PDA

View Full Version : NET News (Feb 18): Has Duke pulled away? Is the fourth No. 1 seed in play?



scottdude8
02-18-2019, 01:26 PM
Just like in the recent AP Poll, in the up-to-date NET rankings (https://www.ncaa.com/rankings/basketball-men/d1/ncaa-mens-basketball-net-rankings) Duke has reclaimed its spot at No. 1. Meanwhile, Kentucky's annihilation of Tennessee on Saturday night dropped Tennessee two spots to No. 6 in the NET, while Kentucky jumped them into No. 5. Over in the B1G, MSU is holding steady at No. 7 and Michigan remains at No. 8.


Sidebar: I continue to think that the NET placing MSU ahead of Michigan, despite them playing in the same conference and MSU having two more losses, including a horrible one at home, is indicative of a flaw in the NET. MSU is better in the analytic rankings, and did face a tougher non-conference schedule, but they didn't actually WIN any of those tougher non-conference games. For comparison, MSU has one quad 1 non-conference win, and their next best non-conference win isn't until NET 114 UCLA, while Michigan has two top-tier non-conference wins against UNC and Villanova, and two more non-conference wins that are better than MSU's win over UCLA. If there's anything that needs to be changed in the NET, I think it has to be weighing actually WINNING games more heavily. End sidebar rant.


Getting into the team sheets (https://extra.ncaa.org/solutions/rpi/Stats%20Library/NET%20Team%20Sheets%20-%20Games%20through%20Feb.%2017,%202019.pdf) today, I wanted to analyze two key questions: first, has Duke "pulled ahead" of the field for the No. 1 overall seed following the Tennessee loss, and second, has the Tennessee loss opened the door for a non-SEC team to take the fourth No. 1 seed.

To answer the first question, I think the answer is definitely yes, and that's where the "talking head" consensus seems to be as well. Duke is No. 1 in every ranking system listed on the Team Sheet except BPI, where they are 2 (Gonzaga is No. 1 there). Not only is Duke 7-1 in Q1 games, every one of their current Q1 games falls within the "top-half of Q1"... remember, each quadrant is split into the top and bottom half this year, essentially making it more like an eight-tier system (it isn't clear how much weight the committee will put on this distinction, but it is there on the sheets). While Gonzaga has the head-to-head victory against us, they lost their only other top-tier Q1 games in the non-conference (they have a victory against Creighton that falls in the bottom half of Q1), and their only other Q1 wins are road WCC games. I would think that sweeping Virginia would have to keep us ahead of Virginia in the seedings barring something unusual. So at the moment, the biggest challenger for the No. 1 overall seed would likely half to come from the SEC, where the Kentucky-Tennessee result definitely helped us. The only way Kentucky could challenge for the top overall seed is if they win out, and even then you would imagine the head-to-head beatdown we served them would come into play.

That leaves Tennessee as the most likely foil for our No. 1 overall seed hopes. But the fact that the loss dropped Tennessee all the way to No. 6 in the NET seems to indicate that there is a major divide between us and the Vols. Tennessee is down to 11 in the KPI, and 7 in the other analytic rankings. The weakness of their SEC schedule is also going to hamper them significantly: none of Tennessee's current conference victories fall into the Q1 category. Now, 5 of their remaining games are predicted to be Q1, giving them room to improve their resume... but in comparison, Michigan has 4 such games while already holding two more Q1 wins, Virginia also has 4 such games remaining and already have three more Q1 wins, and we also have 4 such games remaining with three more Q1 wins in hand.

The conclusion: barring something unusual, Tennessee remains the most likely competition for the No. 1 overall seed, but there's enough space between us and the Vols that they might need to win out to make things interesting.

Now, to the second question, which I think is being massively overlooked by the talking heads, especially Joe Lunardi (who said all weekend how the No. 1 overall seeds are all but "locked", which is a kiss of death!). Obviously at this point the two top SEC teams probably control their own destinies to claim a No. 1 seed, and the same can be said of Virginia and Gonzaga. But, there are at least three scenarios I can think of that would significantly weaken that stranglehold: Virginia has three road Q1 games remaining, so if they lose at least one of those and fall in the ACC Tourney they can't feel secure; a single loss by Gonzaga, while highly unlikely, would likely knock them out of 1 seed contention; and if the SEC beats up on itself down the stretch the SEC champion could conceivably have 4 or 5 conference losses. I'd put the odds at greater than a coin-flip that one of those scenarios occurs and makes the last spot open.

Which leads me to the following conclusion: despite being out of the national conversation recently, Michigan still has a very feasible path to a No. 1 seed. People forget that Michigan still only has 3 losses, and all of them are road, conference Q1 losses (which in my mind are the most "justifiable"). What's more, MSU's recent bad luck losing Nick Ward for the foreseeable future significantly increases the likelihood that Michigan takes both games against Michigan State. Those who are skewed by the historical love of Tom Izzo and MSU over Michigan seem to gloss over the fact that Michigan has taken three straight from MSU, including last year in East Lansing and in the B1G Tournament. (Also forgotten in that discussion: with all the injuries, Cassius Winston is now MSU's only viable offensive threat. Zavier Simpson has historically shut Winston down, going all the way back to high school. The matchup favors Michigan significantly.) If Michigan sweeps MSU in the regular season, perhaps has one hiccup on the road against Maryland or Minnesota (both of which would be Q1 games), and then wins the B1G tournament, I think it would be really hard to argue against a four-loss consensus B1G champion as a 1 seed unless the other contenders have clearly ran ahead of the pack. I think that resume would even stand up extremely well to Gonzaga without an additional loss considering their weak SOS.

Side bar: there's been some talk that UNC could get back into the 1 seed conversation. I find this extremely premature (it'd probably take something very "bold" (https://forums.dukebasketballreport.com/forums/showthread.php?43205-Well-this-certainly-qualifies-as-a-quot-bold-prediction-quot)for this to occur), but the Tar Heels are at No. 9 in the NET, and every one of their losses has come in the Q1. So this also isn't out of the realm of possibility.

All of that is to say that all of the talking heads saying the No. 1 seeds are "locked" seem to be glossing over a LOT of reasonably possible scenarios in which that isn't the case. Shocking, I know!

Love to hear everyone's thoughts on the current NET rankings, their reactions to my analysis, etc. as usual!

Kedsy
02-18-2019, 01:36 PM
Anyone who at this point in the season says that all four #1 seeds are "locked" is ignoring the lessons of history.

Mtn.Devil.91.92.01.10.15
02-18-2019, 01:37 PM
Just like in the recent AP Poll, in the up-to-date NET rankings (https://www.ncaa.com/rankings/basketball-men/d1/ncaa-mens-basketball-net-rankings) Duke has reclaimed its spot at No. 1. Meanwhile, Kentucky's annihilation of Tennessee on Saturday night dropped Tennessee two spots to No. 6 in the NET, while Kentucky jumped them into No. 5. Over in the B1G, MSU is holding steady at No. 7 and Michigan remains at No. 8.


Sidebar: I continue to think that the NET placing MSU ahead of Michigan, despite them playing in the same conference and MSU having two more losses, including a horrible one at home, is indicative of a flaw in the NET. MSU is better in the analytic rankings, and did face a tougher non-conference schedule, but they didn't actually WIN any of those tougher non-conference games. For comparison, MSU has one quad 1 non-conference win, and their next best non-conference win isn't until NET 114 UCLA, while Michigan has two top-tier non-conference wins against UNC and Villanova, and two more non-conference wins that are better than MSU's win over UCLA. If there's anything that needs to be changed in the NET, I think it has to be weighing actually WINNING games more heavily. End sidebar rant.


Getting into the team sheets (https://extra.ncaa.org/solutions/rpi/Stats%20Library/NET%20Team%20Sheets%20-%20Games%20through%20Feb.%2017,%202019.pdf) today, I wanted to analyze two key questions: first, has Duke "pulled ahead" of the field for the No. 1 overall seed following the Tennessee loss, and second, has the Tennessee loss opened the door for a non-SEC team to take the fourth No. 1 seed.

To answer the first question, I think the answer is definitely yes, and that's where the "talking head" consensus seems to be as well. Duke is No. 1 in every ranking system listed on the Team Sheet except BPI, where they are 2 (Gonzaga is No. 1 there). Not only is Duke 7-1 in Q1 games, every one of their current Q1 games falls within the "top-half of Q1"... remember, each quadrant is split into the top and bottom half this year, essentially making it more like an eight-tier system (it isn't clear how much weight the committee will put on this distinction, but it is there on the sheets). While Gonzaga has the head-to-head victory against us, they lost their only other top-tier Q1 games in the non-conference (they have a victory against Creighton that falls in the bottom half of Q1), and their only other Q1 wins are road WCC games. I would think that sweeping Virginia would have to keep us ahead of Virginia in the seedings barring something unusual. So at the moment, the biggest challenger for the No. 1 overall seed would likely half to come from the SEC, where the Kentucky-Tennessee result definitely helped us. The only way Kentucky could challenge for the top overall seed is if they win out, and even then you would imagine the head-to-head beatdown we served them would come into play.

That leaves Tennessee as the most likely foil for our No. 1 overall seed hopes. But the fact that the loss dropped Tennessee all the way to No. 6 in the NET seems to indicate that there is a major divide between us and the Vols. Tennessee is down to 11 in the KPI, and 7 in the other analytic rankings. The weakness of their SEC schedule is also going to hamper them significantly: none of Tennessee's current conference victories falls into the Q1 category. Now, 5 of their remaining games are predicted to be Q1, giving them room to improve their resume... but in comparison, Michigan has 4 such games while already holding two more Q1 wins, Virginia also has 4 such games remaining and already have three more Q1 wins, and we also have 4 such games remaining with three more Q1 wins in hand.

The conclusion: barring something unusual, Tennessee remains the most likely competition for the No. 1 overall seed, but there's enough space between us and the Vols that they might need to win out to make things interesting.

Now, to the second question, which I think is being massively overlooked by the talking heads, especially Joe Lunardi (who said all weekend how the No. 1 overall seeds are all but "locked", which is a kiss of death!). Obviously at this point the two top SEC teams probably control their own destinies to claim a No. 1 seed, and the same can be said of Virginia and Gonzaga. But, there are at least three scenarios I can think of that would significantly weaken that stranglehold: Virginia has three road Q1 games remaining, so if they lose at least one of those and fall in the ACC Tourney they can't feel secure; a single loss by Gonzaga, while highly unlikely, would likely knock them out of 1 seed contention; and if the SEC beats up on itself down the stretch the SEC champion could conceivably have 4 or 5 conference losses. I'd put the odds at greater than a coin-flip that one of those scenarios occurs and makes the last spot open.

Which leads me to the following conclusion: despite being out of the national conversation recently, Michigan still has a very feasible path to a No. 1 seed. People forget that Michigan still only has 3 losses, and all of them are road, conference Q1 losses (which in my mind are the most "justifiable"). What's more, MSU's recent bad luck losing Nick Ward for the foreseeable future significantly increases the likelihood that Michigan takes both games against Michigan State. Those who are skewed by the historical love of Tom Izzo and MSU over Michigan seem to gloss over the fact that Michigan has taken three straight from MSU, including last year in East Lansing and in the B1G Tournament. (Also forgotten in that discussion: with all the injuries, Cassius Winston is now MSU's only viable offensive threat. Zavier Simpson has historically shut Winston down, going all the way back to high school. The matchup favors Michigan significantly.) If Michigan sweeps MSU in the regular season, perhaps has one hiccup on the road against Maryland or Minnesota (both of which would be Q1 games), and then wins the B1G tournament, I think it would be really hard to argue against a four-loss consensus B1G champion as a 1 seed unless the other contenders have clearly ran ahead of the pack. I think that resume would even stand up extremely well to Gonzaga without an additional loss considering their weak SOS.

Side bar: there's been some talk that UNC could get back into the 1 seed conversation. I find this extremely premature (it'd probably take something very "bold" (https://forums.dukebasketballreport.com/forums/showthread.php?43205-Well-this-certainly-qualifies-as-a-quot-bold-prediction-quot)for this to occur), but the Tar Heels are at No. 9 in the NET, and every one of their losses has come in the Q1. So this also isn't out of the realm of possibility.

All of that is to say that all of the talking heads saying the No. 1 seeds are "locked" seem to be glossing over a LOT of reasonably possible scenarios in which that isn't the case. Shocking, I know!

Love to hear everyone's thoughts on the current NET rankings, their reactions to my analysis, etc. as usual!

That all looks about right to me. I know you area Michigan homer, but their profile and schedule suggests a good shot to me as well.

And I will refrain from my diatribe on the word "lock."

MChambers
02-18-2019, 01:42 PM
Anybody want to defend Houston as a #1 seed?

scottdude8
02-18-2019, 01:46 PM
That all looks about right to me. I know you area Michigan homer, but their profile and schedule suggests a good shot to me as well.

And I will refrain from my diatribe on the word "lock."

I tried my best to minimize my Michigan "homer-ing" in my analysis, but there's no doubt it creeped in there. Still, I find it funny how a team with the second-least losses amongst power-conference teams has fallen so completely out of the national discussion. The Penn State loss was bad, but not nearly as bad as a cursory look at the records might suggest (as a Michigan fan site put it, "put Penn State in the America East and they'd be a popular 12-seed"). And people have completely forgotten that this team annihilated three top 20 teams in UNC, Purdue, and Villanova (on the road!) early in the season. They've regressed, yes, but a lot of that can be attributed to three of the team's biggest contributors (Iggy, Poole, and Teske) having significantly increased roles over last season and potentially hitting the mid-season doldrums. If you look purely at the resumes and not at the team names on the marquee, I think Michigan's resume is very competitive against Gonzaga's as well as either SEC champion, and as I highlighted it has the potential to get significantly better.

Also don't forget that recent history shows that Jon Beilein's teams also play their best basketball towards the end of the season (he's very K-like in that way). Michigan didn't lose a game after February 6 until the National Championship game last year, and the year before they only lost two road games between February 4 and the Sweet 16. There's a lot of basketball to be played to determine the No. 1 seeds, to be sure. Hence why Lunardi drove me INSANE this whole weekend.

scottdude8
02-18-2019, 01:48 PM
Anybody want to defend Houston as a #1 seed?

Ummmm... no (haha).

uh_no
02-18-2019, 02:00 PM
If there's anything that needs to be changed in the NET, I think it has to be weighing actually WINNING games more heavily.


You can't have your cake and eat it too. If you reward winning too heavily, you get teams playing nobody all season, have gaudy records, and actually stink getting high seeds. Michigans NCSOS is among the worst in the country. They ought to be penalized for that.

If all you care about is winning, then go by record. If all you care about is evaluation of a team's actual strength, then go KP. NET takes both into account....and it still says MSU is better than michigan.

ElliottHoo
02-18-2019, 02:08 PM
And people have completely forgotten that this team annihilated three top 20 teams in UNC, Purdue, and Villanova (on the road!) early in the season.

The yo-yoing of public perception on Nova (and therefore UM’s win) is fascinating. Early on, consensus was that Nova was primed to return strong so the win was viewed as huge, plus UM had the huge unbeaten streak, and UM was viewed as a title-favorite.

But then at around the same time, UM finally lost a few games AND the pundits realized Nova had fallen out of the top 25 and maybe wasn’t as good as originally expected (this actually happened a couple weeks before, iirc, but the pundits didn’t realize it and the implications for a bit). Then the narrative shifts to “UM is good, BUT...” weak SoS, declining offense, whatever. And UM drops out of the discussion of elite teams despite basically still being right there.

Since then, UM’s done solid work in-conference, other teams (especially MSU) are maybe weakening a bit, and here comes Nova back to significance (at least until the loss to St Johns). Any minute now, some pundit will realize the Nova win is now suddely a good one again and UM deserves to be in the discussion of top teams.

scottdude8
02-18-2019, 02:13 PM
You can't have your cake and eat it too. If you reward winning too heavily, you get teams playing nobody all season, have gaudy records, and actually stink getting high seeds. Michigans NCSOS is among the worst in the country. They ought to be penalized for that.

If all you care about is winning, then go by record. If all you care about is evaluation of a team's actual strength, then go KP. NET takes both into account...and it still says MSU is better than michigan.

Very fair criticism of my statement. Let me make the following adjustment: I think the NET needs to better reward victories over high-quality opponents in comparison to overall SOS, particularly when it comes to the non-conference. MSU played a quantifiably better non-conference schedule than Michigan, for sure, but in no way did they have a better non-conference performance considering who amongst those teams they actually beat. The bottom of Michigan's non-conference schedule may have been weaker than MSU's, but they also won their marquee games. I would argue that wins over marquee opponents in the non-conference is more important than whether your "tune-up" games are against 100-200 NET teams or 200-300 NET teams, although this is a point upon which reasonable people can definitely disagree. I see your point about the cake, but I think that this isn't a binary choice... there is a middle ground where these two factors can be appropriately weighted so that they both are taken into account, but in the correct ratio.

This is one of many places where knowing more details into how the NET is calculated would be helpful, but alas that is not to be. However, considering seedings are not simply determined by the NET, but by human analysis of the NET and the corresponding team sheets, one would hope that this difference in non-conference resume would be reflected in the seedings by looking at the team sheets. But most prognosticators don't seem to think this will be the case, as MSU seems to continually be ranked ahead of Michigan on the 2-seed line (which would have a non-trivial affect on the location of their first weekend games, just like it did last year). That is discouraging.

P.S. Case and point regarding MSU/Michigan non-conference strength of schedule: Michigan had 7 Q4 non-conference games and 1 Q3, while MSU had 4 Q4 and 3 Q3. That seems to be the biggest factor in their relative strength of schedule. Now, do we really think that, for the top-tier teams, there is that much difference between a Q3 and Q4 game, especially when they're all at home (they were)? That differentiator is probably more important in determining the at-large teams as opposed to top of the bracket seeding.

JasonEvans
02-18-2019, 02:31 PM
Ummmm... no (haha).

Yeah, I mean what is the case for Houston other than "we have won a lot of games against decent but not great teams"? Sure, 25-1 is impressive, but when your best win is a home game against LSU and your second best is a road win at Cincy I have a hard time thinking you've played a schedule worthy of a top seed. The loss to Temple isn't awful, but it isn't great either.

If Houston was unbeaten then they could say, "we've beaten everyone!" But once they took one loss, I think they are no better than a #2 or #3.

What's more, all the advanced metrics are not impressed with the Cougars. We are talking about a team that is #14 in T-rank, #18 in Pom, and #18 in BPI. Nevada has a much better case for a #1 seed (and I don't think they have much of a case either).

-Jason "Can Houston be our #2 seed... please?!?!?" Evans

JasonEvans
02-18-2019, 02:47 PM
I'll be so bold as to say this... Duke has a 2 game cushion at this point.

By that I mean I think we could afford to lose 2 more games and still be a #1 seed. I don't think we will lose any games, but here is how the 2 game cushion scenario would play out.


1) Lose 1 more regular season game and 1 in the ACC tourney - In this case, Duke would win the ACC regular season (at worst, in a convincing 2-0 tiebreaker against Virginia, but likely win it outright) and then drop a game in the ACC tourney. Given the many, many impressive wins on our schedule and the way the advanced metrics love us, I think we would be certain to get a #1 seed.

2) Lose 2 regular season games but then win the ACC tourney - While it is possible this scenario denies us a regular season title (though I think it is likely that 3 total ACC losses will at least give you a tie for the ACC reg season title), the win in the ACC tourney would present us as one of the "hot teams" in the nation and would give us a #1 seed.

Now, I know nothing happens in a vaccum and we need to see what other teams are doing to have a real sense of this. I also know that wins and losses are not everything. Duke could win games but look different than they have and appear to be struggling against inferior competition, but I consider this scenario so unlikely as to be nearly moot.

Bottom line -- I'm feeling really good about how the seeding is shaping up for this year's team.

-Jason "what would be our preferred region? Washington, not Louisville, right? DC is certainly closer to Duke but I wonder if they would move us to Lou so they could let Virginia stay really close to home" Evans

ChillinDuke
02-18-2019, 02:50 PM
Anyone who at this point in the season says that all four #1 seeds are "locked" is ignoring the lessons of history.

Seriously.

It's so obvious to me but others (Joseph B. Lunardi, e.g.) throw around this term "lock" so prematurely. Is it really that hard to envision Duke losing 2 out of its next 3? Where is Duke projected then? And with @UNC still on the schedule. Plus the ACCT. We are absolutely not "locked" into anything. And stranger things have happened.

That said, it's looking likely that Duke has built up a reasonable amount of wiggle room to remain on the 1-line. I think if Duke loses only one more game for the remainder of the ACC schedule, we'd be locked at that point regardless of what happens in the ACCT. At that point, we'd have at least 3 more Q1 wins. I'd be comfortable using the word lock at that time.

But essentially until the final regular season game (@UNC), I'm not really sure the word lock can reasonably be uttered. And if not for Duke, certainly not for anyone.

- Chillin

freshmanjs
02-18-2019, 02:52 PM
I'll be so bold as to say this... Duke has a 2 game cushion at this point.

By that I mean I think we could afford to lose 2 more games and still be a #1 seed. I don't think we will lose any games, but here is how the 2 game cushion scenario would play out.


1) Lose 1 more regular season game and 1 in the ACC tourney - In this case, Duke would win the ACC regular season (at worst, in a convincing 2-0 tiebreaker against Virginia, but likely win it outright) and then drop a game in the ACC tourney. Given the many, many impressive wins on our schedule and the way the advanced metrics love us, I think we would be certain to get a #1 seed.

2) Lose 2 regular season games but then win the ACC tourney - While it is possible this scenario denies us a regular season title (though I think it is likely that 3 total ACC losses will at least give you a tie for the ACC reg season title), the win in the ACC tourney would present us as one of the "hot teams" in the nation and would give us a #1 seed.

Now, I know nothing happens in a vaccum and we need to see what other teams are doing to have a real sense of this. I also know that wins and losses are not everything. Duke could win games but look different than they have and appear to be struggling against inferior competition, but I consider this scenario so unlikely as to be nearly moot.

Bottom line -- I'm feeling really good about how the seeding is shaping up for this year's team.

-Jason "what would be our preferred region? Washington, not Louisville, right? DC is certainly closer to Duke but I wonder if they would move us to Lou so they could let Virginia stay really close to home" Evans


A 4 loss Duke would be a 100% lock for a #1 seed. Suspect that even a 6 loss Duke team would have a good shot at it.

dukelifer
02-18-2019, 02:53 PM
Just like in the recent AP Poll, in the up-to-date NET rankings (https://www.ncaa.com/rankings/basketball-men/d1/ncaa-mens-basketball-net-rankings) Duke has reclaimed its spot at No. 1. Meanwhile, Kentucky's annihilation of Tennessee on Saturday night dropped Tennessee two spots to No. 6 in the NET, while Kentucky jumped them into No. 5. Over in the B1G, MSU is holding steady at No. 7 and Michigan remains at No. 8.


Sidebar: I continue to think that the NET placing MSU ahead of Michigan, despite them playing in the same conference and MSU having two more losses, including a horrible one at home, is indicative of a flaw in the NET. MSU is better in the analytic rankings, and did face a tougher non-conference schedule, but they didn't actually WIN any of those tougher non-conference games. For comparison, MSU has one quad 1 non-conference win, and their next best non-conference win isn't until NET 114 UCLA, while Michigan has two top-tier non-conference wins against UNC and Villanova, and two more non-conference wins that are better than MSU's win over UCLA. If there's anything that needs to be changed in the NET, I think it has to be weighing actually WINNING games more heavily. End sidebar rant.


Getting into the team sheets (https://extra.ncaa.org/solutions/rpi/Stats%20Library/NET%20Team%20Sheets%20-%20Games%20through%20Feb.%2017,%202019.pdf) today, I wanted to analyze two key questions: first, has Duke "pulled ahead" of the field for the No. 1 overall seed following the Tennessee loss, and second, has the Tennessee loss opened the door for a non-SEC team to take the fourth No. 1 seed.

To answer the first question, I think the answer is definitely yes, and that's where the "talking head" consensus seems to be as well. Duke is No. 1 in every ranking system listed on the Team Sheet except BPI, where they are 2 (Gonzaga is No. 1 there). Not only is Duke 7-1 in Q1 games, every one of their current Q1 games falls within the "top-half of Q1"... remember, each quadrant is split into the top and bottom half this year, essentially making it more like an eight-tier system (it isn't clear how much weight the committee will put on this distinction, but it is there on the sheets). While Gonzaga has the head-to-head victory against us, they lost their only other top-tier Q1 games in the non-conference (they have a victory against Creighton that falls in the bottom half of Q1), and their only other Q1 wins are road WCC games. I would think that sweeping Virginia would have to keep us ahead of Virginia in the seedings barring something unusual. So at the moment, the biggest challenger for the No. 1 overall seed would likely half to come from the SEC, where the Kentucky-Tennessee result definitely helped us. The only way Kentucky could challenge for the top overall seed is if they win out, and even then you would imagine the head-to-head beatdown we served them would come into play.

That leaves Tennessee as the most likely foil for our No. 1 overall seed hopes. But the fact that the loss dropped Tennessee all the way to No. 6 in the NET seems to indicate that there is a major divide between us and the Vols. Tennessee is down to 11 in the KPI, and 7 in the other analytic rankings. The weakness of their SEC schedule is also going to hamper them significantly: none of Tennessee's current conference victories fall into the Q1 category. Now, 5 of their remaining games are predicted to be Q1, giving them room to improve their resume... but in comparison, Michigan has 4 such games while already holding two more Q1 wins, Virginia also has 4 such games remaining and already have three more Q1 wins, and we also have 4 such games remaining with three more Q1 wins in hand.

The conclusion: barring something unusual, Tennessee remains the most likely competition for the No. 1 overall seed, but there's enough space between us and the Vols that they might need to win out to make things interesting.

Now, to the second question, which I think is being massively overlooked by the talking heads, especially Joe Lunardi (who said all weekend how the No. 1 overall seeds are all but "locked", which is a kiss of death!). Obviously at this point the two top SEC teams probably control their own destinies to claim a No. 1 seed, and the same can be said of Virginia and Gonzaga. But, there are at least three scenarios I can think of that would significantly weaken that stranglehold: Virginia has three road Q1 games remaining, so if they lose at least one of those and fall in the ACC Tourney they can't feel secure; a single loss by Gonzaga, while highly unlikely, would likely knock them out of 1 seed contention; and if the SEC beats up on itself down the stretch the SEC champion could conceivably have 4 or 5 conference losses. I'd put the odds at greater than a coin-flip that one of those scenarios occurs and makes the last spot open.

Which leads me to the following conclusion: despite being out of the national conversation recently, Michigan still has a very feasible path to a No. 1 seed. People forget that Michigan still only has 3 losses, and all of them are road, conference Q1 losses (which in my mind are the most "justifiable"). What's more, MSU's recent bad luck losing Nick Ward for the foreseeable future significantly increases the likelihood that Michigan takes both games against Michigan State. Those who are skewed by the historical love of Tom Izzo and MSU over Michigan seem to gloss over the fact that Michigan has taken three straight from MSU, including last year in East Lansing and in the B1G Tournament. (Also forgotten in that discussion: with all the injuries, Cassius Winston is now MSU's only viable offensive threat. Zavier Simpson has historically shut Winston down, going all the way back to high school. The matchup favors Michigan significantly.) If Michigan sweeps MSU in the regular season, perhaps has one hiccup on the road against Maryland or Minnesota (both of which would be Q1 games), and then wins the B1G tournament, I think it would be really hard to argue against a four-loss consensus B1G champion as a 1 seed unless the other contenders have clearly ran ahead of the pack. I think that resume would even stand up extremely well to Gonzaga without an additional loss considering their weak SOS.

Side bar: there's been some talk that UNC could get back into the 1 seed conversation. I find this extremely premature (it'd probably take something very "bold" (https://forums.dukebasketballreport.com/forums/showthread.php?43205-Well-this-certainly-qualifies-as-a-quot-bold-prediction-quot)for this to occur), but the Tar Heels are at No. 9 in the NET, and every one of their losses has come in the Q1. So this also isn't out of the realm of possibility.

All of that is to say that all of the talking heads saying the No. 1 seeds are "locked" seem to be glossing over a LOT of reasonably possible scenarios in which that isn't the case. Shocking, I know!

Love to hear everyone's thoughts on the current NET rankings, their reactions to my analysis, etc. as usual!

Seems very early for such a discussion. Just hoping Duke wins on Wed. No locks there.

Troublemaker
02-18-2019, 02:58 PM
-Jason "what would be our preferred region? Washington, not Louisville, right? DC is certainly closer to Duke but I wonder if they would move us to Lou so they could let Virginia stay really close to home" Evans

Easily DC. Kentucky is allowed to be the 2 seed in the Louisville region. And even if UK isn't placed there, their fans might show up anyway to root against Duke.

I believe teams are allowed to tell the committee their preferences for sites these days, so if Duke is ahead of UVA in the overall seeding, we will get DC.

scottdude8
02-18-2019, 03:06 PM
Easily DC. Kentucky is allowed to be the 2 seed in the Louisville region. And even if UK isn't placed there, their fans might show up anyway to root against Duke.

I believe teams are allowed to tell the committee their preferences for sites these days, so if Duke is ahead of UVA in the overall seeding, we will get DC.

I looked this up a little while back, and if memory serves only the No. 1 overall seed is explicitly allowed to give the committee its preference, and everything else is done purely by distance. If that is still true it's more motivation to claim that top spot.

ChillinDuke
02-18-2019, 03:17 PM
A 4 loss Duke would be a 100% lock for a #1 seed. Suspect that even a 6 loss Duke team would have a good shot at it.

I dunno. I can still craft scenarios where a 4-loss Duke team looks questionable, certainly not a 100% lock. What if one of those losses is a home loss to Wake Forest? Or even Miami?

What if UVA, Tenn, Mich, and Gonzaga win out?

Look, I get that this is contrived. But the point is 100% is still an overestimate. If Duke beats UNC this week, I will probably agree that a 4-loss Duke team is a 95% (still trying to decide if the most contrived 2-loss remainder would be enough to knock us off - it might be).

- Chillin

HereBeforeCoachK
02-18-2019, 03:19 PM
I dunno. I can still craft scenarios where a 4-loss Duke team looks questionable, certainly not a 100% lock. What if one of those losses is a home loss to Wake Forest? Or even Miami?

What if UVA, Tenn, Mich, and Gonzaga win out?

Look, I get that this is contrived. But the point is 100% is still an overestimate. If Duke beats UNC this week, I will probably agree that a 4-loss Duke team is a 95% (still trying to decide if the most contrived 2-loss remainder would be enough to knock us off - it might be).

- Chillin

And of course, it depends on how the other number ones, the twos, even threes...finish their seasons.

uh_no
02-18-2019, 03:20 PM
I looked this up a little while back, and if memory serves only the No. 1 overall seed is explicitly allowed to give the committee its preference, and everything else is done purely by distance. If that is still true it's more motivation to claim that top spot.

Alright...hear me out....here's how bracketing should go....this would be incredible must-watch TV and wouldn't involve stretching the "bracket reveal" over 19 hours like CBS does.


-populate the bottom 32 teams in the bracket by s curve
-starting with the 8 seeds, best to worst, they each choose which bracket they want to go in
-then the 7 seeds, best to worst
-...
-finally the 1 seeds choose which bracket they want, meaning the #4 overall team automatically gets the last remaining spot

In terms of locations, starting at the 4 seeds, we choose pod locations as well
the best 4 seed chooses his pod, then next best, etc
then the 3 seeds, and 2 seeds.

There is one caveat...if a pod location is already filled, you can "steal" a pod from a lower ranked team, white elephant style. That team then can steal from a lower ranked, or choose a still available location.

When you get to 1 seeds, you choose your regional as well as choosing/stealing your pod.


I would watch every bit of that as teams try to screw eachother stealing pods and choosing their opponents.

Wander
02-18-2019, 03:29 PM
first, has Duke "pulled ahead" of the field for the No. 1 overall seed following the Tennessee loss

IMO, Duke was already significantly ahead of the field for the overall #1 seed even before the Tennessee loss. I don't think most people realize how good our schedule has been. Realistically, I think we might have a two game cushion for the OVERALL #1 seed, not just a #1 seed (that would depend on how other teams perform, but I think it is likely that Tennessee, UVA, and other contenders all take at least one more loss).

Nothing is guaranteed until we see we don't take some bad upset losses, but we're in great shape.

freshmanjs
02-18-2019, 03:43 PM
I dunno. I can still craft scenarios where a 4-loss Duke team looks questionable, certainly not a 100% lock. What if one of those losses is a home loss to Wake Forest? Or even Miami?

What if UVA, Tenn, Mich, and Gonzaga win out?

Look, I get that this is contrived. But the point is 100% is still an overestimate. If Duke beats UNC this week, I will probably agree that a 4-loss Duke team is a 95% (still trying to decide if the most contrived 2-loss remainder would be enough to knock us off - it might be).

- Chillin

If we lose our "easiest" remaining games -- let's say Wake at home and an early loss in ACCT, we would still be ahead of Gonzaga (no matter what they do). Committee cares much more about wins than losses. We have the best wins of anyone by far. It's not close. A four loss team would add to that by beating at least 2 of Virginia Tech away, Syracuse away, or UNC. A 4 loss Duke would have a better resume than we have now, not worse. Regardless of who we lose to on our schedule.


And of course, it depends on how the other number ones, the twos, even threes...finish their seasons.

No, in this case I don't think it does. Any 2 losses on our schedule would keep us ahead of Gonzaga, Michigan, MSU regardless of what they do.

HereBeforeCoachK
02-18-2019, 03:46 PM
If we lose our "easiest" remaining games -- let's say Wake at home and an early loss in ACCT, we would still be ahead of Gonzaga (no matter what they do). Committee cares much more about wins than losses. We have the best wins of anyone by far. It's not close. A four loss team would add to that by beating at least 2 of Virginia Tech away, Syracuse away, or UNC. A 4 loss Duke would have a better resume than we have now, not worse. Regardless of who we lose to on our schedule.



No, in this case I don't think it does. Any 2 losses on our schedule would keep us ahead of Gonzaga, Michigan, MSU regardless of what they do.

Not sure about that if the Zags, the Vols and/or the Wolverines go undefeated from here on.......but my other point was, if all of those teams stumble, then Duke might have a 3 game pad.

uh_no
02-18-2019, 04:01 PM
If we lose our "easiest" remaining games -- let's say Wake at home and an early loss in ACCT, we would still be ahead of Gonzaga (no matter what they do). Committee cares much more about wins than losses. We have the best wins of anyone by far. It's not close. A four loss team would add to that by beating at least 2 of Virginia Tech away, Syracuse away, or UNC. A 4 loss Duke would have a better resume than we have now, not worse. Regardless of who we lose to on our schedule.



No, in this case I don't think it does. Any 2 losses on our schedule would keep us ahead of Gonzaga, Michigan, MSU regardless of what they do.

Seth Greenberg would beg to differ. He knows exactly how to massage his losses to offset big wins and still miss the tournament.

fuse
02-18-2019, 04:38 PM
Alright...hear me out...here's how bracketing should go...this would be incredible must-watch TV and wouldn't involve stretching the "bracket reveal" over 19 hours like CBS does.


-populate the bottom 32 teams in the bracket by s curve
-starting with the 8 seeds, best to worst, they each choose which bracket they want to go in
-then the 7 seeds, best to worst
-...
-finally the 1 seeds choose which bracket they want, meaning the #4 overall team automatically gets the last remaining spot

In terms of locations, starting at the 4 seeds, we choose pod locations as well
the best 4 seed chooses his pod, then next best, etc
then the 3 seeds, and 2 seeds.

There is one caveat...if a pod location is already filled, you can "steal" a pod from a lower ranked team, white elephant style. That team then can steal from a lower ranked, or choose a still available location.

When you get to 1 seeds, you choose your regional as well as choosing/stealing your pod.


I would watch every bit of that as teams try to screw eachother stealing pods and choosing their opponents.

This would make for brilliant TV.
I can’t see anyone agreeing to this format- I like the creativity nonetheless.

Kedsy
02-18-2019, 04:43 PM
Also don't forget that recent history shows that Jon Beilein's teams also play their best basketball towards the end of the season (he's very K-like in that way).

I've enjoyed your NET analysis, and I understand you freely admit to being a Michigan homer but, come on, two seasons doesn't comprise a "recent history" from which we can draw any actual conclusions.

Below is a table showing Beilein's Michigan teams and their record before 2/1 vs. after (and including) 2/1:



Year After 2/1 Pct Before 2/1 Pct
2018 15 2 0.882 18 6 0.750
2017 12 4 0.750 14 8 0.636
2016 6 8 0.429 17 5 0.773
2015 3 8 0.273 13 8 0.619
2014 12 5 0.706 16 4 0.800
2013 11 7 0.611 20 1 0.952
2012 8 4 0.667 16 6 0.727
2011 8 5 0.615 13 9 0.591
2010 4 7 0.364 11 10 0.524
2009 7 6 0.538 14 8 0.636
2008 5 6 0.455 5 16 0.238


I get that the "before" columns include the cupcake November/December games, but still there's nothing in the above table that suggests Beilein's teams are historically better at the end of the season.

uh_no
02-18-2019, 04:43 PM
This would make for brilliant TV.
I can’t see anyone agreeing to this format- I like the creativity nonetheless.

Can you imagine coaches getting on air trying to convince K that Duke really doesn't want to face them and should go to another bracket?

House P
02-18-2019, 05:02 PM
Anybody want to defend Houston as a #1 seed?

I don't think Houston deserves a 1 seed, but the best case to be made probably involves comparing their Quad1/Quad2 record vs the other contenders for a one seed. Based on this (http://barttorvik.com/teamsheets.php), Houston has the most Quad1/Quad2 wins and the second fewest Quad1/Quad2 losses (Nevada is the only top team with no Quad1/2 losses, but Nevada is also the only top 20 NET team with a Quad 3 loss).

Here is the combined Quad1/Quad2 records of the current top 12 teams in the NET rankings.



Team
Quad 1/Quad 2 Record


Duke
14-2


Gonzaga
10-2


Virginia
12-2


Houston
14-1


Kentucky
11-4


Tennessee
11-2


Michigan St
13-5


Michigan
14-3


UNC
10-5


Texas Tech
12-5


Nevada
8-0


Purdue
12-7



That being said, Houston's record also reflects what I think is the biggest limitation of the Quadrant system. Namely, in my mind there is far too big of a variance within Quadrant 1.

For example, a road win vs Tulsa (#74 NET, #108 KenPom), a neutral win vs Alabama (#50 NET, #56 KenPom) and a win @ Duke (#1 NET, #1 KenPom) all count as Quad 1 wins. According to KenPom, a top 10 team (with an AdjEM of 25) would have about an 80% chance of winning at Tulsa or on a neutral court vs Alabama, but would only have about a 15% of winning at Duke. No way that these games should be in the same quadrant.

What does this have to do with Houston? Houston is 4-1 vs Quad 1 teams, which gives them the second best winning percentage in the country vs Q1 teams (Duke currently has the best Q1 record at 7-1). Throw in a perfect 10-0 record vs Q2 and no Q3/4 losses and Houston's resume looks pretty good. However, Houston's resume looks less impressive when you consider the list of their Q1 games

Road vs BYU (#74 KenPom)
Road vs Temple (#75 KenPom)
Road vs Central Florida (#49 KenPom)
Home vs LSU (#21 KenPom)
Home vs Cincinnati (#31 KenPom)

According to KenPom, a top 5 team should be favored by 8-12 points in these games. So, it is not all that impressive that Houston has won four of these five games.

Mtn.Devil.91.92.01.10.15
02-18-2019, 05:30 PM
Anyone who at this point in the season says that all four #1 seeds are "locked" is ignoring the lessons of history.

Or etymology.

I would even be okay with "virtual lock."

**steams to himself over the way words no longer seem to mean anything anymore**

rasputin
02-18-2019, 05:34 PM
Or etymology.

I would even be okay with "virtual lock."

**steams to himself over the way words no longer seem to mean anything anymore**

Literally.

Nugget
02-18-2019, 05:41 PM
Very fair criticism of my statement. Let me make the following adjustment: I think the NET needs to better reward victories over high-quality opponents in comparison to overall SOS, particularly when it comes to the non-conference. MSU played a quantifiably better non-conference schedule than Michigan, for sure, but in no way did they have a better non-conference performance considering who amongst those teams they actually beat. The bottom of Michigan's non-conference schedule may have been weaker than MSU's, but they also won their marquee games. I would argue that wins over marquee opponents in the non-conference is more important than whether your "tune-up" games are against 100-200 NET teams or 200-300 NET teams, although this is a point upon which reasonable people can definitely disagree. I see your point about the cake, but I think that this isn't a binary choice... there is a middle ground where these two factors can be appropriately weighted so that they both are taken into account, but in the correct ratio.

This is one of many places where knowing more details into how the NET is calculated would be helpful, but alas that is not to be. However, considering seedings are not simply determined by the NET, but by human analysis of the NET and the corresponding team sheets, one would hope that this difference in non-conference resume would be reflected in the seedings by looking at the team sheets. But most prognosticators don't seem to think this will be the case, as MSU seems to continually be ranked ahead of Michigan on the 2-seed line (which would have a non-trivial affect on the location of their first weekend games, just like it did last year). That is discouraging.

P.S. Case and point regarding MSU/Michigan non-conference strength of schedule: Michigan had 7 Q4 non-conference games and 1 Q3, while MSU had 4 Q4 and 3 Q3. That seems to be the biggest factor in their relative strength of schedule. Now, do we really think that, for the top-tier teams, there is that much difference between a Q3 and Q4 game, especially when they're all at home (they were)? That differentiator is probably more important in determining the at-large teams as opposed to top of the bracket seeding.

As you note, the seedings aren't simply determined by the NET, but by humans. So, perhaps to ease your concern, remember that when the Committee did its early bracket a week ago they had Michigan #6 and Michigan St. #8, despite Michigan St.'s relative strength in the computer rankings.

JasonEvans
02-18-2019, 05:46 PM
Alright...hear me out...here's how bracketing should go...this would be incredible must-watch TV and wouldn't involve stretching the "bracket reveal" over 19 hours like CBS does.


-populate the bottom 32 teams in the bracket by s curve
-starting with the 8 seeds, best to worst, they each choose which bracket they want to go in
-then the 7 seeds, best to worst
-...
-finally the 1 seeds choose which bracket they want, meaning the #4 overall team automatically gets the last remaining spot

In terms of locations, starting at the 4 seeds, we choose pod locations as well
the best 4 seed chooses his pod, then next best, etc
then the 3 seeds, and 2 seeds.

There is one caveat...if a pod location is already filled, you can "steal" a pod from a lower ranked team, white elephant style. That team then can steal from a lower ranked, or choose a still available location.

When you get to 1 seeds, you choose your regional as well as choosing/stealing your pod.


I would watch every bit of that as teams try to screw eachother stealing pods and choosing their opponents.

Not only would I watch this, but I would probably be willing to fork out $5-10 for the pay-pew-view version of it!!

Wahoo2000
02-18-2019, 05:46 PM
I looked this up a little while back, and if memory serves only the No. 1 overall seed is explicitly allowed to give the committee its preference, and everything else is done purely by distance. If that is still true it's more motivation to claim that top spot.

I believe it's actually like this: an undetermined number of teams who are "in the mix" for the top overall seed get to submit their preferences to the committee, but ONLY the #1 overall seed is guaranteed to GET their region of choice. Still, the committee has, in hand, the preference of I'd assume at LEAST a couple of other #1 seeds.

As others have stated in this thread, Duke is a *near* lock for a 1 seed, and I don't think any other team aside from Virginia would take DC as their preferred venue. And given the sweep of regular season games, the ONLY way I see UVa finishing ahead of Duke in the pecking order is if UVA wins out all the way through the ACC championship game, coupled with at least one more Duke loss during the regular season. Just seems an unlikely (not impossible, but pretty unlikely) scenario.

yancem
02-18-2019, 05:57 PM
Alright...hear me out...here's how bracketing should go...this would be incredible must-watch TV and wouldn't involve stretching the "bracket reveal" over 19 hours like CBS does.


-populate the bottom 32 teams in the bracket by s curve
-starting with the 8 seeds, best to worst, they each choose which bracket they want to go in
-then the 7 seeds, best to worst
-...
-finally the 1 seeds choose which bracket they want, meaning the #4 overall team automatically gets the last remaining spot

In terms of locations, starting at the 4 seeds, we choose pod locations as well
the best 4 seed chooses his pod, then next best, etc
then the 3 seeds, and 2 seeds.

There is one caveat...if a pod location is already filled, you can "steal" a pod from a lower ranked team, white elephant style. That team then can steal from a lower ranked, or choose a still available location.

When you get to 1 seeds, you choose your regional as well as choosing/stealing your pod.


I would watch every bit of that as teams try to screw eachother stealing pods and choosing their opponents.

That would be awesome! Great drama and suspense.

scottdude8
02-18-2019, 07:07 PM
I've enjoyed your NET analysis, and I understand you freely admit to being a Michigan homer but, come on, two seasons doesn't comprise a "recent history" from which we can draw any actual conclusions.

Below is a table showing Beilein's Michigan teams and their record before 2/1 vs. after (and including) 2/1:



Year After 2/1 Pct Before 2/1 Pct
2018 15 2 0.882 18 6 0.750
2017 12 4 0.750 14 8 0.636
2016 6 8 0.429 17 5 0.773
2015 3 8 0.273 13 8 0.619
2014 12 5 0.706 16 4 0.800
2013 11 7 0.611 20 1 0.952
2012 8 4 0.667 16 6 0.727
2011 8 5 0.615 13 9 0.591
2010 4 7 0.364 11 10 0.524
2009 7 6 0.538 14 8 0.636
2008 5 6 0.455 5 16 0.238


I get that the "before" columns include the cupcake November/December games, but still there's nothing in the above table that suggests Beilein's teams are historically better at the end of the season.

That’s a totally fair point. I think any analysis might require a finer touch than this considering how schedules tend to be backloaded, though.

Regardless, maybe I should’ve said “the last two years Michigan has been fantastic down the stretch” instead of trying to make a larger point. Ironically I have to tone down my tendency to do that in my research all the time, lol.

Bluedog
02-18-2019, 10:28 PM
As you note, the seedings aren't simply determined by the NET, but by humans. So, perhaps to ease your concern, remember that when the Committee did its early bracket a week ago they had Michigan #6 and Michigan St. #8, despite Michigan St.'s relative strength in the computer rankings.

Yeah, just to add to this, even by scottdude8's own analysis of the team sheets, it's clear the committee DOES care a lot about wins and losses instead of a teams "raw" NET rank. So, the efficiency ratings are pretty decent at determining quality of opponent and then the committee looks at how teams did against that quality from a W/L perspective. Seems fair to me and does emphasize wins accordingly. It's not just taking kenpom ratings. It's a far improvement over RPI (which also looked at W/L over teams, but was based on a terrible formula).

ice-9
02-19-2019, 12:39 AM
I don't think Houston deserves a 1 seed, but the best case to be made probably involves comparing their Quad1/Quad2 record vs the other contenders for a one seed. Based on this (http://barttorvik.com/teamsheets.php), Houston has the most Quad1/Quad2 wins and the second fewest Quad1/Quad2 losses (Nevada is the only top team with no Quad1/2 losses, but Nevada is also the only top 20 NET team with a Quad 3 loss).

Here is the combined Quad1/Quad2 records of the current top 12 teams in the NET rankings.



Team
Quad 1/Quad 2 Record


Duke
14-2


Gonzaga
10-2


Virginia
12-2


Houston
14-1


Kentucky
11-4


Tennessee
11-2


Michigan St
13-5


Michigan
14-3


UNC
10-5


Texas Tech
12-5


Nevada
8-0


Purdue
12-7



That being said, Houston's record also reflects what I think is the biggest limitation of the Quadrant system. Namely, in my mind there is far too big of a variance within Quadrant 1.

For example, a road win vs Tulsa (#74 NET, #108 KenPom), a neutral win vs Alabama (#50 NET, #56 KenPom) and a win @ Duke (#1 NET, #1 KenPom) all count as Quad 1 wins. According to KenPom, a top 10 team (with an AdjEM of 25) would have about an 80% chance of winning at Tulsa or on a neutral court vs Alabama, but would only have about a 15% of winning at Duke. No way that these games should be in the same quadrant.

What does this have to do with Houston? Houston is 4-1 vs Quad 1 teams, which gives them the second best winning percentage in the country vs Q1 teams (Duke currently has the best Q1 record at 7-1). Throw in a perfect 10-0 record vs Q2 and no Q3/4 losses and Houston's resume looks pretty good. However, Houston's resume looks less impressive when you consider the list of their Q1 games

Road vs BYU (#74 KenPom)
Road vs Temple (#75 KenPom)
Road vs Central Florida (#49 KenPom)
Home vs LSU (#21 KenPom)
Home vs Cincinnati (#31 KenPom)

According to KenPom, a top 5 team should be favored by 8-12 points in these games. So, it is not all that impressive that Houston has won four of these five games.


This whole quadrant thing is so stupid. I get why they had it before, because they didn’t really trust the RPI and wanted a different way of assessing win-loss records.

But with NET they have a solid enough foundation of ranking teams, and baked into NET is strength of schedule.

So this whole quadrant business...not only is it wildly misleading as shown above, it’s also double counting.

English
02-19-2019, 01:51 PM
Yeah, just to add to this, even by scottdude8's own analysis of the team sheets, it's clear the committee DOES care a lot about wins and losses instead of a teams "raw" NET rank. So, the efficiency ratings are pretty decent at determining quality of opponent and then the committee looks at how teams did against that quality from a W/L perspective. Seems fair to me and does emphasize wins accordingly. It's not just taking kenpom ratings. It's a far improvement over RPI (which also looked at W/L over teams, but was based on a terrible formula).

One thing that I've heard some smart people hypothesizing, is how the Committee will use/emphasize the NET in its first year as a means of "validating" it, so to speak. That is, if a team like Wofford or Buffalo or some other highly-ranked (WRT NET), little-known/respected team ends up on the bubble, does the Committee put them in wherein years past, they'd likely be left out in favor of a middling P5 team? They'd basically have to, or else risk the masses saying "wait, you have this algorithm and you're not even using it? It must be broken." That also applies to seeding this year. I wonder how closely the Committee actually sticks with the NET rankings, at the behest of the NCAA, to confirm its utility. I don't really have a guess or a recommendation beyond to simply say I think it should be a tool, not the tool.


This whole quadrant thing is so stupid. I get why they had it before, because they didn’t really trust the RPI and wanted a different way of assessing win-loss records.

But with NET they have a solid enough foundation of ranking teams, and baked into NET is strength of schedule.

So this whole quadrant business...not only is it wildly misleading as shown above, it’s also double counting.

I agree, breaking quadrants into 50-team or 100-team groupings isn't particularly useful. I think the NCAA and Committee see that, too, and that's why they're including sub-quads on the team sheets. It's clear to me that the concept is to provide a uniform way to evaluate teams against very uneven schedules, and in that sense, I see the benefit. Instead of leaving it entirely up to each member of the selection committee to decide how they want to evaluate each team's resume (i.e., what's most important--W/L, SOS, venue of W/Ls, AdjEff, Last 10 gms, etc.), this at least gives things a defined structure from which to approach selection/seeding. Again, a tool, not the tool.

HereBeforeCoachK
02-19-2019, 02:24 PM
This whole quadrant thing is so stupid. I get why they had it before, because they didn’t really trust the RPI and wanted a different way of assessing win-loss records.

But with NET they have a solid enough foundation of ranking teams, and baked into NET is strength of schedule.

So this whole quadrant business...not only is it wildly misleading as shown above, it’s also double counting.

Agree with all of this.....the Quad thing has always been so absurdly NON precise.....especially in a day and age of very precise and advanced metrics. This is like shutting down a super computer and pulling out a damned abacus.....

scottdude8
02-19-2019, 02:25 PM
I agree, breaking quadrants into 50-team or 100-team groupings isn't particularly useful. I think the NCAA and Committee see that, too, and that's why they're including sub-quads on the team sheets. It's clear to me that the concept is to provide a uniform way to evaluate teams against very uneven schedules, and in that sense, I see the benefit. Instead of leaving it entirely up to each member of the selection committee to decide how they want to evaluate each team's resume (i.e., what's most important--W/L, SOS, venue of W/Ls, AdjEff, Last 10 gms, etc.), this at least gives things a defined structure from which to approach selection/seeding. Again, a tool, not the tool.

I agree that the quadrant system is a positive step compared to having nothing of the sort, but it can get dangerous when applied blindly... this has been one of my big points with regards to SOS and a differentiation between Q3 and Q4 teams, for example. I think that having a uniform way of saying this team had "x" number of "this quality" wins is extremely useful when there is a quantitative, as opposed to qualitative, definition of "quality" (you can argue about whether or not that quantitative definition is valid or not, but as I've learned in the academic world, you need a shared set of "first principles" in order to have any sort of reasonable debate). Now, where there needs to be some nuance is in how that information is used.

For instance, if you're deciding between two teams for a final at large spot, it's entirely possible that they had similar struggles against "top-tier" opponents as defined by Q1 and Q2, but one team played more Q3 teams (and had more success against them), whereas the other team beat up on a bunch of Q4 teams. This is an important differentiator at this level of the bracket, and should be used (who's to say that team who played a ton of Q4 teams wouldn't have suffered a hiccup against a Q3 team that the former team risked and successfully avoided?). However, if you're looking at the top of the bracket, there really isn't much difference between a Q3 and a Q4 win (indeed, I'd argue that if you have a Q3 or Q4 loss, the chances of you being in competition for a top 2 or 3 seed are relatively low), but in this arena a road win against a top 10 NET team is much more impressive than a home win against a top 30 NET team, and that difference could/should play a role in differentiating two otherwise similar teams. So the more fine-scale differentiation is necessary.

All of this is to say the following: If the NET does a better job of ranking teams in a manner that agrees with reality than the RPI, it has already served a valuable purpose by better defining these quadrants. The NET rankings themselves can also do a good job of putting teams into approximate "tiers". However, to differentiate teams within a "tier" requires a more nuanced use not only of the rankings, but of the Team Sheets, and potentially weighing different factors in different manners depending on the tier you're looking at. There's never going to be a panacea for the ills of seeding the NCAA Tournament, but we can take steps closer and closer to having a well-understood and agreed upon set of "first principles" as to what matters in making these decisions. In this way the NET is a good first step forward compared to the RPI, and the improved Team Sheets are another. Now we need to see how the committee actually uses this information on Selection Sunday to see if these improvements will bear any fruit.

freshmanjs
02-19-2019, 04:39 PM
Not sure about that if the Zags, the Vols and/or the Wolverines go undefeated from here on....but my other point was, if all of those teams stumble, then Duke might have a 3 game pad.

It's even stronger than what I wrote yesterday. One of the following must be true of a 4 loss Duke team:

A) Duke wins at least 3 out of @VaTech, @Syracuse, UNC, @UNC OR
B) Duke wins the ACC tournament and wins at least 2 out of @VaTech, @Syracuse, UNC, @UNC

We should also keep in mind that, based on the other conference tournaments and current records:

There will be at best:

B1G: 4 loss Michigan or 6 loss Michigan State
SEC: 3 loss Tennessee or 5 loss Kentucly

JasonEvans
02-20-2019, 11:34 AM
Houston is 4-1 vs Quad 1 teams, which gives them the second best winning percentage in the country vs Q1 teams (Duke currently has the best Q1 record at 7-1).

Was looking at this post again and wanted to ask folks a question...

Is win percentage more important than volume of wins? For example, a team that is 4-1 vs Q1 teams... is that better than a team that is 7-2 versus Q1 teams? How about a team that is 11-3? I think a lot of weight should be given to teams that play a tough schedule. If you only play 5 Q1 games (and none against teams in the top half of Q1), I think you are sorta disqualified from consideration for a #1 seed.

-Jason "that said, I recognize the limitations some teams have in their conference opponents... but you still gotta play some decent teams to get a top seed" Evans

CDu
02-20-2019, 11:40 AM
Was looking at this post again and wanted to ask folks a question...

Is win percentage more important than volume of wins? For example, a team that is 4-1 vs Q1 teams... is that better than a team that is 7-2 versus Q1 teams? How about a team that is 11-3? I think a lot of weight should be given to teams that play a tough schedule. If you only play 5 Q1 games (and none against teams in the top half of Q1), I think you are sorta disqualified from consideration for a #1 seed.

-Jason "that said, I recognize the limitations some teams have in their conference opponents... but you still gotta play some decent teams to get a top seed" Evans

Are you looking for opinion, or how you think it happens in practice?

I guess my answer in either case is "it's subjective." If you are undefeated on the season and play just 3 Q1 teams, but they are the #2, #3, and #4 teams in the country and you beat them all by 15+ points? You probably deserve a 1 seed. If you are 4-1 against back end of Q1, with a bunch of close wins against Q2 and Q3? You probably don't deserve it.

As for how the committee will use it? Probably only as tiebreakers. In other words, I imagine they'll stick more closely to the NET rank, and adjust on a seed line (or bump around a seed or so) based on that as a tiebreaker. And they'll use whatever tiebreaker the committee comes to a consensus on when comparing two similarly-rated (by NET) teams. And that may or may not be based on the Q1 record.

I think folks get too tied into absolutes in these discussions, when in practice there is a ton of subjectivity that goes into seedings. Trying to find a consistent theme in all of this is a fool's errand. Each committee is different, and even within a single year each decision is different.

Hingeknocker
02-20-2019, 11:42 AM
Was looking at this post again and wanted to ask folks a question...

Is win percentage more important than volume of wins? For example, a team that is 4-1 vs Q1 teams... is that better than a team that is 7-2 versus Q1 teams? How about a team that is 11-3? I think a lot of weight should be given to teams that play a tough schedule. If you only play 5 Q1 games (and none against teams in the top half of Q1), I think you are sorta disqualified from consideration for a #1 seed.

-Jason "that said, I recognize the limitations some teams have in their conference opponents... but you still gotta play some decent teams to get a top seed" Evans

If it's going to be the case that Quadrants matter, then I think that volume should be much more important than win percentage. I personally don't think Quadrants should matter at all - they are arbitrary categories that hinder, rather than help the overall understanding of a team's quality. But, accepting the fact they do matter, it seems logical that playing more games against Q1 teams would be a better thing.

I say this based on a reading of how the committee discusses strength of schedule and the quality of resume when they select/seed teams. They seem to be impressed by "volume" metrics when it comes to Q1 victories.

AGDukesky
02-20-2019, 11:42 AM
Was looking at this post again and wanted to ask folks a question...

Is win percentage more important than volume of wins? For example, a team that is 4-1 vs Q1 teams... is that better than a team that is 7-2 versus Q1 teams? How about a team that is 11-3? I think a lot of weight should be given to teams that play a tough schedule. If you only play 5 Q1 games (and none against teams in the top half of Q1), I think you are sorta disqualified from consideration for a #1 seed.

-Jason "that said, I recognize the limitations some teams have in their conference opponents... but you still gotta play some decent teams to get a top seed" Evans

I might go with number of games over .500. 7-2 is definitely better than 4-1 but 6-6 is not. At some point the sheer volume of games is more impressive assuming you win more than you lose.

scottdude8
02-20-2019, 01:01 PM
Was looking at this post again and wanted to ask folks a question...

Is win percentage more important than volume of wins? For example, a team that is 4-1 vs Q1 teams... is that better than a team that is 7-2 versus Q1 teams? How about a team that is 11-3? I think a lot of weight should be given to teams that play a tough schedule. If you only play 5 Q1 games (and none against teams in the top half of Q1), I think you are sorta disqualified from consideration for a #1 seed.

-Jason "that said, I recognize the limitations some teams have in their conference opponents... but you still gotta play some decent teams to get a top seed" Evans

I think that this, just like all these quantifications, needs to be used depending on the situation. If you're trying to determine whether a team deserves a No. 1 seed, that team should have enough of a sample size against similar, top-tier opponents (i.e. Q1 games) to prove that they're one of the best four teams in college basketball. I don't think 5 games is enough to determine that (it's actually for this reason I'm surprised Gonzaga has earned "lock No. 1 seed" status in so many peoples eyes... they're non-conference performance was not nearly as good this year as it has been in past seasons where they earned a top seed). However, if you're thinking about a 2,3,4 seed, then there's an argument to be had... I might be swayed by the argument that a team who went 4-1 vs. the Q1 would conceivably have a better record against the Q1 if they had more opportunities then, say, a major conference team that went 5-5 in such games.

All of this keeps going back to the point that if the committee were smart they wouldn't go for sweeping, one-size-fits-all ways of ranking/seeding the teams, but rather a more nuanced perspective recognizing that different features of a team's resume are more/less important depending on what type of teams you are comparing them against.

uh_no
02-20-2019, 01:04 PM
All of this keeps going back to the point that if the committee were smart they wouldn't go for sweeping, one-size-fits-all ways of ranking/seeding the teams, but rather a more nuanced perspective recognizing that different features of a team's resume are more/less important depending on what type of teams you are comparing them against.

I mean that's what they DO do. NET and RPI are just tools the committee have at their disposal. How much weight they put in them is up to them. In the end, they may rely on those rankings more than we would like, but it's never been the case that the bracket is simply RPI order....especially when talking about the top seed line.