PDA

View Full Version : Question about NET rankings and performance metric sites like kenpom.com



CDC84
03-07-2019, 04:59 PM
It still remains a mystery how the NCAA committee is creating the NET rankings. Here are the rankings:

https://www.ncaa.com/rankings/basketball-men/d1/ncaa-mens-basketball-net-rankings

Here are the teams sorted by quad 1 wins, quad 2 wins, etc.

https://herosports.com/ncaa-tournament/college-basketball-net-rankings-ahah

Gonzaga is at #1. Critics who claim they don't deserve a number one seed are constantly referring to their LACK of quad 1 wins compared to other contenders (mainly from BCS leagues).

However, if you go to kenpom.com, the Zags are #2 in the nation behind UVA. Their offensive efficiency is the 3rd best in the Pomeroy era going back to 2002. They are #1 in offense by a MILE. The Zags are now #15 in the nation in defensive efficiency. As we all know, history has shown that being in the top 20 in both areas makes you a national title contender. If you're in the top 10 in both areas, you're a favorite. But with Gonzaga, their offensive numbers are so historically great and off the charts that it gives them wiggle room with their defense. This, IMO, suggests that even though teams like Duke and UVA are in the top 10 in both offense and defense, Gonzaga are equal challengers to these 2 teams.

The NET rankings are really starting to make me wonder that the committee is taking performance metrics sites very seriously, and that to collect a bunch of quad 1 wins or whatever is not as important to the committee as these critics think. Yes, having a ton of quad 1 wins helps. But what is now also very important is how you have PERFORMED against your schedule beyond just wins and losses. In other words, I could see some CBS guy question the chairman about why GU got a #1 over this team or that team when they didn't have nearly as many quad 1 wins. The chairman will respond, "Gonzaga had a number of good quad one wins, but what impressed the committee was how they performed against their schedule. Their offensive efficiency numbers are off the charts and make up for the lack of quad 1 wins that other teams might possess. Their defense is in the top 15. Their performance is that of a #1 seed, and a huge part of the NET isn't just your wins and losses, but how you perform against your schedule."

I could also see the chairman saying this: "Gonzaga also has zero bad losses, and they were missing key players in their 2 losses. Some of the teams that might have more quad 1 wins than Gonzaga have bad losses on their resume. Our new NET system punishes teams more for losing to bad teams."

Any thoughts?? Because I can guarantee this is going to be brought up during the bracketology programs. Some teams like Tennessee, Kentucky, Michigan and Michigan State are going to end up on the 2 line if Gonzaga wins out. Some people will be asking why GU is on the one line and not them.

willandi
03-07-2019, 05:07 PM
It still remains a mystery how the NCAA committee is creating the NET rankings. Here are the rankings:

https://www.ncaa.com/rankings/basketball-men/d1/ncaa-mens-basketball-net-rankings

Here are the teams sorted by quad 1 wins, quad 2 wins, etc.

https://herosports.com/ncaa-tournament/college-basketball-net-rankings-ahah

Gonzaga is at #1. Critics who claim they don't deserve a number one seed are constantly referring to their LACK of quad 1 wins compared to other contenders (mainly from BCS leagues).

However, if you go to kenpom.com, the Zags are #2 in the nation behind UVA. Their offensive efficiency is the 3rd best in the Pomeroy era going back to 2002. They are #1 in offense by a MILE. The Zags are now #15 in the nation in defensive efficiency. As we all know, history has shown that being in the top 20 in both areas makes you a national title contender. If you're in the top 10 in both areas, you're a favorite. But with Gonzaga, their offensive numbers are so historically great and off the charts that it gives them wiggle room with their defense. This, IMO, suggests that even though teams like Duke and UVA are in the top 10 in both offense and defense, Gonzaga are equal challengers to these 2 teams.

The NET rankings are really starting to make me wonder that the committee is taking performance metrics sites very seriously, and that to collect a bunch of quad 1 wins or whatever is not as important to the committee as these critics think. Yes, having a ton of quad 1 wins helps. But what is now also very important is how you have PERFORMED against your schedule beyond just wins and losses. In other words, I could see some CBS guy question the chairman about why GU got a #1 over this team or that team when they didn't have nearly as many quad 1 wins. The chairman will respond, "Gonzaga had a number of good quad one wins, but what impressed the committee was how they performed against their schedule. Their offensive efficiency numbers are off the charts and make up for the lack of quad 1 wins that other teams might possess. Their defense is in the top 15. Their performance is that of a #1 seed, and a huge part of the NET isn't just your wins and losses, but how you perform against your schedule."

I could also see the chairman saying this: "Gonzaga also has zero bad losses, and they were missing key players in their 2 losses. Some of the teams that might have more quad 1 wins than Gonzaga have bad losses on their resume. Our new NET system punishes teams more for losing to bad teams."

Any thoughts?? Because I can guarantee this is going to be brought up during the bracketology programs. Some teams like Tennessee, Kentucky, Michigan and Michigan State are going to end up on the 2 line if Gonzaga wins out. Some people will be asking why GU is on the one line and not them.

Personally, I think the quad stuff is another way for the committee to reward P 6 teams. All the teams in the power conferences play more teams likely to be quad 1 or 2.

It's just another BS way to shaft the small conferences.

CDC84
03-07-2019, 06:41 PM
We'll see how the committee treats GU, but they had them as a one in that initial sweet 16. I just don't see many respected bracketogists who don't have Gonzaga as a #1 seed. I read a Pat Forde article where he said that they are not just a shoe in to be a one seed (based on their margin of victory over the WCC's top teams), but that GU may not be the 4th overall number 1 seed as well. I really wonder if some of these bracketologists and many of the analysts who seriously monitor college basketball are getting the hunch that by Gonzaga standing firm at #1 in the NET rankings despite their low level of quad 1 wins that the committee is seriously considering and valuing kenpom.com and noticing where the Zags rank at that site on offense, defense, and overall nationally. Also, Gonzaga just doing things like absolutely annihilating a "good" BYU team by 30 in one of the toughest arenas in the nation. That's 1991 UNLV type stuff. If you look at the history of college ball, when top teams typically beat opponents by 30 points, most of the time it is because the team is bad and has a losing record. Gonzaga is repeatedly beating league teams by 30 who possess well above .500 records.

On a side note....the most interesting case to follow on Selection Sunday will be if Wofford wins the Southern Tourney. They just got done sweeping a league with 3 other, high quality non-BCS teams in East Tennessee, Furman and UNC-Greensboro who may not make the NCAA's but are just as good as some of the BCS teams who will be getting at large bids (although some people think the Southern could get two teams if Wofford wins the tourney....it just depends how it shakes out.). If Wofford wins the Southern, which would mean completely sweeping a high level non-BCS league, I'm telling ya, if they land anything worse than a 7 seed, the committee should ashamed of themselves as they are punishing highly seeded teams. I could see those guys putting them in the 8/9 game. No number one seed wants to play one of the greatest 3 point shooting teams in the past 20 years of college basketball. Those guys will pull up at 35 feet and just sting you. The three point shot is the single greatest equalizer when it comes to a "less talented" team beating an elite number one seed the opening weekend.

LTownZag
03-07-2019, 08:11 PM
As an aside regarding our #15 Kenpom Defense rank, I have a hard time believing we are worse than Houston, VCU, Clemson, Florida, Wisconsin, and the now defenestrated Michigan State. I think our proper defense with Crandall playing is somewhere around #10.

NEC26
03-07-2019, 08:44 PM
Personally, I think the quad stuff is another way for the committee to reward P 6 teams. All the teams in the power conferences play more teams likely to be quad 1 or 2.

It's just another BS way to shaft the small conferences.

The more metrics you add to the equation the more the committee can pick and choose which ones they can use to justify picking some sub .500 conference record team that has lost the majority of their tough games but doggone it they played a lot of good teams. It gets to the point that playing tough teams means more than ever winning.

Reborn
03-07-2019, 08:53 PM
IMO, I believe that Kenpom ranking system is the very best ranking system and the best way to sort out where teams belong in the NCAA Tournament. I have some differences when he gets past #10, but it's just my eye test (for example I think LSU and Nevada should be ranked higher and St Mary's lower. If the committee goes by Kenpom St Mary's will be in. I just can't see them in the tournament based on the two games I saw them against Gonzaga.

The Zags, most likely, will be a number # 1 seed in the West. I am hoping we get placed in San Jose.

Go Zags!!!

LTownZag
03-07-2019, 08:59 PM
Any thoughts?? Because I can guarantee this is going to be brought up during the bracketology programs. Some teams like Tennessee, Kentucky, Michigan and Michigan State are going to end up on the 2 line if Gonzaga wins out. Some people will be asking why GU is on the one line and not them.


I think the 4 "quads" of wins is a step in the right direction, but it's still a very blunt and coarse-grained response to determining game quality. For example, playing a home game against VA, DUKE, or VA is worth the same (a Q1 game) as playing a home game against UCF, Marquette, or Utah State.

It's arbitrary to have these random "steps" or jumps, where a win against the 30th team means a lot, but a win against the 31st means so much less.

Why not have 10 deciles or 100 "per-cents" instead of 4 quadrants?

It's of particular harm to GU since several opponents ended up just outside the desired quadrant.

BYU is 1 spot away from having been a Q1 win (@ BYU)
UW is 3 spots away from having been a Q1 win (@Spokane)
SMC is 9 spots away from having been a Q1 Win (@ Spokane)

Beating the 74th or 75th team in a road game is a Q1 win. Beating BYU (#76) in Provo isn't.

amaronizag
03-07-2019, 09:28 PM
"It's arbitrary to have these random "steps" or jumps, where a win against the 30th team means a lot, but a win against the 31st means so much less."

Well stated LTown
If you're going to rank teams (#1, #2, #3, etc), then you should use the rank # to value team wins rather than quadrants. If you beat the number #1 team in an away game, you should get 1 point. If you beat the #1 team in a home game you should get maybe 1.2 points. If you beat the #10 ranked team in an away game you get 10 points. If you beat the #10 team in a home game you should get 12 points. If you beat the #300 ranked team you get 300 points, etc. Total the points at the end of the year and the teams with the fewest points have the highest rankings. Rank teams according to points earned.

I'm not suggesting that is the ONLY method that should be used for ranking teams, but it could replace the quadrant part of the NET system. I favor using metric based systems like KenPom, and other parts of the net system, but think the quad system stinks.

willandi
03-07-2019, 10:50 PM
The more metrics you add to the equation the more the committee can pick and choose which ones they can use to justify picking some sub .500 conference record team that has lost the majority of their tough games but doggone it they played a lot of good teams. It gets to the point that playing tough teams means more than ever winning.

My quote from the SMC thread:

Forget quad 1 etc. That is just set up because th P 6 schools have more quad chances. It really is aimed at eliminating the mid major schools.

Let each conference send their choice, regular season or tournament, but with no names visible.
Eliminate all teams that are lower than .500 in conference record.
insert the NEXT 37 teams (I think that is right) based on the analytics.
Seed all the 68 teams based on the analytics
Reveal the names of the teams.

Let the chips fall where they may. Every team in every conference would know what they have to accomplish to get to the NCAA. They would have to get the conference auto bid OR they would have to have a winning record in conference and have a high enough metric to make the cut.

Period.

webspinnre
03-08-2019, 08:18 AM
I'm definitely on board with eliminating any team that doesn't finish .500 or higher in their conference.

I think the quadrant setup with NET is a significant improvement over the old RPI approach. Kenpom (and Sagarin, etc) are still better, but at least the quadrant approach using a analytic rating system isn't completely useless like RPI was.

bartruff1
03-08-2019, 08:28 AM
I would like to see both the Conference Champions and the Conference Tournament Champions get a bid...

I don't care about the best 36 teams getting a at large ...I would reward the Conference season champions...

I imagine the Committee will use all comparable data not just the NET ratings to look for strong correlations....

strikenowhere
03-08-2019, 08:39 AM
I would like to see both the Conference Champions and the Conference Tournament Champions get a bid...

I don't care about the best 36 teams getting a at large ...I would reward the Conference season champions...

I imagine the Committee will use all comparable data not just the NET ratings to look for strong correlations....

This seems to create a complete disincentive for regular season champions to even participate in the conference championship, especially for mid-major/small major conferences. With the P5(6?) they are at least playing for seeding.

willandi
03-08-2019, 08:42 AM
I would like to see both the Conference Champions and the Conference Tournament Champions get a bid...

I don't care about the best 36 teams getting a at large ...I would reward the Conference season champions...

I imagine the Committee will use all comparable data not just the NET ratings to look for strong correlations....

I wouldn't mind that at all either.

I think that MOST of the people that watch do so to see their own school, especially the smaller schools, compete and not to watch the same blue bloods year after year after year.

It won't happen though because it would limit the P 6 schools to only getting an extra 6-10 teams in, depending on tournament upsets (unless the gave the regular season champ a bye into the NCAA's).

It surprises me, at times, that more States, and Congress, haven't gotten involved. The NCAA is funneling MILLIONS of dollars, every year, primarily to the same small subset of conferences and schools, rather than spreading the money out. That, in part, is also what causes some of the disparity throughout all levels of NCAA athletics.

mgadfly
03-08-2019, 08:48 AM
If GU wins the conference tourney, they get a 1-seed. That is the most visible treatment of mid-major conferences, the one that every college basketball fan, no matter how casual, will judge the committee's treatment of non major conferences. The real sleight of hand comes on the seeding of the rest of the non major schools and the elimination of teams on the bubble.

GU is going to take some money from major conference schools, and everyone is fine with that at this point. It's allowing any of the other teams and conference to take money out of the power conferences pockets that isn't going to happen. The non-majors will end up with 4 or fewer at-large bids, and somehow, those 2 of those 4 will be matched up against other non-majors. That way, barring low majors beating majors near the majors campus, there will only be a few teams getting shares for conferences other than the top 6 in the round of 32 and Sweet 16. And the committee will defend its theft by pointing to its treatment of GU.

bartruff1
03-08-2019, 08:54 AM
This seems to create a complete disincentive for regular season champions to even participate in the conference championship, especially for mid-major/small major conferences. With the P5(6?) they are at least playing for seeding.

I don't know why they would do that...they would still be competing for seeding and venue …..and pride...

Gonzaga is in and I suspect they are competing in the WCC Tournament ...that is probably true of all of the top ten teams...

willandi
03-08-2019, 09:04 AM
I don't know why they would do that...they would still be competing for seeding and venue …..and pride...

Gonzaga is in and I suspect they are competing in the WCC Tournament ...that is probably true of all of the top ten teams...

If Gonzaga knew that they are in and will get a 1 seed, by virtue of the previous games, they might sit out so a second WCC team would win the conference tournament.

The REAL drawback would be that additional week and 1/2 without games.

jazzdelmar
03-08-2019, 09:05 AM
GU will be fine.....but NEVER, NEVER expect a BCS money grubbing chair or committee member to speak up for us. If anyone would, it would be CBS. Feel for the East Tennessee, Furman and UNC-Greensboros of the world. Gonna get screwed so some 6-12 lame P5 gets in, if only to earn first rd $$$$ for said money grubbing chair or committee member.

NEC26
03-08-2019, 09:08 AM
I'm definitely on board with eliminating any team that doesn't finish .500 or higher in their conference.

I think the quadrant setup with NET is a significant improvement over the old RPI approach. Kenpom (and Sagarin, etc) are still better, but at least the quadrant approach using a analytic rating system isn't completely useless like RPI was.

I agree that that NET looks like a better system. So why not eliminate the RPI altogether. Too many metrics just muddies the waters and allows them the ability to pick whatever metric helps their cause.
You see this every year with somebody justifying one team then turning around and ignoring that metric for another. If you look at how many power 6 conference teams get in it is much higher in the last 20 years.

To the point nearly no smaller conference schools have a chance at being selected if they don't win their conference tournament. Teams that can't win at a .500 clip in their conference have not earned a spot in the NCAA tournament PERIOD!

Mr Vulture
03-08-2019, 09:19 AM
GU will be fine.....but NEVER, NEVER expect a BCS money grubbing chair or committee member to speak up for us. If anyone would, it would be CBS. Feel for the East Tennessee, Furman and UNC-Greensboros of the world. Gonna get screwed so some 6-12 lame P5 gets in, if only to earn first rd $$$$ for said money grubbing chair or committee member.

Truth! Preach it Jazz... :agreed:

jazzdelmar
03-08-2019, 09:24 AM
Truth! Preach it Jazz... :agreed:

Can I get an amen? :lmao: We should do a podcast, we should. It'll be about more than nothing....

bartruff1
03-08-2019, 09:39 AM
If Gonzaga knew that they are in and will get a 1 seed, by virtue of the previous games, they might sit out so a second WCC team would win the conference tournament.

The REAL drawback would be that additional week and 1/2 without games.

No one knows....not even the shadow ….if they are going to get a number 1 seed until the all the Conference Tournament are over...in addition there are likely legal obligations to the Conference... to the sponsors …. the venue...and revenue considerations …..I cannot imagine under any likely circumstances that a Conference Campions would set out the Conference Tournament…..nor do I believe for a single second that they would throw the tournament in order to get two bids...... as I am sure some in here would think... :)

mgadfly
03-08-2019, 09:52 AM
No one knows....not even the shadow ….if they are going to get a number 1 seed until the all the Conference Tournament are over...in addition there are likely legal obligations to the Conference and revenue considerations …..I cannot imagine under any likely circumstances that a Conference Campions would set out the Conference Tournament…..nor do I believe for a single second that they would throw the tournament in older to get two bids as I am sure some in here would think... :)

I agree that no one would be throwing games. But imagine if you are Few and Perkins rolled his ankle in the semifinals on Monday night. He says he is ready to go, but ... Do you sit Tillie, Perkins and other banged up players in an abundance of caution and do your best to win with the rest of the guys? If you win, you keep the #1 seed. If you lose, you potentially earn your conference millions of dollars in additional TV revenue. While I agree with you, I can certainly see the concern in certain situations.

realtydog
03-08-2019, 09:56 AM
Can I get an amen? :lmao: We should do a podcast, we should. It'll be about more than nothing....

1 seed 2 seed 3 seed 4......who gives a crap

if you think this is a national championship team (at least final four) does it matter ? tournament success seems to rely on three things---being a really good team to start the tourney----then getting "hot" at the right time---and lastly (and maybe most important) some good old fashioned LUCK

you need any two of the three to advance in a game it seems---sometimes GU has been the better team --who was not "hot" but...had some luck.....etc.

who cares if GU plays a #10 or #7 in the second round rather than a #8 or #9 -------the #10 seed may be hotter or playing better than the other 3...etc and so on

webspinnre
03-08-2019, 10:19 AM
I agree that that NET looks like a better system. So why not eliminate the RPI altogether. Too many metrics just muddies the waters and allows them the ability to pick whatever metric helps their cause.
You see this every year with somebody justifying one team then turning around and ignoring that metric for another. If you look at how many power 6 conference teams get in it is much higher in the last 20 years.

To the point nearly no smaller conference schools have a chance at being selected if they don't win their conference tournament. Teams that can't win at a .500 clip in their conference have not earned a spot in the NCAA tournament PERIOD!

As far as I know, the RPI is dead, which is great news. That was a worthless system. NET isn't great, but it's at least somewhat useful.

CDC84
03-08-2019, 11:01 AM
I'm definitely on board with eliminating any team that doesn't finish .500 or higher in their conference.

I think the quadrant setup with NET is a significant improvement over the old RPI approach. Kenpom (and Sagarin, etc) are still better, but at least the quadrant approach using a analytic rating system isn't completely useless like RPI was.

There is a lot of talk that this could very well occur with Indiana, because although their league record is crappy, they have a lot of big time wins. It will be interesting if the committee puts them in. But they will probably need to do some good work in the Big ten tourney. Also, many prominent analysts believe that this year's Big ten is the most competitively balanced power conference league in the history of college basketball. There is no team that sucks. A team like Michigan can go into Penn State and get annihilated.

JPtheBeasta
03-08-2019, 11:31 AM
Coach Few and the players value winning the conference championship. Coach Few has said that he sets 4 goals: win their preseason tournament (eg Maui Invitational), win the conference regular season title, win the conference championship, and play well in March.

jazzdelmar
03-08-2019, 12:53 PM
Coach Few and the players value winning the conference championship. Coach Few has said that he sets 4 goals: win their preseason tournament (eg Maui Invitational), win the conference regular season title, win the conference championship, and play well in March.

I think "play well in March" has been supplanted by "win the NC;" and I think the 4 goals are far from equal. New day in Zagland, JP.

Zagceo
03-08-2019, 12:55 PM
4 NBA players on roster...its NATTY or BUST

willandi
03-08-2019, 01:35 PM
1 seed 2 seed 3 seed 4......who gives a crap

if you think this is a national championship team (at least final four) does it matter ? tournament success seems to rely on three things---being a really good team to start the tourney----then getting "hot" at the right time---and lastly (and maybe most important) some good old fashioned LUCK

you need any two of the three to advance in a game it seems---sometimes GU has been the better team --who was not "hot" but...had some luck.....etc.

who cares if GU plays a #10 or #7 in the second round rather than a #8 or #9 -------the #10 seed may be hotter or playing better than the other 3...etc and so on

Forgot #4. Quality and consistent officiating.

maynard g krebs
03-08-2019, 01:37 PM
As an aside regarding our #15 Kenpom Defense rank, I have a hard time believing we are worse than Houston, VCU, Clemson, Florida, Wisconsin, and the now defenestrated Michigan State. I think our proper defense with Crandall playing is somewhere around #10.

These numbers are so closely bunched that it's pretty much moot. Zags' adjusted number is 92.7 pts per 100 possessions; #10 is at 91.5, so we're talking about less than a point per game.

CDC84
03-08-2019, 01:38 PM
4 NBA players on roster...its NATTY or BUST

And more experience (NCAA tournament or otherwise) than Duke.....

I said this the other day, but the one thing I am certain of about this team is that they are far less prone to an opening weekend upset than the other two other number one seeded teams GU has had. They just have too much talent, and the offense is too elite. I suppose a team like Wofford could get a 9 seed and sting GU for 16 threes, but if the NCAA committee knows what it is doing (which they sometimes don't), Wofford shouldn't be getting a 9 seed.

maynard g krebs
03-08-2019, 01:41 PM
I would like to see both the Conference Champions and the Conference Tournament Champions get a bid...



Problem w/ that, imo, is that it would lead to "upsets" in the 1 bid leagues being almost universal. Almost all the 1 bid leagues would mysteriously wind up getting two.

bartruff1
03-08-2019, 03:37 PM
There have been dozens of opportunities tor teams to have done that in the past under the current system...including Gonzaga and Nevada this year....but they won't throw a game...…and I doubt anyone else has... I have a different projection on the world...

maynard g krebs
03-09-2019, 04:49 PM
There have been dozens of opportunities tor teams to have done that in the past under the current system...including Gonzaga and Nevada this year....but they won't throw a game...…and I doubt anyone else has... I have a different projection on the world...

I wasn't talking about teams throwing games. I was talking about officiating, and have observed too many orchestrated results to stay in denial about it. Sadly, I live in a world where Enron and Monsanto do actually exist, so I don't have the luxury of a "projection on the world".

Pleasant Peninsula
03-09-2019, 05:36 PM
I ... have observed too many orchestrated results to stay in denial about it.

Do tell. Don't keep this damning evidence about the orchestration of results to yourself! Share it with us, so we can a) learn from it and/or b) peer review your findings.

maynard g krebs
03-10-2019, 04:19 PM
Do tell. Don't keep this damning evidence about the orchestration of results to yourself! Share it with us, so we can a) learn from it and/or b) peer review your findings.

All I can do here is to quote Japanese writer Haruki Murakama: "if you can't understand it without an explanation, you can't understand it with an explanation".

Or from Eldard Hubbard: "never explain. Your friends do not need it, and your enemies will not believe you anyway."

MontanaCoyote
03-10-2019, 06:13 PM
All I can do here is to quote Japanese writer Haruki Murakama: "if you can't understand it without an explanation, you can't understand it with an explanation".

Or from Eldard Hubbard: "never explain. Your friends do not need it, and your enemies will not believe you anyway."

We have a like saying in politics; “If you’re explaining, you’re losing.”

sheps001
03-10-2019, 09:01 PM
I would be interested to see what is the correlation between Ken Pom and actual performance in the tourney. I remember we were his no1 in the year when we went to the finals and should have won but for Goss turning his ankle and the refs taking us out of the game but what do you expect, Duke and UNC always get the calls when its against teams like us. Also too Karnowski got poked in the eye the game before and was seeing double or triple during the final game, thats why he was missing shots and free throws. Just wait, though, we have a top five recruiting class coming in next year and loaded with top big men. If they can get a decent shooting guard who is good at 3's and a very good point guard, either through recruiting or transfer then we definitely have a shot at the finals again.