Something is not right with the NET

[shiverz45]

SC is 9-1 against quadrant 1 opponents. Stanford is 5-2 and UCONN is 2-1 yet both are still ranked higher in the NET. SC is number 1 in Massey and RPI. I don't know what is being factored in to yield this result but whatever it is weighs too much in my opinion.

 

[Gamecockfan77]

SC is 9-1 against quadrant 1 opponents. Stanford is 5-2 and UCONN is 2-1 yet both are still ranked higher in the NET. SC is number 1 in Massey and RPI. I don't know what is being factored in to yield this result but whatever it is weighs too much in my opinion.

It'll work itself out.

 

[shiverz45]

It'll work itself out.

I really don't care if we are ranked #1 or not. I am just curious about the statistics they are using to yield that result. I know it is just one factor to determining seeding so it really doesn't matter that much.

 

[LETTERL]

I really don't care if we are ranked #1 or not. I am just curious about the statistics they are using to yield that result. I know it is just one factor to determining seeding so it really doesn't matter that much.

I'm curious about this, too. I mean, I know what they are using to calculate it but some of the rankings just don't make sense.

 

[Plebe]

SC is 9-1 against quadrant 1 opponents. Stanford is 5-2 and UCONN is 2-1 yet both are still ranked higher in the NET. SC is number 1 in Massey and RPI. I don't know what is being factored in to yield this result but whatever it is weighs too much in my opinion.

Oh dear, how will we justify raising another fake banner if the NET doesn't have us at number 1.

 

[CocoHusky]

SC is 9-1 against quadrant 1 opponents. Stanford is 5-2 and UCONN is 2-1 yet both are still ranked higher in the NET. SC is number 1 in Massey and RPI. I don't know what is being factored in to yield this result but whatever it is weighs too much in my opinion.

I forgot to mention I had a meeting with the people that run the NET.

 

[visitingcock]

1	2	Stanford	Pac-12	13-2	9-1	0-0	4-1	0-0
2	1	UConn	Big East	10-1	3-1	0-0	7-0	0-0
3	3	South Carolina	SEC	13-1	6-0	2-0	5-1	0-0
4	4	Oregon	Pac-12	11-3	5-2	0-0	6-1	0-0
5	5	Baylor	Big 12	11-2	4-1	0-0	7-1	0-0
6	6	Louisville	ACC	16-0	4-0	1-0	11-0	0-0
7	7	UCLA	Pac-12	9-2	4-1	0-0	5-1	0-0
8	10	Maryland	Big Ten	12-2	5-1	2-1	5-0	0-0
9	11	Ohio St.	Big Ten	10-1	3-1	0-0	7-0	0-0
10	8	Indiana	Big Ten	9-4	4-2	0-0	5-2	0-0
11	9	NC State	ACC	11-1	3-1	0-0	8-0	0-0
12	12	Michigan	Big Ten	10-1	4-1	0-0	6-0	0-0
13	15	Arizona	Pac-12	11-2	2-1	0-0	9-1	0-0
14	13	Tennessee	SEC	11-3	3-1	0-0	8-2	0-0
15	17	Texas A&M	SEC	15-1	7-1	0-0	8-0	0-0
16	16	South Fla.	AAC	10-1	4-0	0-0	6-1	0-0
17	14	Georgia	SEC	13-3	5-1	0-0	8-2	0-0
18	18	Gonzaga	WCC	14-2	7-1	3-1	4-0	0-0
19	19	Rutgers	Big Ten	5-3	1-2	0-0	4-1	0-0
20	20	Kentucky	SEC	12-4	3-3	0-0	9-1	0-0
21	22	Georgia Tech	ACC	9-3	4-1	0-0	5-2	0-0
22	21	West Virginia	Big 12	12-2	2-1	2-0	8-1	0-0
23	23	Missouri St.	MVC	8-2	4-0	1-1	3-1	1-0
24	25	Arkansas	SEC	12-6	3-4	1-1	8-1	0-0
25	24	SFA	Southland	13-2	5-2	2-0	6-0	2-0

 

[visitingcock]

I really don't care if we are ranked #1 or not. I am just curious about the statistics they are using to yield that result. I know it is just one factor to determining seeding so it really doesn't matter that much.

I'm more curious about Oregon at number 4.

 

[Dream Jobbed 2.0]

Sorry looks like you’re gonna get a #1 in the NET 2021 Rankings banner

 

[visitingcock]

Sorry looks like you’re gonna get a #1 in the NET 2021 Rankings banner

Has the last game of the season been played? Dadgummit, Nobody told me.

I guess we'll just be Massey Champs this year. Like the Ducks last year.

 

[ConwayGmck]

Yep things may work themselves out, but yesterday was a good example of your issues: this was BEFORE yesterday's results - and CT's loss to Arkansas - was added to the equation, but Stanford and USC was #2 and #3 in the NCAA Evaluation Tool (NET) ratings - they had the higher power ratings than CT, the higher offensive ratings - CT had the higher defensive rating, but ANY offense/defense rating should have been a red light for CT, because while both Stanford and USC had top 10 SOS ratings, CT's SOS rating wasn't even in the TOP 70(!!!!)

And yet, despite ALL those higher ratings for BOTH Stanford and South Carolina, Connecticut had the #1 overall rating. And note I didn't even touch on your points above that you addressed, as to the overall number of Q1 NET opponents both teams had played against CT's total: they played as many if not MORE Q1 opponents thus far than CT had played against Q1 AND Q2 opponents combined.

It goes back to that mysterious, nefarious, murky nebulous "Team Value Index" (TVI) which was ONCE one of FIVE (5) elemental parts of the equation, but now is supposed to be one of just TWO (2) - starting this season for both men's and women's BB, they removed winning %, adjusted winning %, and scoring margin of victory, which was the holy grail of sorts for CT fans since CT plays in the AAC/Big East. Those metrics are no longer a part of the NET.

The TVI is explained as a component which "rewards" teams for beating quality opponents, especially away from home. But the exact formula of this component has NEVER been released to the public, and apparently never will be, leaving a degree of undisclosed wiggle room in just how the evaluators "evaluate" this aspect. Is it based on an actual mathematical metric, is it based on more "touchy-feely" subjectivity? They ain't telling.

The second metric to the new NET is an adjusted net efficiency rating. The adjusted efficiency is described by the NCAA as "a team’s net efficiency, adjusted for strength of opponent and location (home/away/neutral) across all games played. For example, a given efficiency value (net points per 100 possessions) against stronger opposition rates higher than the same efficiency against lesser opponents and having a certain efficiency on the road rates higher than the same efficiency at home".

I'm guessing that any references to "quality" of opponents, "strength of opponent" and "stronger opposition" refers to either individual team NET ratings or Strength of Schedule (SOS) ratings, since the whole point of the NET was to replace the RPI, so they aren't using RPI as a reference because then what's the point. But if THIS is the case, HOW can they use CT's SOS and NET ratings of their opponents against Stanford and South Carolina, and then rank CT ABOVE both Stanford and South Carolina. That's as effective of a 1+1=3 math as I've ever seen.

So I thought before yesterday, that a HUGE factor to this ghostly TVI must be being undefeated (But if that was the case, then having lost to Arkansas yesterday should have plummetted CT. And BEFORE yesterday, both NC State and Louisville were undefeated and had better cases for being above CT in the ratings, but they weren't). To a degree, I guess it did but not by much - there's STILL other teams with only 1 loss but with better rated SOS and better Q1 NET opponent results than CT has, but are rated beneath them. So I still am wary and suspicious of the NET, and always will be for ANY ranking formula that refuses to openly disclose its rationale for choosing who will be #1, and who will not be.....

 

[ConwayGmck]

Yep things may work themselves out, but yesterday was a good example of your issues: this was BEFORE yesterday's results - and CT's loss to Arkansas - was added to the equation, but Stanford and USC was #2 and #3 in the NCAA Evaluation Tool (NET) ratings - they had the higher power ratings than CT, the higher offensive ratings - CT had the higher defensive rating, but ANY offense/defense rating should have been a red light for CT, because while both Stanford and USC had top 10 SOS ratings, CT's SOS rating wasn't even in the TOP 70(!!!!)

And yet, despite ALL those higher ratings for BOTH Stanford and South Carolina, Connecticut had the #1 overall rating. And note I didn't even touch on your points above that you addressed, as to the overall number of Q1 NET opponents both teams had played against CT's total: they played as many if not MORE Q1 opponents thus far than CT had played against Q1 AND Q2 opponents combined.

I have to correct myself substantially here - because my above comment is royally screwed up. The screw-job stems from my having major issues with BOTH the Massey power ratings AND with the NCAA Evaluation Tool (NET) ratings, and how both jive with themselves based on actual team data.

The premise of this thread was to question Massey's Power ratings. I agreed with that questioning, and the first two paragraphs I posted were based almost entirely on Massey's power ratings, but instead of naming the source in question "Massey's Power Ratings", I named it the NCAA Evaluation Tool (NET) by mistake. I guess having two separate sources of issues running around at the same time in my mind.

So when I'm talking about Stanford and South Carolina being #2 and #3 behind Connecticut yesterday, I meant in Massey's ratings, NOT in the NET rankings. The "power ratings", defensive and offensive ratings, and SOS rankings all refer to Massey's own rating system, not NET.

And THEN after that I started diving into my issues with the NET system. So for those who might have the spare time to actually read my essay-long posts and start getting confused, don't feel bad: I re-read what I posted and got confused too, and that's why I had to explain myself.

As it is, SINCE yesterday's results have been tallied up, Stanford has jumped ahead of Connecticut at #1 in the NET, and South Carolina is still #3. USC is 9-1 versus Q1 NET opponents thus far, while CT is 2-1 versus Q1 NET opponents. Both USC and CT have lost 1 game - USC to a top-10 NET NC State, CT to a top 25 NET Arkansas. USC has played the #2 SOS thus far, while CT has played the #68 SOS. But somehow, CT is #2 in NET and USC is #3.

But since USC's loss was at home, and CT's loss was on the road, I guess a overload of weight is attached to where a team loses as opposed to whom that team loses to, in the NET's equation.....

In Massey's ratings, South Carolina is now #1, Stanford is #2, and CT has dropped to #3. I guess in that system losing game places a hit to a team's power rating, but remaining undefeated sure hasn't helped Louisville - they are now #7 and even dropped a spot from yesterday.....

 

[Alydar]

Yep things may work themselves out, but yesterday was a good example of your issues: this was BEFORE yesterday's results - and CT's loss to Arkansas - was added to the equation, but Stanford and USC was #2 and #3 in the NCAA Evaluation Tool (NET) ratings - they had the higher power ratings than CT, the higher offensive ratings - CT had the higher defensive rating, but ANY offense/defense rating should have been a red light for CT, because while both Stanford and USC had top 10 SOS ratings, CT's SOS rating wasn't even in the TOP 70(!!!!)

And yet, despite ALL those higher ratings for BOTH Stanford and South Carolina, Connecticut had the #1 overall rating. And note I didn't even touch on your points above that you addressed, as to the overall number of Q1 NET opponents both teams had played against CT's total: they played as many if not MORE Q1 opponents thus far than CT had played against Q1 AND Q2 opponents combined.

It goes back to that mysterious, nefarious, murky nebulous "Team Value Index" (TVI) which was ONCE one of FIVE (5) elemental parts of the equation, but now is supposed to be one of just TWO (2) - starting this season for both men's and women's BB, they removed winning %, adjusted winning %, and scoring margin of victory, which was the holy grail of sorts for CT fans since CT plays in the AAC/Big East. Those metrics are no longer a part of the NET.

The TVI is explained as a component which "rewards" teams for beating quality opponents, especially away from home. But the exact formula of this component has NEVER been released to the public, and apparently never will be, leaving a degree of undisclosed wiggle room in just how the evaluators "evaluate" this aspect. Is it based on an actual mathematical metric, is it based on more "touchy-feely" subjectivity? They ain't telling.

The second metric to the new NET is an adjusted net efficiency rating. The adjusted efficiency is described by the NCAA as "a team’s net efficiency, adjusted for strength of opponent and location (home/away/neutral) across all games played. For example, a given efficiency value (net points per 100 possessions) against stronger opposition rates higher than the same efficiency against lesser opponents and having a certain efficiency on the road rates higher than the same efficiency at home".

I'm guessing that any references to "quality" of opponents, "strength of opponent" and "stronger opposition" refers to either individual team NET ratings or Strength of Schedule (SOS) ratings, since the whole point of the NET was to replace the RPI, so they aren't using RPI as a reference because then what's the point. But if THIS is the case, HOW can they use CT's SOS and NET ratings of their opponents against Stanford and South Carolina, and then rank CT ABOVE both Stanford and South Carolina. That's as effective of a 1+1=3 math as I've ever seen.

So I thought before yesterday, that a HUGE factor to this ghostly TVI must be being undefeated (But if that was the case, then having lost to Arkansas yesterday should have plummetted CT. And BEFORE yesterday, both NC State and Louisville were undefeated and had better cases for being above CT in the ratings, but they weren't). To a degree, I guess it did but not by much - there's STILL other teams with only 1 loss but with better rated SOS and better Q1 NET opponent results than CT has, but are rated beneath them. So I still am wary and suspicious of the NET, and always will be for ANY ranking formula that refuses to openly disclose its rationale for choosing who will be #1, and who will not be.....

NET considers MOV, home and road, and looks at a team's offensive and defensive pts/poss. I think it will help performance quality. Too many good teams stroll through three periods against weak conference foes and then use a frantic 4th to come back and win by 4 when they should have won by 20. I'm really not too informed about NET but the quality of a win or a loss is factored in while all RPI cared about was who won.

 

[LETTERL]

NET considers MOV, home and road, and looks at a team's offensive and defensive pts/poss. I think it will help performance quality. Too many good teams stroll through three periods against weak conference foes and then use a frantic 4th to come back and win by 4 when they should have won by 20. I'm really not too informed about NET but the quality of a win or a loss is factored in while all RPI cared about was who won.

Well, phooey! It doesn't matter whether a team wins by 1 or by 25...a win is a win is a win. Period.

 

[ConwayGmck]

NET considers MOV, home and road, and looks at a team's offensive and defensive pts/poss. I think it will help performance quality. Too many good teams stroll through three periods against weak conference foes and then use a frantic 4th to come back and win by 4 when they should have won by 20. I'm really not too informed about NET but the quality of a win or a loss is factored in while all RPI cared about was who won.

They did last year. They do not now:

No longer will the NET use winning percentage, adjusted winning percentage and scoring margin. The change was made after the committee consulted with Google Cloud Professional Services, which worked with the NCAA to develop the original NET.

 

[visitingcock]

They did last year. They do not now:

They have added migratory bird flight patterns, tea leaves and crystal balls

 

[Alydar]

Well, phooey! It doesn't matter whether a team wins by 1 or by 25...a win is a win is a win. Period.

So for you ranking is easy, the team with the most wins is the champ? Everyone else are just losers?

 

[LETTERL]

So for you ranking is easy, the team with the most wins is the champ? Everyone else are just losers?

No, not at all. That's not what I'm saying. There are a lot of variables that go into determining the final score of a basketball game...you just can't look at the final score and say "Well, Team A should have really dominated Team B"...especially in a season when a lot of teams have to pause and get out of rhythm. I suspect that is not being taken into account by these NET rating algorithms. Wins are important, for sure,

It's nice if you can dominate a weak team by 35 or 40 points in a league that doesn't have quality depth...but in the major conferences there are often six or seven teams that can beat anyone on any given night. Because a lot of these teams beat up on each other, the overall records look less than stellar, which affects the computer ratings. Then a win by a team like Virginia Tech, for example, crushes a team like NC State's NET rating, when VaTech could very well have had 4 or 5 more wins if -- oh, never mind. I give up.

It is a complicated issue...trying to figure out exactly who the best teams are. Hopefully, we will have a national tournament this year to settle the issue once and for all.

 

[newfrontier]

They have added migratory bird flight patterns, tea leaves and crystal balls

don't forget their additional use of tarot cards, divining rods, palm reading, astrology, Feng shui, numerology, origami, se'ances, bone casting (tosses), trances, cloud shapes, incantations, UFO sightings, dream interpretation, spiritual treks, Kau Chim, body markings, secret but fair elections, the movement of suspended objects, fire gazing, double blind testing, reflective objects, clarvoyance, exorcism, wire tapping, interpretation of drawings, spirit boards, face reading, hypnosis, animal entrails, confession, dripping wax, alchemy, spinning wheels, ESP, chants, mind reading, rock-paper-scissors, the drawing of straws, "Eeny, meeny, miny, moe", lotteries, coin tossing, short skirts, cleavage, prayer, meditation, musical chairs, leeches, sting operations, rolling dice, water boarding, a few double agents, the "smell test", drawings from a hat, blood tests, star gazing, urinalysis, animal sacrifices, weekly visits to the Oracle of Delphi and above all, wishful thinking.

I mean, these guys are good, really good.

 

[visitingcock]

Does anybody know how the NET is used in the election process?

Is it merely a guide to the selection committee is it mandatory?

After the conference champs are all accounted for, does the selection committee just go down the ranking list until they reach the 64th entrant and stop - with the 65th team being SOL? Or do they have discretion to depart from the ranking list when making selections?

just for fun, I reviewed the RPI rankings. They are wildly different from the NET. Wildly! Glad it's not being used anymore, though it would probably be more legit looking by season's end

Oregon is ranked 41! Here is the RPI Top 25:

​	​	​	​	​	SOS​	SOS Rank​	​
​		1​	South Carolina	13-1​	0.7344​	1​	0.6757​	Sec	8-0​
​		2​	Louisville	16-0​	0.7097​	12​	0.6130​	Acc	9-0​
​	​	3​	Maryland	12-2​	0.6948​	8​	0.6253​	Big10	8-1​
​	​	4​	Missouri St.	9-2​	0.6705​	10​	0.6192​	Mvc	5-0​
​	​	5​	Bucknell	7-0​	0.6699​	64​	0.5599​	Patr	7-0​
​	​	6​	Ohio St.	10-1​	0.6674​	35​	0.5788​	Big10	6-1​
​		7​	Texas A&M	15-1​	0.6658​	50​	0.5676​	Sec	6-1​
​	​	8​	West Virginia	12-2​	0.6645​	14​	0.6102​	Big12	6-2​
​	​	9​	N.C. State	11-1​	0.6590​	52​	0.5662​	Acc	6-1​
​	​	10​	Stanford	14-2​	0.6552​	39​	0.5765​	Pac12	11-2​
​	​	11​	Kentucky	12-4​	0.6551​	9​	0.6234​	Sec	5-3​
​	​	12​	Connecticut	10-1​	0.6500​	69​	0.5556​	Aac	0-0​
​	​	13​	Kent St.	5-2​	0.6401​	25​	0.5891​	Midam	4-0​
​	​	14​	Syracuse	7-3​	0.6401​	24​	0.5915​	Acc	5-3​
​	​	15​	Rice	10-1​	0.6397​	100​	0.5372​	Cusa	6-0​
​	​	16​	Lehigh	6-3​	0.6397​	3​	0.6434​	Patr	6-3​
​	​	17​	Georgia Tech	9-3​	0.6380​	13​	0.6118​	Acc	7-2​
​	​	18​	Fairfield	6-2​	0.6376​	16​	0.6065​	Maac	5-1​
​	​	19​	Alabama	12-3​	0.6325​	30​	0.5836​	Sec	5-3​
​	​	20​	Gonzaga	15-2​	0.6321​	87​	0.5462​	Wcc	10-0​
​	​	21​	Georgia	13-3​	0.6287​	33​	0.5795​	Sec	5-3​
​	​	22​	Arkansas	12-6​	0.6261​	17​	0.6005​	Sec	2-5​
​	​	23​	Iowa St.	11-5​	0.6247​	18​	0.6005​	Big12	7-2​
​	​	24​	UCLA	10-2​	0.6210​	73​	0.5542​	Pac12	8-2​
​	​	25​	Baylor	11-2​	0.6197​	82​	0.5494​	Big12	6-1​