Seeding Outlook

[caw]

Looks like the majority of sites have them as a 4 but it's close. Average seed is the highest 4 so real close to a 3.

The Bracket Project's Bracket Matrix - 2023

There was pretty significant movement from yesterday, I think when I went to that site before the game there were maybe five brackets with UConn as a three and 1-2 with UConn as a 5. Now there are about 17 with UConn as a three, zero with UConn as a 5 and 1 with UConn as a 2. I know that website was updated today but I am unclear on how often the websites it aggregates are updated.

Either way, seems to be about 1/3 UConn 2/3 UT for the final three seed.

 

[caw]

Quadrants are dumb and frankly should go away. They should just do like they did with RPI with the NET and break it down v. Top 25, Top 50, and Top 100, and sub 100, and then break them down Home, Road, Neutral. It's just more informative. Some Q1 games are not vs. likely tournament teams, but most Top 50 are. That's just more useful.

The quad **** is frankly more confusing than helpful. A good team should never lost to the 135th best team yet somehow at home that's Q2 (meh) and on the road Q3 (bad).

Agreed, just look at Houston's resume.

Q1 wins - @Oregon, @ UVA, Neutral SMC, @ Memphis
Q2 wins - Oral Roberts, Kent State, @ Cincy, @ Tulane, @ UCF, @ Wichita St, @ Temple, Memphis.

That is some garbage aside from @ UVA and Neutral to SMC. Seriously debatable if any of those teams get an at large, Memphis is a bubble team and probably a 10 seed right now I guess.


I mean that just doesn't scream #1 seed.

 

[tzznandrew]

Agreed, just look at Houston's resume.

Q1 wins - @Oregon, @ UVA, Neutral SMC, @ Memphis
Q2 wins - Oral Roberts, Kent State, @ Cincy, @ Tulane, @ UCF, @ Wichita St, @ Temple, Memphis.

That is some garbage aside from @ UVA and Neutral to SMC. Seriously debatable if any of those teams get an at large, Memphis is a bubble team and probably a 10 seed right now I guess.


I mean that just doesn't scream #1 seed.

Break it down the old way:

v. Top 25: n-SMC
v. Top 50: @ UVA, Memphis, @ Memphis, @ Oregon, Oral Roberts
v. Top 100: Kent State, @ Cincinnati, Cincinnati, Tulane, @ Tulane, UCF, @ UCF

Tulane will potentially fall out (99 in NET).

That's a better snapshot of their resume, right?

That said...this is a very good team. They were E8 last year, Final Four the year before, and S16 the year before the pandemic. A real battle tested team with a coach I really respect. Would prefer not to see them until the E8.

 

[auror]

Break it down the old way:

v. Top 25: n-SMC
v. Top 50: @ UVA, Memphis, @ Memphis, @ Oregon, Oral Roberts
v. Top 100: Kent State, @ Cincinnati, Cincinnati, Tulane, @ Tulane, UCF, @ UCF

Tulane will potentially fall out (99 in NET).

That's a better snapshot of their resume, right?

That said...this is a very good team. They were E8 last year, Final Four the year before, and S16 the year before the pandemic. A real battle tested team with a coach I really respect. Would prefer not to see them until the E8.

It's not better, no. Because a win against a 50-75 team in the road is harder than a win against a 25-50 team at home That formating implies otherwise.

 

[ElGreco]

Am I the only one that thinks if we win the BET and other things fall our way we could be a 1 seed? I understand it may take a bunch of dominoes, but the parity is incredible this year.

Never ever gonna happen this year even though everyone and the metrics love us. Kansas lost yesterday and they moved to Ted overall one

 

[tzznandrew]

It's not better, no. Because a win against a 50-75 team in the road is harder than a win against a 25-50 team at home That formating implies otherwise.

I think winning against a tournament team anywhere is better and more valuable than winning against a non-tournament team.

And I think breaking down the H/R/N of any win in 25/50/100 sets gives you enough context to draw the appropriate conclusion.

I don't think there's an appreciable or useful difference between some of the Q2/Q3 distinctions.

 

[uconnbill]

The committee places zero value on how you finish.

How can you think that? I have always been told that a hot team does get a bump

 

[auror]

I think winning against a tournament team anywhere is better and more valuable than winning against a non-tournament team.

And I think breaking down the H/R/N of any win in 25/50/100 sets gives you enough context to draw the appropriate conclusion.

I don't think there's an appreciable or useful difference between some of the Q2/Q3 distinctions.

If it's easier to beat a projected 10 seed at home than a bubble team on the road, why is that a more valuable win? Winning on the road is really hard, and the old top 50, etc. system did not do justice to road wins against good but not great teams. UConn's win @Nova would have just been a really mediocre top 100 win in your system. In quads, we know it's Q2 borderline Q1. Adding the @ context in a top 50 list means the person analyzing has to mentally do the adjustment on how difficult the win is. The simplicity of the quad system is that the adjustment is baked in.

We're not just making up these "difficulties", either. They're empirically derived based on historical data. The quadrant themselves are arbitrary (based on quad 1 being beating a top 30 team at home equivalency, quad 2 being top 75 at home, and expanding from there), but the cutoffs of the quadrants are really the derived equivalent home/road/neutral ranks.

At some point the winning % for tournament teams becomes like 95%+ against Q3 and Q4 teams, but that's why when most people analyze resumes they just ignore those quads and only focus on the bad losses. Just like you would with "200+ wins" or whatever. But yes, Q2 wins are all more difficult than Q3 wins (once you're confident enough in the sample size, but the same is true for top X lists).

 

[businesslawyer]

I think winning against a tournament team anywhere is better and more valuable than winning against a non-tournament team.

And I think breaking down the H/R/N of any win in 25/50/100 sets gives you enough context to draw the appropriate conclusion.

I don't think there's an appreciable or useful difference between some of the Q2/Q3 distinctions.

And there is between the 49th and 51st ranked teams is useful? You have to draw lines somewhere if you are going to group team. Where you draw them is totally arbitrary. (But looking at whether games were home or away is absolutely appropriate.)

 

[shizzle787]

How can you think that? I have always been told that a hot team does get a bump

They got rid of that a few years ago.

 

[ToolzieFan]

Re: quadrants are dumb.

 

[auror]

Re: quadrants are dumb.

Well maybe our win at Nova will get some extra consideration lol (and also there's good chance it becomes quad 1 by Sunday). I like Seth Burn a lot. He's the pioneer of WAB, which I think is the best resume metric. But from what I can tell this is a 1 year sample size with a clickbaity tagline (the fully explained stat is "47% of Q1 games are easier than one outlier Q2 game"). He's trying to get mainstream adoption of WAB and trying to go viral on this to spread some awareness.

In which case, the problem isn't the quad system, but that the NET fails to properly value some outlier teams, especially ones with multiple sustained injuries.

If anything, I'd say the graphic validates the quadrants. There are some very clear outliers, but otherwise, In a one year sample, 90% of the blue is more difficult than 90% of the red.

 

[Rico444]

What confuses me about the NET is that people say it's not comprehensive enough to really matter when it comes to seeding (we're 6 in the country but most brackets have us as a 4). However, the main thing we use to determine seeding is the quadrant victories, which use the NET to sort teams.

 

[tzznandrew]

And there is between the 49th and 51st ranked teams is useful? You have to draw lines somewhere if you are going to group team. Where you draw them is totally arbitrary. (But looking at whether games were home or away is absolutely appropriate.)

of course making a distinction somewhere is arbitrary, I just think the Quadrant stuff, which treats playing the #1 team on the road the same as the #70 team, does a worse job getting a snapshot of a team's performance than just breaking it down by number and then by h/r/n. Beating Houston, Kansas St. and Xavier on the road—or at home, frankly—is just more impressive than winning on the road against VCU, Kent State, and Washington State.

I think it's also just more honest about a teams accomplishments—easier for a casual fan to grasp—when presented like :

TEAM A
Wins v. NET Top 25: 3

Home: 1 Houston​

Road: 17 Kansas State, 22 Xavier​

TEAM B
Wins v. NET Top 25: 0
Wins v. NET Top 50: 0
Wins v. NET Top 100: 3

Road: 67 VCU, 69 Kent State, 70 Washington State​

Rather than

TEAM A
Q1 Wins: 3

H-1 Houston, @ 17 Kansas State, @ 22 Xavier​

TEAM B
Q1 Wins: 3

@ 67 VCU, @ 69 Kent State, 70 Washington​

 

[businesslawyer]

of course making a distinction somewhere is arbitrary, I just think the Quadrant stuff, which treats playing the #1 team on the road the same as the #70 team, does a worse job getting a snapshot of a team's performance than just breaking it down by number and then by h/r/n. Beating Houston, Kansas St. and Xavier on the road—or at home, frankly—is just more impressive than winning on the road against VCU, Kent State, and Washington State.

I think it's also just more honest about a teams accomplishments—easier for a casual fan to grasp—when presented like :

TEAM A
Wins v. NET Top 25: 3

Home: 1 Houston​

Road: 17 Kansas State, 22 Xavier​

TEAM B
Wins v. NET Top 25: 0
Wins v. NET Top 50: 0
Wins v. NET Top 100: 3

Road: 67 VCU, 69 Kent State, 70 Washington State​

Rather than

TEAM A
Q1 Wins: 3

H-1 Houston, @ 17 Kansas State, @ 22 Xavier​

TEAM B
Q1 Wins: 3

@ 67 VCU, @ 69 Kent State, 70 Washington​

The point that your example makes is not that grouping by NET rankings is better or worse than using quadrants. The point your are making, which I’ve said repeatedly before, is that using groupings is only a starting point. Once you’ve used it to group teams that may be close to each other, you then need to look at wins and losses individually and compare. Which by all reports the Committee does.

 

[UConNick]

A 3-seed in Bennett’s latest bracket on the Athletic, popping up just in front of Tennessee (who’s his highest 4).

We’ve also drawn ever closer to Tennessee on bracketmatrix, Tenn’s average seed there a 3.22 while we’re a 3.68. And that’s only still taking into account brackets released as of yesterday, I believe. That gap will narrow, I bet, as today’s brackets are factored in.

I’m not one who thinks we can get all the way to a 2, but a path to the 3-seed is forming. We know Saturday won’t matter to the committee (though it should. I do think conference tourney champions should receive a slight boost for winning that, even if it’s just re-evaluating their metrics factoring in 3-4 more quality wins). But I think getting past PC and topping Marquette would be enough to solidify us as a 3. I would lean towards a 4 being our floor if PC got us, but as I’ve said numerous times, I don’t trust the committee.

 

[pj]

What confuses me about the NET is that people say it's not comprehensive enough to really matter when it comes to seeding (we're 6 in the country but most brackets have us as a 4). However, the main thing we use to determine seeding is the quadrant victories, which use the NET to sort teams.

Yes, but there is a problem in that the way the NET is calculated aggravates the problem with quadrant scoring. NET is already rigged to reward strength of schedule (the two factors in NET are "the Team Value Index (TVI), which is a result-based feature that rewards teams for beating quality opponents, particularly away from home, as well as an adjusted net efficiency rating" -- College basketball's NET rankings, explained -- net efficiency rating is similar to Kenpom and then they bias it to reward strength of schedule). Then the quadrant scores work from NET, and quadrant scoring is rigged to favor schedules with high-quadrant opponents.

So you have a compounding of the favoritism to strength of schedule, which rewards teams from strong conferences. This year it favors the Big 12.

Wouldn't be surprising if Kansas is a #1 seed and Houston a #2 due to this effect. UCLA also underrated by quadrant scores I think.

 

[pj]

Re: quadrants are dumb.

Great tweet! Interesting that winning at Villanova is the toughest Q2 win.

 

[auror]

What confuses me about the NET is that people say it's not comprehensive enough to really matter when it comes to seeding (we're 6 in the country but most brackets have us as a 4). However, the main thing we use to determine seeding is the quadrant victories, which use the NET to sort teams.

I do not believe the quadrant system is the main thing used to determine seeding. From many reports of people who should know or are told, the committee dives a lot deeper into individual resumes (and members personal eye test) than just relying on the metrics or quads.

So the quads are what the horse race bracketologists and amateur forum goers rely on, but not as much the committee. But I'm sure they're used to some extent, I would just say definitely not the most important thing. They're a useful distillation of information, especially since as a range there is a margin of error built-in (only the values at the extremes are subject to really being wrong).

 

[navery12]

Not much of note tonight that can affect UConn. New Mexico plays Wyoming and is a 9 point favorite on KenPom. They're at #50 in the NET rankings so a bad performance from them could push Providence back up to 50 and make Thursday a Q1 game again

 

[HuskyHoosier]

We can root for the zags and st Mary’s to lose and give us a better chance at zags 2 or eliminate st Mary’s from competing for a 3.

 

[SubbaBub]

Top 4 in the east is all that matters, Top 5 in the East would be fine, if we are being honest. Being a 3 (or 2) gives us a slightly better Rd of 16 game assuming we aren't doomed.

 

[Skies2018]

Agreed, just look at Houston's resume.

Q1 wins - @Oregon, @ UVA, Neutral SMC, @ Memphis
Q2 wins - Oral Roberts, Kent State, @ Cincy, @ Tulane, @ UCF, @ Wichita St, @ Temple, Memphis.

That is some garbage aside from @ UVA and Neutral to SMC. Seriously debatable if any of those teams get an at large, Memphis is a bubble team and probably a 10 seed right now I guess.


I mean that just doesn't scream #1 seed.

Houston is overrated. They played a Mickey Mouse schedule all year. They’re gonna burn out in the Sweet 16 is my prediction.

 

[navery12]

 

[UCFan1513]

Tonight we are all San Francisco Dons