What drives the Quad System calculations?

[JonathanW]

As we all know, that USC loss is like a lead weight hanging around UK's neck. So I was looking at why is USC a Q4 loss?

So USC sits at 277, going 1-2 vs Q1, 1-3 vs Q2, 1-5 vs Q3, and 5-1 vs Q4.
It looks evident that Q3 record is what kills them. But how do they compare to other bad Power 6 schools?

198 0-3 / 1-5 / 3-1 / 5-2 Boston College (missing a Q1 win, has 1 more Q4 loss, but better vs Q3)
200 1-6 / 0-4 / 1-4 / 5-0 Florida St (1 less Q2 win, but not have Q4 loss)
213 0-8 / 0-2 / 2-0 / 5-2 Oregon St (Q3 is better than USC, but 1 extra Q4 loss, AND 0 Q1+Q2 wins compared to 2 for USC)
216 1-5 / 0-3 / 1-3 / 5-0 Minnesota (no Q4 losses, but 1 Q1+Q2 wins vs 2 for USC)
240 0-5 / 0-5 / 1-3 / 5-2 Georgetown (0 Q1+Q2 wins, almost as bad vs Q3, and 1 more Q4 loss than USC)
277 1-2 / 1-3 / 1-5 / 5-1 USC
290 0-3 / 1-7 / 1-3 / 1-4 Cal (so USC is only 13 spots ahead of a team that has only won 20% of their Q4 games?)
335 0-4 / 0-7 / 0-2 / 2-4 Louisville

Looking at the above data alone, I would rank those teams BC, Minnesota, USC, FSU, Oregon St, Georgetown, Cal, UL
Minnesota is equal to or better than FSU in all 4 Quads, but sits 16 spots behind FSU.
The closest two IMO are FSU and Oregon St, so it appears FSU should drop.
So I would estimate USC should be in the 205-220 range (relative to those other P6 schools), not almost 280.

I am a Statistician. So analyzing and interpreting data is what I've done for 30 years, and these data do not make sense!

 

[HerrosHeroes]

Duke and North Carolina Statisticians

 

[JonathanW]

Other interesting things about the current Quad ranks:

St. Marys is 7, yet is 1-2 / 5-0 / 5-2 / 6-0. So losing record vs Q1 AND 2 Q3 losses. Their are 5 teams behind them that have a better Q1 record AND 0 Q2/Q3/Q4 losses. Missouri is 48, yet has a similar Q1 record (2-5) with no Q2/Q3/Q4 losses.

TCU is 13, with a Q4 loss like UK has. Not saying we should be where they are (yet). But the difference between us and them is 5-3 compared to 1-5 vs Q1. So with say 6 more Q1 wins we could get that high.

Boise St is at 22, with a Q4 loss. Except they have 0 Q1 wins. 0-2 / 8-2 / 2-0 / 5-1. Compared to UK 1-5 / 3-0 / 4-0 / 6-1. You look at that, and we should be ahead of Boise St. We are better vs Q1 and Q2, same vs Q3 and Q4.

Ohio St is at 26, with a Q4 loss. They are 2-7 / 3-1 / 0-0 / 6-1. So one more Q1 win for them, but 1 more Q2 loss. We should be about equal with them in the ranking, but are 7 spots behind.

 

[Cowtown Cat]

Other interesting things about the current Quad ranks:

St. Marys is 7, yet is 1-2 / 5-0 / 5-2 / 6-0. So losing record vs Q1 AND 2 Q3 losses. The next 3 teams behind them (& probably more) have a better Q1 record AND 0 Q2/Q3/Q4 losses.

TCU is 13, with a Q4 loss like UK has. Not saying we should be where they are (yet). But the difference between us and them is 5-3 compared to 1-5 vs Q1. So with say 6 more Q1 wins we could get that high.

Boise St is at 22, with a Q4 loss. Except they have 0 Q1 wins. 0-2 / 8-2 / 2-0 / 5-1. Compared to UK 1-5 / 3-0 / 4-0 / 6-1. You look at that, and we should be ahead of Boise St. We are better vs Q1 and Q2, same vs Q3 and Q4.

Ohio St is at 26, with a Q4 loss. They are 2-7 / 3-1 / 0-0 / 6-1. So one more Q1 win for them, but 1 more Q2 loss. We should be about equal with them in the ranking, but are 13 spots behind.

Damn, that’s insane. Thanks for the info. Very interesting.

 

[bbncal02]

Other interesting things about the current Quad ranks:

St. Marys is 7, yet is 1-2 / 5-0 / 5-2 / 6-0. So losing record vs Q1 AND 2 Q3 losses. Their are 5 teams behind them that have a better Q1 record AND 0 Q2/Q3/Q4 losses. Missouri is 48, yet has a similar Q1 record (2-5) with no Q2/Q3/Q4 losses.

TCU is 13, with a Q4 loss like UK has. Not saying we should be where they are (yet). But the difference between us and them is 5-3 compared to 1-5 vs Q1. So with say 6 more Q1 wins we could get that high.

Boise St is at 22, with a Q4 loss. Except they have 0 Q1 wins. 0-2 / 8-2 / 2-0 / 5-1. Compared to UK 1-5 / 3-0 / 4-0 / 6-1. You look at that, and we should be ahead of Boise St. We are better vs Q1 and Q2, same vs Q3 and Q4.

Ohio St is at 26, with a Q4 loss. They are 2-7 / 3-1 / 0-0 / 6-1. So one more Q1 win for them, but 1 more Q2 loss. We should be about equal with them in the ranking, but are 7 spots behind.

Welcome to the NCAA….

 

[BigBlueFanGA]

As we all know, that USC loss is like a lead weight hanging around UK's neck. So I was looking at why is USC a Q4 loss?

So USC sits at 277, going 1-2 vs Q1, 1-3 vs Q2, 1-5 vs Q3, and 5-1 vs Q4.
It looks evident that Q3 record is what kills them. But how do they compare to other bad Power 6 schools?

198 0-3 / 1-5 / 3-1 / 5-2 Boston College (missing a Q1 win, has 1 more Q4 loss, but better vs Q3)
200 1-6 / 0-4 / 1-4 / 5-0 Florida St (1 less Q2 win, but not have Q4 loss)
213 0-8 / 0-2 / 2-0 / 5-2 Oregon St (Q3 is better than USC, but 1 extra Q4 loss, AND 0 Q1+Q2 wins compared to 2 for USC)
216 1-5 / 0-3 / 1-3 / 5-0 Minnesota (no Q4 losses, but 1 Q1+Q2 wins vs 2 for USC)
240 0-5 / 0-5 / 1-3 / 5-2 Georgetown (0 Q1+Q2 wins, almost as bad vs Q3, and 1 more Q4 loss than USC)
277 1-2 / 1-3 / 1-5 / 5-1 USC
290 0-3 / 1-7 / 1-3 / 1-4 Cal (so USC is only 13 spots ahead of a team that has only won 20% of their Q4 games?)
335 0-4 / 0-7 / 0-2 / 2-4 Louisville

Looking at the above data alone, I would rank those teams BC, Minnesota, USC, FSU, Oregon St, Georgetown, Cal, UL
Minnesota is equal to or better than FSU in all 4 Quads, but sits 16 spots behind FSU.
The closest two IMO are FSU and Oregon St, so it appears FSU should drop.
So I would estimate USC should be in the 205-220 range (relative to those other P6 schools), not almost 280.

I am a Statistician. So analyzing and interpreting data is what I've done for 30 years, and these data do not make sense!

I agree

 

[The_Answer1313]

As we all know, that USC loss is like a lead weight hanging around UK's neck. So I was looking at why is USC a Q4 loss?

So USC sits at 277, going 1-2 vs Q1, 1-3 vs Q2, 1-5 vs Q3, and 5-1 vs Q4.
It looks evident that Q3 record is what kills them. But how do they compare to other bad Power 6 schools?

198 0-3 / 1-5 / 3-1 / 5-2 Boston College (missing a Q1 win, has 1 more Q4 loss, but better vs Q3)
200 1-6 / 0-4 / 1-4 / 5-0 Florida St (1 less Q2 win, but not have Q4 loss)
213 0-8 / 0-2 / 2-0 / 5-2 Oregon St (Q3 is better than USC, but 1 extra Q4 loss, AND 0 Q1+Q2 wins compared to 2 for USC)
216 1-5 / 0-3 / 1-3 / 5-0 Minnesota (no Q4 losses, but 1 Q1+Q2 wins vs 2 for USC)
240 0-5 / 0-5 / 1-3 / 5-2 Georgetown (0 Q1+Q2 wins, almost as bad vs Q3, and 1 more Q4 loss than USC)
277 1-2 / 1-3 / 1-5 / 5-1 USC
290 0-3 / 1-7 / 1-3 / 1-4 Cal (so USC is only 13 spots ahead of a team that has only won 20% of their Q4 games?)
335 0-4 / 0-7 / 0-2 / 2-4 Louisville

Looking at the above data alone, I would rank those teams BC, Minnesota, USC, FSU, Oregon St, Georgetown, Cal, UL
Minnesota is equal to or better than FSU in all 4 Quads, but sits 16 spots behind FSU.
The closest two IMO are FSU and Oregon St, so it appears FSU should drop.
So I would estimate USC should be in the 205-220 range (relative to those other P6 schools), not almost 280.

I am a Statistician. So analyzing and interpreting data is what I've done for 30 years, and these data do not make sense!

Well a couple of things. This is a problem with the committee as well as analyzing it like this.
You cannot just group things into four quads and think everything is equal. The ranges are far too wide and it leaves out important information when things are done this way.

Team A could be 0-4 in Q1 games but lost to say the top 4 teams and all on the road.
Team B could also be 0-4 in Q1 games but lost all of them at home to teams barely in Q1.

Looking just at records, people would conclude they are the same. But they are far from that. These Quads are just an easy way for humans to look at these things but anytime you put things into groupings like this, you miss information.

The other thing is the quad system is based on your NET ranking. This of course factors in several different things such as efficiency numbers. It factors in offensive efficiency, defensive efficiency, strength of schedule, whether the game was home/away/neutral etc.

USC is ranking where they are ranked because of all of that.

Furthermore, Kenpom has them at 244th. Sagarin has them 205th tho.
Massey composite rankings which averages a bunch of systems = 227th.

Regardless anything south of 200 isn't good.

But as far as our rating is concerned the reason we aren't higher has less to do with one game (USC) and more to do with our inability to beat anyone of note except UT yet.

 

[Padsfs07]

The NET isn't just about Quadrant wins and losses. They are two seperate entities.

The NET takes into account game results, strength of schedule, game location, scoring margin, net offensive and defensive efficiency, and the quality of wins and losses.

Once they make a list, then the quadrants come into play.

 

[The_Answer1313]

The NET isn't just about Quadrant wins and losses. They are two seperate entities.

The NET takes into account game results, strength of schedule, game location, scoring margin, net offensive and defensive efficiency, and the quality of wins and losses.

Once they make a list, then the quadrants come into play.

Exactly.

The never released the actual NET formula but those were the factors that go into it. Tho I do believe they removed scoring margin as of this season

It never made much sense to me because if you are factoring in offensive and defensive efficiency you are factoring in scoring margin anyways lol. You have a good offensive efficiency number and a good defensive efficiency number. Subtract the two and that's your efficiency margin (the number you see on Kenpom home page). That's basically scoring margin on a per possession basis.

 

[St.PatterSoN-54-]

 

[BigBlueFanGA]

Well a couple of things. This is a problem with the committee as well as analyzing it like this.
You cannot just group things into four quads and think everything is equal. The ranges are far too wide and it leaves out important information when things are done this way.

Team A could be 0-4 in Q1 games but lost to say the top 4 teams and all on the road.
Team B could also be 0-4 in Q1 games but lost all of them at home to teams barely in Q1.

Looking just at records, people would conclude they are the same. But they are far from that. These Quads are just an easy way for humans to look at these things but anytime you put things into groupings like this, you miss information.

The other thing is the quad system is based on your NET ranking. This of course factors in several different things such as efficiency numbers. It factors in offensive efficiency, defensive efficiency, strength of schedule, whether the game was home/away/neutral etc.

USC is ranking where they are ranked because of all of that.

Furthermore, Kenpom has them at 244th. Sagarin has them 205th tho.
Massey composite rankings which averages a bunch of systems = 227th.

Regardless anything south of 200 isn't good.

But as far as our rating is concerned the reason we aren't higher has less to do with one game (USC) and more to do with our inability to beat anyone of note except UT yet.

But that didn't happen. Go to https://stats.ncaa.org/selection_rankings/nitty_gritties/29668 and try to explain Boise St being that far above UK. I'll wait.

For that matter, how is Georgia St that far ahead of USCjr? It makes no sense.

 

[The_Answer1313]

But that didn't happen. Go to https://stats.ncaa.org/selection_rankings/nitty_gritties/29668 and try to explain Boise St being that far above UK. I'll wait.

They are probably rated higher for the same reason they are higher in kenpom. They have better efficiency numbers than we do.

In this they are 22. We are 33’
Kenipom has them 27 and us 29.

I don’t have the NET formula so I don’t know the difference but given both are efficiency based that’s probably what it boils down to.

 

[The_Answer1313]

FWIW when it comes to mid majors I do feel like computer systems tend to overate them.

But humans vastly underrated these teams.

The lack of sample size against good competition I think makes it hard for anyone to assess these teams.

Is st marys good? Yes probably. Are they the 7th best team in the country? Not so sure about that

 

[The_Answer1313]

Lastly there’s a limit to all of this.

I’d argue there’s not a whole difference between being 22nd or 33rd in this thing.

The teams could easily be reversed.

This isn’t exact. No one can definitely say oh the 25th best team in this ranking is better than say the 35th.

I would look at it in terms of ranges and that both UK and Boise might be anywhere say 20 to 30th

 

[BigBlueFanGA]

They are probably rated higher for the same reason they are higher in kenpom. They have better efficiency numbers than we do.

In this they are 22. We are 33’
Kenipom has them 27 and us 29.

I don’t have the NET formula so I don’t know the difference but given both are efficiency based that’s probably what it boils down to.

Interesting. The team sheets don't mention efficiency. That should have little to do with it when you play a crap schedule. It isn't possible to "adjust" accurately but wins, losses, home, away and relative strength of schedule are known. That should be the litmus test.

 

[crestcat]

If Cow had taught those Dudes to be unselfish earlier none of them numbers would matter. LOL

 

[The_Answer1313]

another thing about this is you don’t even need the quad system. The quad system based on NET isn’t telling you anything additional than what NET is already saying. In fact it’s removing info and losing information

 

[The_Answer1313]

Interesting. The team sheets don't mention efficiency. That should have little to do with it when you play a crap schedule. It isn't possible to "adjust" accurately but wins, losses, home, away and relative strength of schedule are known. That should be the litmus test.

It’s possible. It might not be entirely 100% but then again nothing is.

We moved from RPI to NET. It’s definitely a step up in the right direction.

IMO next thing is just can the committee completely and use an aggregate of good rating systems to seed people

 

[The_Answer1313]

Another thing about seeding we need to understand is the committee isn’t picking the top at large teams. They are picking the teams with the best resume. The two isn’t exactly the same

This is why sometimes you have games where the worse seed is actually the favorite or the line is closer than what should be expected. And then naturally these become the upset picks for nearly everyone

 

[SaltyCat22]

The Man. When The Man is so cruel to calculate my stocks to the one hundred thousandths of a penny you know we are going up against the might of the Military Industrial Complex.

 

[ncaaky]

We moved from RPI to NET. It’s definitely a step up in the right direction.

IMO next thing is just can the committee completely and use an aggregate of good rating systems to seed people

The best way to seed the teams would be to let Vegas rate the teams. That would give the fairest and most balanced brackets. It would be pretty darn simple as well.

 

[The_Answer1313]

The best way to seed the teams would be to let Vegas rate the teams. That would give the fairest and most balanced brackets. It would be pretty darn simple as well.

I agree.

 

[CincinnatiWildcat]

The best way to seed the teams would be to let Vegas rate the teams. That would give the fairest and most balanced brackets. It would be pretty darn simple as well.

You are assuming that if we let Vegas rank the teams they would have the incentive to make a fair bracket and not one with match-ups that optimize their own profits, or that those two things are the same. You would just end up with the bubble teams that get in being whoever Vegas inside information tells them will get the most bets placed on their game.

 

[phyrekat]

The Man. When The Man is so cruel to calculate my stocks to the one hundred thousandths of a penny you know we are going up against the might of the Military Industrial Complex.

Can I subscribe to your newsletter?

 

[TucsonCat]

The best way to seed the teams would be to let Vegas rate the teams. That would give the fairest and most balanced brackets. It would be pretty darn simple as well.

Vegas is in the business of predicting. Tournament should be about actual accomplishment as well. I hate to use us as an example. But Vegas would probably put us as 5 seed right now with our roster talent and play of late. But we haven't earned that. Our on court resume isn't there, yet.

For all the gnashing of teeth and outrage on the internet, the committee does a fine job.

 

[fatguy87]

No matter the metric or measure, USC sucks. The loss to them is a stain on our resume.

 

[JonathanW]

Vegas is in the business of predicting. Tournament should be about actual accomplishment as well. I hate to use us as an example. But Vegas would probably put us as 5 seed right now with our roster talent and play of late. But we haven't earned that. Our on court resume isn't there, yet.

For all the gnashing of teeth and outrage on the internet, the committee does a fine job.

Mostly agree. The tournament is about performance, the past 4 months. Although IMO it should more heavily weight the last month. Or even the last week, for example if 2 teams had exact same resumes, except one struggled the last week and one didn't.
The tournament is not meant, and shouldn't be, to be about who is expected to perform the best in it.
The 13 team is a good example, a team that had the overall resume and was surging when Noel got hurt, BUT when you looked at it after the Noel-injury, that group wasn't tournament worthy.

 

[Blueaz]

As we all know, that USC loss is like a lead weight hanging around UK's neck. So I was looking at why is USC a Q4 loss?

So USC sits at 277, going 1-2 vs Q1, 1-3 vs Q2, 1-5 vs Q3, and 5-1 vs Q4.
It looks evident that Q3 record is what kills them. But how do they compare to other bad Power 6 schools?

198 0-3 / 1-5 / 3-1 / 5-2 Boston College (missing a Q1 win, has 1 more Q4 loss, but better vs Q3)
200 1-6 / 0-4 / 1-4 / 5-0 Florida St (1 less Q2 win, but not have Q4 loss)
213 0-8 / 0-2 / 2-0 / 5-2 Oregon St (Q3 is better than USC, but 1 extra Q4 loss, AND 0 Q1+Q2 wins compared to 2 for USC)
216 1-5 / 0-3 / 1-3 / 5-0 Minnesota (no Q4 losses, but 1 Q1+Q2 wins vs 2 for USC)
240 0-5 / 0-5 / 1-3 / 5-2 Georgetown (0 Q1+Q2 wins, almost as bad vs Q3, and 1 more Q4 loss than USC)
277 1-2 / 1-3 / 1-5 / 5-1 USC
290 0-3 / 1-7 / 1-3 / 1-4 Cal (so USC is only 13 spots ahead of a team that has only won 20% of their Q4 games?)
335 0-4 / 0-7 / 0-2 / 2-4 Louisville

Looking at the above data alone, I would rank those teams BC, Minnesota, USC, FSU, Oregon St, Georgetown, Cal, UL
Minnesota is equal to or better than FSU in all 4 Quads, but sits 16 spots behind FSU.
The closest two IMO are FSU and Oregon St, so it appears FSU should drop.
So I would estimate USC should be in the 205-220 range (relative to those other P6 schools), not almost 280.

I am a Statistician. So analyzing and interpreting data is what I've done for 30 years, and these data do not make sense!

There is page explaining it… w and Ls are only part. Efficiency is a big part.
It is explained very vaguely… WITHOUT the actual way it’s calculated… that way, the prejudices can be hidden

 

[Blueaz]

No matter the metric or measure, USC sucks. The loss to them is a stain on our resume.

But not as huge, as some like to make it seem

 

[HerrosHeroes]

But not as huge, as some like to make it seem

And not the way they shot that night.
11 of 20 from 3 for 55%
As Coach Cal says: “They grew beer muscles.”

 

[The_Answer1313]

Another thing to note is the NET and Quad stuff is just one thing.
At the end, it's still a bunch of humans in a room making the actual seeding decisions. It's not like they are going to run 1 through 32 in the NET and seed exactly.