Seeding Outlook | Page 3 | The Boneyard

Seeding Outlook

tzznandrew said:
First things first: get Albany as a 2/3 should be the goal. You play 14 seed rather than 13. In the second round, the 6/11 teams are always preferable to the 5/12—and we avoid the 1 even longer.

It's the S-curve after that, so more luck. Want to be slotted opposed to whichever 1-seed is closest east. Seems most likely to be Purdue.

But at this stage, since we're not in discussion for a 1, the goal has to be Albany with the 1 on the other side of the bracket.
Expand ...
Under-discussed and astute point.

Forget all the MSG talk (which will be at the whims of the Committee). A #3 seed puts us in a much, much better position to get to the Final Four than a #4 seed.

  • #13 seeds tend to be a lot more game than #14
  • though it isn't always the case, this year there is a significant dropoff after the first ~20 or so teams. The 6 seed line is weak this year.
  • likewise, though it isn't always the case, this year there is a pretty significant dropoff after the first handful of teams. Facing 2-seed Kansas State is a lot more appealing than #1 seed Kansas in the Sweet 16 if we get there
 
Tenspro2002 said:
Under-discussed and astute point.

Forget all the MSG talk (which will be at the whims of the Committee). A #3 seed puts us in a much, much better position to get to the Final Four than a #4 seed.

  • #13 seeds tend to be a lot more game than #14
  • though it isn't always the case, this year there is a significant dropoff after the first ~20 or so teams. The 6 seed line is weak this year.
  • likewise, though it isn't always the case, this year there is a pretty significant dropoff after the first handful of teams. Facing 2-seed Kansas State is a lot more appealing than #1 seed Kansas in the Sweet 16 if we get there
Expand ...
Yes, all but a hand full of national championships have been won by top 3 seeds.
 
CTBasketball said:
We are not jumping to the 2 line. Pray for a BET title and a 3 seed in the East.
Expand ...
I don't think we can get a 2 either now that I look at it: we can't pass Kansas, Texas, Kansas State, UCLA, Purdue, or Baylor based on Quad victories/losses. It is highly unlikely we pass any one of the other two we need to get a Quad 3 loss: Houston or Alabama. Both would have to run into sub-100 NET teams and lose. I don't see that happening.
 
shizzle787 said:
I don't think we can get a 2 either now that I look at it: we can't pass Kansas, Texas, Kansas State, UCLA, Purdue, or Baylor based on Quad victories/losses. It is highly unlikely we pass any one of the other two we need to get a Quad 3 loss: Houston or Alabama. Both would have to run into sub-100 NET teams and lose. I don't see that happening.
Expand ...
There's a very outside chance of the 2. The 3 is our goal. It's gettable. We may be sittin on one right now. But of course we need to hold on.
 
.-.
tzznandrew said:
There's a very outside chance of the 2. The 3 is our goal. It's gettable. We may be sittin on one right now.
Expand ...
I think we are #12 on the S-Curve ahead of Tennessee (they have 3 Q2 losses to our 1 Q3) and slightly behind Gonzaga.
 
shizzle787 said:
I don't think we can get a 2 either now that I look at it: we can't pass Kansas, Texas, Kansas State, UCLA, Purdue, or Baylor based on Quad victories/losses. It is highly unlikely we pass any one of the other two we need to get a Quad 3 loss: Houston or Alabama. Both would have to run into sub-100 NET teams and lose. I don't see that happening.
Expand ...

I don't know how the committee is going to look at it, but that's not at all how I see it.

Just looking at the two weakest teams on your "we can't pass" list:
  • Kansas State has 8 losses and against shared opponents with UConn went 4-2 with losses to Butler and Iowa State and wins over Oklahoma St (twice), Iowa State, and Florida. UConn went 5-0 against those opponents. K State is not finishing strong, 6-6 in their last 12. They are Kenpom #18.
  • Baylor has 9 losses and against shared opponents with UConn went 0-3 with losses to Marquette and Iowa State (twice). UConn went 2-1 against those opponents. Baylor is 2-3 in their last 5. They are Kenpom #14.
I don't see how either team could be ranked ahead of UConn. We have a better overall record, look better against shared opponents, are Kenpom #4 and are finishing strong having won 8 of our last 9.
 
pj said:
I don't know how the committee is going to look at it, but that's not at all how I see it.

Just looking at the two weakest teams on your "we can't pass" list:
  • Kansas State has 8 losses and against shared opponents with UConn went 4-2 with losses to Butler and Iowa State and wins over Oklahoma St (twice), Iowa State, and Florida. UConn went 5-0 against those opponents. K State is not finishing strong, 6-6 in their last 12.
  • Baylor has 9 losses and against shared opponents with UConn went 0-3 with losses to Marquette and Iowa State (twice). UConn went 2-1 against those opponents. Baylor is 2-3 in their last 5.
I don't see how either team could be ranked ahead of UConn.
Expand ...
The committee places zero value on how you finish.

Kansas State has zero losses in Quad 3 or 4. They have one loss in Quad 2. If they lose to TCU, that will be a Quad 1 loss. They would finish at 9-8 in Quad 1.

If we win out, we would finish 8-6 in Quad 1 but we would have a Quad 3 loss. 9-8 vs. 8-6 is basically a tossup. The differentiator is having a Quad 2 loss (them) vs. a Quad 3 loss (us).

Baylor has zero losses in Quads 2, 3, or 4. They are 11-9 in Quad 1. If they lose to Iowa State they will be 11-10 in Quad 1. 11-10 in Quad 1 with no other losses is slightly better than 8-6 in Quad 1 with a Quad 3 loss.
 
  • Like
Reactions: caw
shizzle787 said:
The committee places zero value on how you finish.

Kansas State has zero losses in Quad 3 or 4. They have one loss in Quad 2. If they lose to TCU, that will be a Quad 1 loss. They would finish at 9-8 in Quad 1.

If we win out, we would finish 8-6 in Quad 1 but we would have a Quad 3 loss. 9-8 vs. 8-6 is basically a tossup. The differentiator is having a Quad 2 loss (them) vs. a Quad 3 loss (us).

Baylor has zero losses in Quads 2, 3, or 4. They are 11-9 in Quad 1. If they lose to Iowa State they will be 11-10 in Quad 1. 11-10 in Quad 1 with no other losses is slightly better than 8-6 in Quad 1 with a Quad 3 loss.
Expand ...
I think we may be ahead of KSU right now. We may hop Baylor. We may hop one merely on procedural grounds, too, given how clumped up the BXII is.

If we win out, we're a 3 at least. Other things would need to shake out for us to get a 2.

A 3 is fine. We've won a title from there. We can start in Albany from there.
 
I’m all for better seeding, the easier the opponent in theory.

Once the get the monkey off their back in the first round, we can exhale and enjoy the games a bit more.
 
.-.
shizzle787 said:
The committee places zero value on how you finish.

Kansas State has zero losses in Quad 3 or 4. They have one loss in Quad 2. If they lose to TCU, that will be a Quad 1 loss. They would finish at 9-8 in Quad 1.

If we win out, we would finish 8-6 in Quad 1 but we would have a Quad 3 loss. 9-8 vs. 8-6 is basically a tossup. The differentiator is having a Quad 2 loss (them) vs. a Quad 3 loss (us).

Baylor has zero losses in Quads 2, 3, or 4. They are 11-9 in Quad 1. If they lose to Iowa State they will be 11-10 in Quad 1. 11-10 in Quad 1 with no other losses is slightly better than 8-6 in Quad 1 with a Quad 3 loss.
Expand ...

This worship of the Quadrant scores strikes me as idiocy. By quadrant score we are #13 nationally and on the 4 line, but if we had won one more game we'd be #7 nationally and on the 2 line. Our last three losses came by 3 points or less and one could easily have swung the other way.

Moreover, the quadrant scores really reward teams for playing in a conference with a lot of top 50 teams. If you play all Q1 teams and go .500, your quadrant score is 1.50 and you would be ranked the #16 team in the country, ahead of Virginia, St Mary's, Tennessee, Duke, Kentucky, and many others, and get a 4 seed. Whereas if you play all Q4 teams and go undefeated, your quadrant score is 1.00 and you are ranked #42 and are on the bubble or an 11 seed.

By the eye test, UConn is better than a 4 seed, and better than K State or Baylor.
 
pj said:
This worship of the Quadrant scores strikes me as idiocy. By quadrant score we are #13 nationally and on the 4 line, but if we had won one more game we'd be #7 nationally and on the 2 line. Our last three losses came by 3 points or less and one could easily have swung the other way.

Moreover, the quadrant scores really reward teams for playing in a conference with a lot of top 50 teams. If you play all Q1 teams and go .500, your quadrant score is 1.50 and you would be ranked the #16 team in the country, ahead of Virginia, St Mary's, Tennessee, Duke, Kentucky, and many others, and get a 4 seed. Whereas if you play all Q4 teams and go undefeated, your quadrant score is 1.00 and you are ranked #42 and are on the bubble or an 11 seed. ... Since Q1 is basically the top 50 teams, going 50-50 against Q1 should rank you around #25 and a 7 seed, not #16 and a 4 seed.

By the eye test, UConn is better than a 4 seed, and better than K State or Baylor.
Expand ...
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).
 
HooperScooper said:
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.

Expand ...

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.
 
tzznandrew said:
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).
Expand ...
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.
 
caw said:
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.
Expand ...
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.
 
.-.
tzznandrew said:
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.
Expand ...
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.
 
CardiacAndre7 said:
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.
Expand ...
Never ever gonna happen this year even though everyone and the metrics love us. Kansas lost yesterday and they moved to Ted overall one :) ;)
 
auror said:
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.
Expand ...
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.
 
Last edited:
tzznandrew said:
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.
Expand ...
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).
 
Last edited:
tzznandrew said:
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.
Expand ...
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.)
 
.-.
ToolzieFan said:
Re: quadrants are dumb.

Expand ...

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.
 
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.
 
businesslawyer said:
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.)
Expand ...
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​
 
tzznandrew said:
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​
Expand ...
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.
 
.-.
You must log in or register to reply here.

Forum statistics

Threads
168,207
Messages
4,556,897
Members
10,442
Latest member
Virginiafan


Top Bottom