday. I was surprised to see that, because we beat a horrible opponent (#340 UMES) at home. With RPI the opposite happened and we dropped from #4 to #15 because our opponent was so bad. I noticed yesterday too that other undefeated teams with weak schedules (such as NC State/Pitt) were extremely overvalued in NET compared to other rankings. To me and a lot of others, the system seems flawed in this regard.

The best part about this is we already played the tougher half of our OOC schedule. If we can eat up cupcakes during December without dropping in the rankings (like what will happen with RPI), that'd be pretty amazing.

Now, to be honest it wouldn't surprise me if the NCAA adjusted the numbers to more heavily weigh in level of opponent. They are keeping some parts of the formula a secret, so they could change it gradually and we wouldn't notice. At least for now though, things may be looking up.

Looking at it daily is probably going to be an unhealthy practice for me.

Creighton jumped up 10 spots in a day they didn't play and one of their previous opponents lost.

Yeah I agree that looking daily doesn't tell the whole story. However, I (and others) noticed the trend of level of opponent mattering significantly less than it did for RPI. Could be a good thing for teams like SJU (6-0, currently #18) and a bad thing for teams like Hall (4-2, currently #86).

[stever20]

I posted this over at Redmen, any thoughts?

As you may know, the NCAA released their first NET ratings a few days ago. They have received a lot of criticism, however the system could work in our favor.

There is something pretty interesting with the (partially) secret formula- it appears to only marginally take into account level of opponent. Instead, it focuses a lot more on whether you win/lose and by what margin- regardless of what team you played.

Obviously, given our weak OOC.... this could actually be beneficial.

Take, for example, last night:
https://www.ncaa.com/rankings/basketbal ... t-rankings

SJU is now #18, up from #32 yesterday. I was surprised to see that, because we beat a horrible opponent (#340 UMES) at home. With RPI the opposite happened and we dropped from #4 to #15 because our opponent was so bad. I noticed yesterday too that other undefeated teams with weak schedules (such as NC State/Pitt) were extremely overvalued in NET compared to other rankings. To me and a lot of others, the system seems flawed in this regard.

The best part about this is we already played the tougher half of our OOC schedule. If we can eat up cupcakes during December without dropping in the rankings (like what will happen with RPI), that'd be pretty amazing.

Now, to be honest it wouldn't surprise me if the NCAA adjusted the numbers to more heavily weigh in level of opponent. They are keeping some parts of the formula a secret, so they could change it gradually and we wouldn't notice. At least for now though, things may be looking up.

Looking at it daily is probably going to be an unhealthy practice for me.

Creighton jumped up 10 spots in a day they didn't play and one of their previous opponents lost.

The Clemson loss was on Monday.....

some teams right in front of them lost which was a big thing I think....

[stever20]

the other thing that I'd say is we can't see how close the ratings are.... sometimes RPI ratings were like between 31-34 only .0029 separating them.

[Fieldhouse Flyer]

No. 1 Ohio State? NCAA's initial NET basketball ratings raise questions - Myron Medcalf, ESPN - November 26, 2018

On last year's team sheets, rankings from KenPom.com, Jeff Sagarin, ESPN's Strength of Record, the KPI and ESPN's BPI were mentioned. They'll be on this year's team sheets for committee members, too. Just like the RPI, the NET is just one component with an impact on what happens on Selection Sunday.

But if Monday's unveiling is an indicator of future results, the NCAA's attempt to rectify a situation with the NET could backfire.

......

Overreaction to the NCAA's new NET rankings, timeline on Silvio De Sousa and the worst injury so far – Matt Norlander, CBS Sports – 5 hours ago

Monday's big unveiling of the NET rankings brought a lot of backlash -- probably more than even the NCAA was expecting.

The NCAA wants the rankings out now so people can react, discuss, critique -- and the powers-that-be affiliated with the NET can learn from it. Everyone associated with the NET knew the debut would have outliers attached to it. These rankings were not, and won't be for at least a month, expected to look close to KenPom.com, the Sagarin Ratings, KPI, BPI and others.

"In any rankings system when you've played less than 20 percent of the season, you're going to have some head-scratchers," Gavitt said. "This is a year-long education process. I think we feel comfortable with the decisions to date and we'll continue to evaluate it. This is very much a forward-looking analytic. We're trying to track all of this and evaluate it for the future as we go along. We are not going to make any changes to the NET during the season. It wouldn't be until after the season, after selections, that we consider any tweaks, and don't anticipate those right now, but will certainly evaluate it."

I asked Gavitt if there was consideration to publishing the final ranking for 2017-18 in the NET, which the NCAA has on hand. He said that specific discussion hasn't been had but he would take it into consideration. If it's done, it could ease the fever around the NET.

"There's a lot of questions about transparency and sharing information," Gavitt said. "This is what it's all about. There's certainly, just like the RPI, there's not enough games for it all to make sense [in late November], but not starting to share it would potentially lead people to make up things in their mind that are just not accurate, like there's something to hide here or we're not being as open as we've tried to be about the committee's resources."

[Fieldhouse Flyer]

Predictably hilarious, but that was expected this early in the season.

1. Ohio State
10. Loyola Marymount
12. Belmont
[92. Syracuse]

https://www.cbssports.com/college-basketball/news/ncaas-new-net-rankings-debut-ohio-state-shockingly-no-1-with-duke-no-6-and-kentucky-no-61/

Syracuse 72, Ohio State 62 –AP/ESPN Recap – November 28, 2019

COLUMBUS, Ohio -- The Orange shot nearly 46 percent from beyond the arc as Syracuse upset AP No. 16 Ohio State 72-62 in the Big Ten/ACC Challenge on Wednesday night.

Given the early-season volatility of the NET, it will be interesting to see where Ohio State and Syracuse land in next week’s NET rankings (and AP Poll).

[xusandy]

I've come to believe that both the KenPom and Sagarin ratings are pretty accurate, and by the end of the season don't differ by more than a few spots. To me, the NET league rankings look pretty reasonable, though there are a lot of individual rankings that seem way off (OMG, XU is not in the top 100 in the NET rankings, and the Friars are right there with us?!)

But the accuracy of the NET can't be judged after 20% of the games, with relatively few away games by P6 schools, and no tough league games to factor into the metric. I suspect that be the end of the season, the NET, KenPom, and Sagarin ratings will converge. Big deal, in another 3-5 years we'll be talking about how to schedule to maximize your NET rating, how many points you need to beat level 1 or level 2 cupcakes by, how many defensive stops you make with exactly 3 of your starters on the florr, etc. etc. ad nauseum.