Big East Polls, NET Rankings and Team Sheets

[DoctorV]

Yes and no. This was definitely the case when RPI was the main metric, but now that they are using an efficiency based metric, how much you win or lose by is just about as important. Colorado State didn't miss the tournament last year because they didn't beat enough good teams. They had a higher winning percentage, better Q1+2 record, and no bad losses compared to fellow MWC league mate Utah State.

But USU was ranked #47 in the NET and CSU was #63. If it was just about winning enough games and beating good teams, CSU's RPI of #41 would've got them in ahead of #56 USU, but now it's about minimizing loss margins and winning in blowout fashion that matters more.

Solid info. This seems like an unfortunate consequence of metrics system.

I could look it up but you already know the answer- did SDSU win both regular season and conference tourneys? Why were CSUs metrics so poor in comparison to USU? If I recall they started off really well so I assume they played a really weak OOC schedule and their bye games were against very weak teams so this was the difference?

Since you are obviously a metrics guy I think this is a good question for you-
Do you look at this case and see a flaw in the system? Would you have preferred that someone in that room had stepped up and caught it and CSU was put in before USU or you think the committee got it right and have no problem with it?

[MU82]

Yes and no. This was definitely the case when RPI was the main metric, but now that they are using an efficiency based metric, how much you win or lose by is just about as important. Colorado State didn't miss the tournament last year because they didn't beat enough good teams. They had a higher winning percentage, better Q1+2 record, and no bad losses compared to fellow MWC league mate Utah State.

But USU was ranked #47 in the NET and CSU was #63. If it was just about winning enough games and beating good teams, CSU's RPI of #41 would've got them in ahead of #56 USU, but now it's about minimizing loss margins and winning in blowout fashion that matters more.

I don't like that a coach who's winning by 14 with 1:30 left needs to keep his foot on the gas so he can win by 21. I don't like that a coach who's losing by 15 has to keep his better players in a lost-cause game so it doesn't end up being a 20-point loss. It's a flaw in the system IMHO, and it could get players hurt.

[Herman Cain]

Big East NET Rankings as of games of December 6, 2021
New Old
5   5   Villanova   
11   13   UConn   
23   21   Seton Hall   
31   30   Xavier   
34   37   Providence   
71   78   Marquette   
90   89   Creighton   
96   94   DePaul   
106   108   St. John's
199   197   Butler   
248   247   Georgetown

[brewcity77]

Solid info. This seems like an unfortunate consequence of metrics system.

I could look it up but you already know the answer- did SDSU win both regular season and conference tourneys? Why were CSUs metrics so poor in comparison to USU? If I recall they started off really well so I assume they played a really weak OOC schedule and their bye games were against very weak teams so this was the difference?

Since you are obviously a metrics guy I think this is a good question for you-
Do you look at this case and see a flaw in the system? Would you have preferred that someone in that room had stepped up and caught it and CSU was put in before USU or you think the committee got it right and have no problem with it?

SDSU did win both. There were three main things that jumped out. Looking at comparable events that likely led to their kenpom difference (which is the closest analogue to NET), two things stood out.

Where I see real differences are the average margin of loss (9.8 for USU, 12.5 for CSU) and the average winning margin against their 6 worst opponents (33.3 for USU, 27.7 for CSU). That really jumps out because of those 6, 5 were common opponents, so it's a pretty clear direct comparison that shows blowing teams out by larger margins matters.

My problem with NET is that it is primarily based on efficiency margin differential. On the surface, that's fine, but look at what a couple blowouts can do to a team. Last year, Louisville was projected to lose at Wisconsin and at UNC, but because they lost at Wisconsin by 37 instead of 4 and at UNC by 45 instead of 1, they saw a combined kenpom hit of 29 rank spots between those two games. They lost games they were supposed to lose, but the margins likely cost them 20-30 rank spots and a NCAA bid. Margin should matter, but the problem with a season-long metric is that it views the difference between 1 and 11 points the same way it does 31 and 41 points. There should be a regressive model where the bigger the margin, the less it matters. 1 vs 11 is tangibly different, more so than say 20 vs 40, both commanding, blowout margins.

My second issue with NET is their "cap" of 10 points. They claim they have a maximum game margin difference of 10 points, but when the primary evaluation tool is an efficiency margin that works on margin differential, the 10 point cap is actually a bonus added on top of the existing efficiency margin.

I do think NET is better than RPI. I like advanced metrics and think they have improved the game. But I think there's still room for improvement.

[mubb3434]

Butler moved from 199 to 121 in NET after their win at Oklahoma. What a MASSIVE result for the league.

Marquette moves up 4 spots to 67

Providence moves up to 31 after their solid win vs. a solid Vermont squad

Really good day for the Big East!

[mug644]

Hopefully the Big East can get another 4 wins tonight. Looks like UConn-West Virginia and MU-KSU should be the toughest matchups. Xavier should handle Ball St, but maybe we can count on G'town taking care of business against UMBC.

[Herman Cain]

Big East NET rankings as of games of December 7, 2021
New Old
5   5   Villanova   
10   11   UConn   
25   23   Seton Hall   
31   34   Providence   
33   31   Xavier   
67   71   Marquette   
83   96   DePaul   
87   90   Creighton   
113   106   St. John's
121   199   Butler   
245   248   Georgetown   

[MarquetteMike1977]

Big East NET rankings as of games of December 7, 2021
New Old
5   5   Villanova   
10   11   UConn   
25   23   Seton Hall   
31   34   Providence   
33   31   Xavier   
67   71   Marquette   
83   96   DePaul   
87   90   Creighton   
113   106   St. John's
121   199   Butler   
245   248   Georgetown

Good Stuff Thanks

[mubb3434]

Marquette moves from 67 to 60 in the NET this morning.

Of note, Xavier moved to 20 from 33 & Georgetown moved to 205 from 245.

[MU Fan in Connecticut]

Marquette moves from 67 to 60 in the NET this morning.

Of note, Xavier moved to 20 from 33 & Georgetown moved to 205 from 245.

I was going to say I found it weird that X is 8-1 and beat Ohio State who has two losses yet Ohio State is somehow ranked and X is only receiving votes. B10 bias?

[JamilJaeJamailJrJuan]

I was going to say I found it weird that X is 8-1 and beat Ohio State who has two losses yet Ohio State is somehow ranked and X is only receiving votes. B10 bias?

B10 bias, and OSU beat Duke - and all media has Duke bias.  So yeah.

[Herman Cain]

Big East NET rankings as of games of December 8,2021
New Old
6   5   Villanova   
14   10   UConn
20   33   Xavier   
26   25   Seton Hall   
35   31   Providence   
60   67   Marquette   
79   83   DePaul   
84   87   Creighton
111   113   St. John's
118   121   Butler   
205   245   Georgetown   

[goldeneagle91114]

Big East NET rankings as of games of December 8,2021
New Old
6   5   Villanova   
14   10   UConn
20   33   Xavier   
26   25   Seton Hall   
35   31   Providence   
60   67   Marquette   
79   83   DePaul   
84   87   Creighton
111   113   St. John's
118   121   Butler   
205   245   Georgetown

I'm sorry, i know it's been provided before. But, what is the typical NET ranking (range) for an NCAA at large bid?

[fjm]

I'm sorry, i know it's been provided before. But, what is the typical NET ranking (range) for an NCAA at large bid?

Brew mentioned only twice has a team worse than 60 made the ncaa tourney and both in 2019.

Don't recall where he mentioned it.

[brewcity77]

I'm sorry, i know it's been provided before. But, what is the typical NET ranking (range) for an NCAA at large bid?

Inside the top-40 is generally very safe. 40-50 is good shape for a high-major. 50-60 you better have a really good resume. Past 60, you're probably getting left out unless you have 21+ wins and at least a couple marquee victories.

[The Sultan]

I can't seem to find it now, but at one point there was a RPI forecast where you could predict RPI based on the outcomes of games.  Do they have something for NET that you know of?

[TAMU, Knower of Ball]

I can't seem to find it now, but at one point there was a RPI forecast where you could predict RPI based on the outcomes of games.  Do they have something for NET that you know of?

That was RPI wizard. I don't think there's an net equivalent because to my knowledge the formula for NET has never been released and no one has reverse engineered it

[The Sultan]

That was RPI wizard. I don't think there's an net equivalent because to my knowledge the formula for NET has never been released and no one has reverse engineered it

Gotcha.  I was curious as to what our BE record might need to be to meet the thresholds in brew's post above.  (Assuming a loss to UCLA Saturday.)

[lawdog77]

As an aside, I am trying to find the preseason prognostications of our fellow scoopers. If anyone can link that, it would be greatly appreciated, and I will double my annual scoop donation.

[brewcity77]

I can't seem to find it now, but at one point there was a RPI forecast where you could predict RPI based on the outcomes of games.  Do they have something for NET that you know of?

They don't because the NET formula hasn't been cracked yet. Torvik does have his TeamCast, which has similarities to RPI Forecast, but I haven't tested it for accuracy. It does include some metric predictions, but I'm not sure how useful they are. It essentially gives a game-by-game option and where you'd be seeded based on those results.

That said, I'd be skeptical of it's projections. I went through the simulator with Marquette winning out, so beating UCLA, going 20-0 in the Big East, and winning the BET by beating Providence, UConn, and 'Nova, and while it did have us in the tourney, we were only a 7-seed. But if you want to test it out, here it is:

https://barttorvik.com/teamcast.php?team=Marquette&year=2022

[TAMU, Knower of Ball]

They don't because the NET formula hasn't been cracked yet. Torvik does have his TeamCast, which has similarities to RPI Forecast, but I haven't tested it for accuracy. It does include some metric predictions, but I'm not sure how useful they are. It essentially gives a game-by-game option and where you'd be seeded based on those results.

That said, I'd be skeptical of it's projections. I went through the simulator with Marquette winning out, so beating UCLA, going 20-0 in the Big East, and winning the BET by beating Providence, UConn, and 'Nova, and while it did have us in the tourney, we were only a 7-seed. But if you want to test it out, here it is:

https://barttorvik.com/teamcast.php?team=Marquette&year=2022

Lol. Has there ever been a high major team that went 32-2 and didn't get a 1 seed?

[Shooter McGavin]

Lol. Has there ever been a high major team that went 32-2 and didn't get a 1 seed?

Has! Agreed.

Also it seems the NET adds more weight to wins and losses than Kenpom based on moving up more quickly with the same efficiency data.

[Miss Katie’s]

That said, I'd be skeptical of it's projections. I went through the simulator with Marquette winning out, so beating UCLA, going 20-0 in the Big East, and winning the BET by beating Providence, UConn, and 'Nova, and while it did have us in the tourney, we were only a 7-seed. But if you want to test it out, here it is:

https://barttorvik.com/teamcast.php?team=Marquette&year=2022

Seems legit. 

[Herman Cain]

Big East NET rankings as of games of December 9,2021
New Old
5   6   Villanova   
15   14   UConn   
18   20   Xavier   
23   26   Seton Hall   
36   35   Providence   
60   60   Marquette   
72   79   DePaul   
82   84   Creighton   
94   111   St. John's
120   118   Butler   
206   205   Georgetown

[TheGym]

Big East NET rankings as of games of December 9,2021
New Old
5   6   Villanova   
15   14   UConn   
18   20   Xavier   
23   26   Seton Hall   
36   35   Providence   
60   60   Marquette   
72   79   DePaul   
82   84   Creighton   
94   111   St. John's
120   118   Butler   
206   205   Georgetown

***