ACC NET rankings 1/28

[CorduroyG]

We beat pitt 2x and this is the thanks we get

UNC: 7 (-1)
DUKE: 18 (+1)
CLEM: 28 (+3)
VT: 46 (+8)
WAKE: 47 (-1)
UVA: 48 (+7)
MIA: 63 (0)
PITT: 66 (0)
SYR: 76 (+3)
NCST: 83 (-6)
FSU: 86 (-4)
BC: 90 (-3)
GT: 134 (-3)
ND: 171 (-4)
LOU: 231 (-7)

 

[Nicknack]

Something obviously not right about that.

 

[IthacaMatt]

We have no bad losses. Just by common sense we should be ahead of Pitt and Miami. It's not like either of them are absolutely tearing it up; we have similar records, but we've been both 3 out of 3.

 

[jncuse]

2 Comments

1. We do have a "bad loss" as of now. Florida St at home was a Q3 loss - also decent chances though that works its way up to Q2 which is nice. (Florida St needs to get in top 75 - they are being dragged down by being not very good in OOC)

2. Hate the NET formula or not its absolutely beneficial to us that Pitt stays in the top 75, while we continue to win (hopefully) and improve our resume. Its a good thing that they are 66 and not 76. (You can see all the why's being the 66 in the other thread) Not saying its a fair ranking, but its good for us that they stay in this range and keep winning.

 

[dasher]

I brought this up last week. I don't care about the Net rankings. But Pitt being ahead of us shows flaws in the system to me.

 

[jncuse]

NCST rose 6 spots after being noncompetitive to #73 on the road? Such a trash calculation

They fell 6 spots.
The minus means you went down

 

[CbusCuse]

The NET system obviously LARGELY factors in how much you win and lose by. Which we all know we have gotten killed when we have lost. It seems like a good tool to use for gambling purposes regarding the spread. But not necessarily a good predictor of W/L. Without looking at specific scores of every game it could be a good tool to say OK this is how convincingly the W/L's were decided between two similarly rated teams in another ratings system that uses more traditional W/L and SOS. I would hope the selection committee would factor in other ratings systems first or much more heavily. who you win and lose to should be much more of a consideration than by how bad you beat a cupcake for example. Seriously who cares how much except for gambling purposes?

 

[upperdeck]

Its basic flaw is that game flow does matter.

Some coaches down 10-15 with 1-2 min left call off the dogs and let the game finish out.. Others continue to press or whatever it is they do.

a nice 12 pt win can go to 6 or 20 in 2 possessions and never really be in doubt as a game.

The same that a 20 pt lead can go down to 10 because walkons for one team make 2-3 mistakes and the other teams 2nd string is trying to finish off a game strong.

Winning margin does not always show how the game went.

 

[benjams77]

I have a feeling the Bracketologist are gonna have a hard time this year and be very off... The NET this year has some weird things and it may be too much placed on Blowouts.. Look at Auburn they have Zero quad 1 wins and only played 2 quad 1 games but they are top 10 NET.. Makes no sense.. There are other teams with Michigan St has two Quad 1 wins but has 8 losses and are ranked 25th in NET.. I understand the tough schedule but you gotta win... Colorados NET numbers are similar to ours but they are ranked 29...

I saw Waters retweet some Bracket guy n "X" and he has Cuse 1st 4 out.. Someone asked and the guy said Cuse has really good resume averages abd would actually be the best resume average team left out of the feild if they didnt make it... I just think the NET is gonna be a problem this year

 

[Fly Rodder]

I brought this up last week. I don't care about the Net rankings. But Pitt being ahead of us shows flaws in the system to me.

Pitt beat Duke at Cameron. Should they be ahead of Duke?

 

[NineOneSeven]

(no system is perfect, believe it or not)

 

[blust2i3d4]

Pitt beat Duke at Cameron. Should they be ahead of Duke?

The system only saw their two losses to Syracuse.

 

[Cheriehoop]

Pitt beat Duke at Cameron. Should they be ahead of Duke?

It didn’t hurt Duke that much losing to Pitt nor losing a game to GT and having just a 5 point win in their second meeting etc. Their winning margin seems not to be an emphasis as we think it needs to be for us. They have scored just a total of 16 points more than their opponents in the past 4 games - GT, Pitt, Louisville and Clemson despite winning 3 of them. Strange? Not?

 

[dasher]

Pitt beat Duke at Cameron. Should they be ahead of Duke?

Oh, does PItt have a better record than Duke? Did they beat them sort of easily twice? Pitt has a 3-6 league record. Ours is 5-4. Pitt has a won loss record of 12-8 Ours is 14-6 And of course, we beat them twice.

 

[Fly Rodder]

Oh, does PItt have a better record than Duke? Did they beat them sort of easily twice? Pitt has a 3-6 league record. Ours is 5-4. Pitt has a won loss record of 12-8 Ours is 14-6 And of course, we beat them twice.

I think you'd be happier watching the ACC standings than NET. You're pissed because it doesn't show the ACC standings.

 

[dasher]

I think you'd be happier watching the ACC standings than NET. You're pissed because it doesn't show the ACC standings.

No. Not at all. I guess I could understand it if Pitt had a series of wins against quad 1 or quad 2 teams. They don't. We aren't real good. But we are better than them. In the end, it doesn't mean much. We are probably heading to the NIT and I think they will be staying at home. I guess you and I will have to disagree about whether PItt should have a higher net than us.

 

[IthacaMatt]

2 Comments

1. We do have a "bad loss" as of now. Florida St at home was a Q3 loss - also decent chances though that works its way up to Q2 which is nice. (Florida St needs to get in top 75 - they are being dragged down by being not very good in OOC)

2. Hate the NET formula or not its absolutely beneficial to us that Pitt stays in the top 75, while we continue to win (hopefully) and improve our resume. Its a good thing that they are 66 and not 76. (You can see all the why's being the 66 in the other thread) Not saying its a fair ranking, but its good for us that they stay in this range and keep winning.

Florida State is not necessarily going to be considered a "bad loss" by the committee. They are 6-3 in conference and played UNC tough. If they continue playing like they have lately, they are going to stack wins in the next month.

 

[pwong]

We have no bad losses. Just by common sense we should be ahead of Pitt and Miami. It's not like either of them are absolutely tearing it up; we have similar records, but we've been both 3 out of 3.

All of out losses are 20+ losses? I don’t understand how people don’t get that this is also driving our metrics down

 

[jncuse]

Florida State is not necessarily going to be considered a "bad loss" by the committee. They are 6-3 in conference and played UNC tough. If they continue playing like they have lately, they are going to stack wins in the next month.

That's potentially true as well. When you are down to the final teams the committee claims they do a deeper scrub on wins and losses... which I assume is a deeper dive then whether they are just on one side of Q1/Q2/Q3... which is why sometimes I claim don't worry too much if team is #48 or #52 if they are in the tournament like Oregon... although its still a game I get caught up in.

 

[jncuse]

The NET system obviously LARGELY factors in how much you win and lose by. Which we all know we have gotten killed when we have lost. It seems like a good tool to use for gambling purposes regarding the spread. But not necessarily a good predictor of W/L. Without looking at specific scores of every game it could be a good tool to say OK this is how convincingly the W/L's were decided between two similarly rated teams in another ratings system that uses more traditional W/L and SOS. I would hope the selection committee would factor in other ratings systems first or much more heavily. who you win and lose to should be much more of a consideration than by how badly you beat a cupcake for example. Seriously who cares how much except for gambling purposes?

Part #1 - Teams are largely selected based on quality of wins and losses irrelevant of margin. On a standalone basis NET doesn't matter overly much. That's the good news for us. The committee, whether in the RPI or NET era, has typically always made its calls based on quality of W's and quality of L's... with little attention to the margin of those games (Q1, Q2 end results matter. And Q3-Q4 really only matter if you lose -- whether you squeeze by or destroy them in your wins its largely irrelevant -- and as per my calculations in that other thread our performance against crappy teams is holding us back just as much NET wise even if we won)

In both eras, for P6 teams once you get low 40's -#70 teams that get in could be all over the place for various reason. So our low NET, especially if we can get it to below 60, shouldn't hamper us to much.

 

[IthacaMatt]

All of out losses are 20+ losses? I don’t understand how people don’t get that this is also driving our metrics down

Look, you're over-complicating this a ton. All of you passionate NET / advanced metrics guys.

Teams will win and teams will lose. Past wins will look better - and worse. I think it tends to even out.

Take Florida State, for instance. That is currently a Q3
loss , but they have won 6 out of their last 8 against ACC teams, not tomato cans. They have opportunities for Q1 wins upcoming against Duke and maybe Clemson on the road. They also have NC State at home, which would be another good win.

On the other hand, we thought winning at Pitt would be a Q1 win at the time, but they have slumped to 3-5 in the ACC, and the bloom is off that rose.

Similarly, Clemson looked great early, but has been iosing a lot of games lately. I don't think they've fallen as far as Pitt because they had some quality wins in the pre-conference season, which led to their ranking earlier this year.

When teams are all beating each other, you can't fixate on any one win or loss as the key to get you into the tournament, because the situation is dynamic, and each of those teams' ratings rise and fall every week.

Perhaps stat boys have taken over the selection committee, because I don't see how the ACC can only be the 5th or 6th best conference when they put 3 teams in the Final Four only 2 years ago.

People claim that the bottom of the ACC is soft, but didn't Clemson beat Top 25 Alabama on the road, and register good wins against South Carolina and TCU, and suffered a 2 point loss to Memphis. That's the 10th place team in the "weak" ACC.

Then look at the 11th place team, Pitt: They beat Oregon State, a good intersectional win against a P5 opponent, they beat West Virginia from the 'mighty' Big 12 on the road, and they just beat Duke. How terrible is Pitt?

And don't you think Pitt winning that Duke game, along with their follow-up win at Georgia Tech kind of offsets the computer rankings of our other opponents who may be trending downward? Pitt still is only 3-5 in the ACC, as well as 3-5 in their last 8 games.

Is the ACC going to continue to "get punished" in the computer rankings and the media, or won't these wins kind of even things out?

 

[upperdeck]

Before the ACC tourney starts
if we get to 21 wins, wins will matter more than what our Net is..
if we only get to 20 its gonna be tough
if we get to 22 we are probably in.

7-4 and really every game except UNC the talent is there to win.
4-1 at home 3-3 on the road. If we do better we are OK.

Cornell is getting votes now . if they run the table and lose in the IVY finals do they have a decent net to fall back on? They could win 10 in a row and be top 20 doing that.

 

[jncuse]

Part #2 (continued from above) - Why use NET as an initial ranking system then?
One of the oddities of using NET is that its very much based on margin. And the selection committee very much makes it decisions based on quality of wins and losses, independent (or largely independent) of margin.

So why the heck do they use NET to rank teams? It's a fair question. But it goes back to RPI.

RPI was purely based on winning or losing. So from that perspective it does align as a ranking system with what the selection committee. But the RPI produced a lot of stupid... and it could be more easily gamed by some mid-tier conferences like MVC.

NET is generally more favourable to the P6 conferences. There might be a few anomalies of invidual teams because of margin, but basically the conference that do the best in OOC play will come out as the best in NET. It can't really be games by a mid-conference.

 

[pwong]

Look, you're over-complicating this a ton. All of you passionate NET / advanced metrics guys.

Teams will win and teams will lose. Past wins will look better - and worse. I think it tends to even out.

Take Florida State, for instance. That is currently a Q3
loss , but they have won 6 out of their last 8 against ACC teams, not tomato cans. They have opportunities for Q1 wins upcoming against Duke and maybe Clemson on the road. They also have NC State at home, which would be another good win.

On the other hand, we thought winning at Pitt would be a Q1 win at the time, but they have slumped to 3-5 in the ACC, and the bloom is off that rose.

Similarly, Clemson looked great early, but has been iosing a lot of games lately. I don't think they've fallen as far as Pitt because they had some quality wins in the pre-conference season, which led to their ranking earlier this year.

When teams are all beating each other, you can't fixate on any one win or loss as the key to get you into the tournament, because the situation is dynamic, and each of those teams' ratings rise and fall every week.

Perhaps stat boys have taken over the selection committee, because I don't see how the ACC can only be the 5th or 6th best conference when they put 3 teams in the Final Four only 2 years ago.

People claim that the bottom of the ACC is soft, but didn't Clemson beat Top 25 Alabama on the road, and register good wins against South Carolina and TCU, and suffered a 2 point loss to Memphis. That's the 10th place team in the "weak" ACC.

Then look at the 11th place team, Pitt: They beat Oregon State, a good intersectional win against a P5 opponent, they beat West Virginia from the 'mighty' Big 12 on the road, and they just beat Duke. How terrible is Pitt?

And don't you think Pitt winning that Duke game, along with their follow-up win at Georgia Tech kind of offsets the computer rankings of our other opponents who may be trending downward? Pitt still is only 3-5 in the ACC, as well as 3-5 in their last 8 games.

Is the ACC going to continue to "get punished" in the computer rankings and the media, or won't these wins kind of even things out?

Just because you don’t understand how the NET works, doesn’t mean you need to throw away the entire system. Not every record is the same, you need a system to determine how much a team and result is valued (wins against good teams are valued higher than wins against bad teams, losses against bad teams are worse than losses against medoicre teams, blowout losses are worse than single digit losses)

Is the NET perfect? No. But it’s better than no system at all.

 

[IthacaMatt]

Just because you don’t understand how the NET works, doesn’t mean you need to throw away the entire system. Not every record is the same, you need a system to determine how much a team and result is valued (wins against good teams are valued higher than wins against bad teams, losses against bad teams are worse than losses against medoicre teams, blowout losses are worse than single digit losses)

Is the NET perfect? No. But it’s better than no system at all.

It's seems a lot worse than RPI, wasn't that what the original one was called? I think this whole "5 year rolling average" in an era of NIL and portal is nonsense. That underlies KenPom. We saw him say so himself in that interview with Mike Waters. And KenPom is kind of the step-father of the current system.

P.S. - I am way more of a stat head than you'll ever know.