Top 3 player lineups vs Quad 1 teams | The Boneyard

Top 3 player lineups vs Quad 1 teams

HuskyNan said:
Expand ...

What? And who is this person?

In the 6 games UCONN played vs Lousville, Ohio State, Michigan, SoCal, ND and tennesse these are her stats combined:

True Shooting % 68%.

Usage Percent % 24%

Assist Percent % 20.9%

Turnover Pct % 8.3%

Offensive Rating 125

Defensive Rating 92

Win Share Per 40 33
 
hoophuskee said:
What? And who is this person?

In the 6 games UCONN played vs Lousville, Ohio State, Michigan, SoCal, ND and tennesse these are her stats combined:

True Shooting % 68%.

Usage Percent % 24%

Assist Percent % 20.9%

Turnover Pct % 8.3%

Offensive Rating 125

Defensive Rating 92

Win Share Per 40 33
Expand ...
It’s lineup combinations, not one person’s performance
 
There's a bit of legerdemain involving the conversion of stats from 5 players to 3. But it is still very revealing. I'd begin by considering minutes played by each of these lineups, and then consider the combination of O and D rating. Louisville only has 79 minutes and an excellent D rating. The #2 lineup is KK Azzi and Sarah with a similar O rating and slightly worse D rating but over 151 minutes. Congrats to Louisville on the defense of this group. Too bad they can't sustain it for long.

Also congrats to Texas on an excellent O rating, though the corresponding D rating is a lot worse. The really striking thing is how many UConn lineups make it onto the list, and the cumulative minutes of excellence: 757 minutes. No other team is even close. to this.
 
.-.
It is amusing to see the lengths that people will go to find some stat in which they can point to someone other than UConn being at the top. Obviously there are some individual player stats where the top player is not from UConn but at least Skim recognizes this as a team sport. They can't look overall team record for obvious reasons. They can't even consider only tournament teams are only quad one teams because UConn is undefeated. UCLA teams will of course tout the number of quad one wins, and conveniently ignore the fact that UConn quad one wins are clustered near the top while UCLA's are not.

So someone dreamed up looking at three player combos. Did they try looking at that stat against all teams or all NCAA tournament likely teams? I don't know, but they chose to look at games against quad one teams. Not a bad choice but whenever one delivers a statistic, it's always sensible to question whether that was their original selection or did they try a few and pick the one that made the point they wish to make?

In this particular case, they found that seven of the top 10 or UConn so it's going to be a challenge to conclude anything other than that UConn is at the top of the heap.

Sarah Strong is in three of those seven lineups which is pretty darn good. KK Arnold is in 4, as is Azzi Fudd, and Sarah Williams. Each of home along with Ashland shade are in more of the top 10 lineups then anyone from Louisville UCLA Texas or South Carolina.

Sound like the overall message is UConn dominance of this metric, yet Skim manages to tease out a way to feature a different team. All posters in X are welcome to cherry pick stats that suit what they like to say. I hope they don't mind if others point out how strained it looks. (I'm not arguing with the overall metric, it's interesting and stunningly supportive of the notion that UConn is the best team, I'm just finding it amusing that you can take a table that so UConn dominant and your three take away points ignore that dominance.)
 
Bone Dog said:
Louisville only has 79 minutes and an excellent D rating.
Expand ...
That is an excellent observation.

I'm curious how they decided the cut off at 53.7 minutes, conveniently making sure that the 56 minutes played by Texas qualified. If my math is right, Texas is played nine quad one opponents for a total of 360 minutes, which means this lineup has been in the game a little under 16% of the time. 56 minutes is roughly 2 full games of play which seems pretty limited data to draw conclusions.
 
UConn has played six quad one opponents for a total of 240 minutes

How is it that the combined minutes of the seven UConn entries in this table add up to 757?

When they say Louisville has 79 minutes with this combination, does that mean 79 minutes of game time this combination was on the floor or are they adding up the minutes of each of the three players?
 
hoophuskee said:
Thne why was there a reference of Sarah?
Expand ...
It’s the performance of that particular 3 person lineup rather than just one person’s stats. Is the team more effective with Azzi-Sarah-Blanca on the court or Kk-Azzi-Sarah? It’s just another metric

IMG_4803.jpeg
IMG_4804.jpeg
 
Those are seven of the ten possible combinations of three of the five starters. I bet if you ran the list to 20 the other three combinations would appear. This just means that Connecticut's starting five is head and shoulders above anyone else's.
 
I had not seen this statistic before, but it is an interesting approach to effectiveness and it also largely reflects standard sub patterns for coaches - starters, two subs replacing two starters, and then a rotation of those seven with maybe an 8th and even a 9th rotating to give rest, adjust for fouls, etc. One of the reasons Uconn has 7 of the top 10 is because this is exactly how his standard sub rotation works. And their general dominance in their Quad 1 wins has allowed him to freely cycle through his line-ups, where in tight games his starting 5 might dominate the court time.

And I doubt 53.7 minutes is a random selection - it is probably derived for a set percentage of the average minutes of D1 Quad 1 match-ups played by teams.

I would like to know the date these stats were compiled - it would appear to be quite recent, but even a 7 day difference would have a significant effect for teams whose OOC was non-competitive and are only now getting into the meat of their schedule.

Of particular note to me - UCLA has played a lot of quad1 opponents - that they have only 1 top ten combination listed, and that the minutes played do not dwarf Uconn's is really surprising. TX also with their top 3 rotation at only 56 minutes is also surprising.

And ... that we now think it unsurprising that advance analytics and an obscure (to me at least) statistical analysis is casually reported for WCBB is a great indicator that the dark ages for WBB are slowly receding!
 
.-.
Bone Dog said:
There's a bit of legerdemain involving the conversion of stats from 5 players to 3. But it is still very revealing. I'd begin by considering minutes played by each of these lineups, and then consider the combination of O and D rating. Louisville only has 79 minutes and an excellent D rating. The #2 lineup is KK Azzi and Sarah with a similar O rating and slightly worse D rating but over 151 minutes. Congrats to Louisville on the defense of this group. Too bad they can't sustain it for long.

Also congrats to Texas on an excellent O rating, though the corresponding D rating is a lot worse. The really striking thing is how many UConn lineups make it onto the list, and the cumulative minutes of excellence: 757 minutes. No other team is even close. to this.
Expand ...
The 3-player meta rotations are indeed the collection of 5-player rotation stints having those 3 players up to the 2/4 DePaul game.
  • I verified the minutes played (in Tier I = Quad 1) for the 3-player groups for KK-Azzi-Sarah and KK-Ash-Azzi.
  • The 3-player stats are just aggregating the 5-player rotation stint stats.
In Geno and CD’s rotation schemes, the starters play the most with the “Top 8” and “Upper Bench”.
  • Thus it makes most sense that UConn starters feature the best 3-player combos.
  • That the best 3-player combos in all D1 WCBB are mostly UConn’s arguably means UConn has the best starting five and possibly the best Top (N‎ = ) 8 used for Quad 1 games;
  • I would argue that since most UConn Quad 1 games were early season and UConn has since improved are some of the reasons why Kayleigh’s or Blanca’s 3-player rotations are not in the dataset;
  • It also means Serah has been proving her worth as a starter.
Interestingly, UConn and Louisville both have 140+ rotation stints to-date.
 
NycUcWbbFan said:
The 3-player meta rotations are indeed the collection of 5-player rotation stints having those 3 players up to the 2/4 DePaul game.
  • I verified the minutes played (in Tier I = Quad 1) for the 3-player groups for KK-Azzi-Sarah and KK-Ash-Azzi.
  • The 3-player stats are just aggregating the 5-player rotation stint stats.
In Geno and CD’s rotation schemes, the starters play the most with the “Top 8” and “Upper Bench”.
  • Thus it makes most sense that UConn starters feature the best 3-player combos.
  • That the best 3-player combos in all D1 WCBB are mostly UConn’s arguably means UConn has the best starting five and possibly the best Top (N‎ = ) 8 used for Quad 1 games;
  • I would argue that since most UConn Quad 1 games were early season and UConn has since improved are some of the reasons why Kayleigh’s or Blanca’s 3-player rotations are not in the dataset;
  • It also means Serah has been proving her worth as a starter.
Interestingly, UConn and Louisville both have 140+ rotation stints to-date.
Expand ...
Can I ask why is this of value? Also, I still don't understand why Sarah was singled out but anyhow games are 5-on-5? It is fun to look at for sure. Nan brings up interesting discusion - heck i thought a great triplet group in NBA was Bird, McHale, Parish. For our UCONN teams such as Lobo, Wolters Rizzzotti, to so many others. Its; a fun disussion who is best just like when we discuss who is greatest team etc. But you are heavy into stats; do you see any value in it, other than a fun discussion?

I'm big on having three offenisve players. And the Bulls while Rodman wasn't a scorer he was among the greatest offensive rebounders of all-time. In other words, their Big 3 was different than Magic, Worthy, and Jabbar as ab example. So there can be different ways to lok at a "Big 3."

I just think this current UCONN team has two big time offensive players and then a vaiiety of others that can put up points or make big time offensive plays. It doesn't follow the path of a clear Big 3 and imo that's okay because several can be part of that 3 or 4 or 5 on a given night. Two players that score 9 and 9 can be equivlent to 14 and 4 if other factors the same.
 
hoophuskee said:
Can I ask why is this of value? Also, I still don't understand why Sarah was singled out but anyhow games are 5-on-5? It is fun to look at for sure. Nan brings up interesting discusion - heck i thought a great triplet group in NBA was Bird, McHale, Parish. For our UCONN teams such as Lobo, Wolters Rizzzotti, to so many others. Its; a fun disussion who is best just like when we discuss who is greatest team etc. But you are heavy into stats; do you see any value in it, other than a fun discussion?

I'm big on having three offenisve players. And the Bulls while Rodman wasn't a scorer he was among the greatest offensive rebounders of all-time. In other words, their Big 3 was different than Magic, Worthy, and Jabbar as ab example. So there can be different ways to lok at a "Big 3."

I just think this current UCONN team has two big time offensive players and then a vaiiety of others that can put up points or make big time offensive plays. It doesn't follow the path of a clear Big 3 and imo that's okay because several can be part of that 3 or 4 or 5 on a given night. Two players that score 9 and 9 can be equivlent to 14 and 4 if other factors the same.
Expand ...
3-player stats are more useful &A for cross-D1/ cross-pro ball comparisons but it is noteworthy that UConn’s “beautiful basketball” &B dominates these stats.

Geno and CD have used 141 rotation stints satisfying multiple objectives.
  • With such a high number (due to the fast paced mix-match due to the high quality roster), it is a challenge to communicate what’s in the 141 rotations, without encountering general stats-coma;
  • It’s easier to explain 3-player groupings of the 141 rotations (such as below with KK-Azzi-Sarah and Azzi-Blanca-Sarah);
  • A good 3-player core is a good start to building rotations that are meant to win games (“Winning Time Rotations” + “Pocket Bridge Rotations”) and are also easy mnemonic devices;
  • &A: The 3-player stats framework in the way CBAnalytics uses it gives across-D1 deeper-dive &1 insight (as does RAPM and BPM etc.) that is not available in box scores or individual player stats.
&B Is it useful to Geno and CD?
  • Geno and CD concentrate on producing “plug-and-play” “basketball players” playing “perfect possessions” all capable of producing uniform separation from opponents. Best-in-class (team) metrics affirm their best-in-class approach.
&1 Why I think Sarah and Azzi are first team AA material.

Example 1: KK + Azzi + Sarah
IMG_8343.jpeg


Example 2: Azzi + Blanca + Sarah
IMG_8344.jpeg
 
NycUcWbbFan said:
3-player stats are more useful &A for cross-D1/ cross-pro ball comparisons but it is noteworthy that UConn’s “beautiful basketball” &B dominates these stats.

Geno and CD have used 141 rotation stints satisfying multiple objectives.
  • With such a high number (due to the fast paced mix-match due to the high quality roster), it is a challenge to communicate what’s in the 141 rotations, without encountering general stats-coma;
  • It’s easier to explain 3-player groupings of the 141 rotations (such as below with KK-Azzi-Sarah and Azzi-Blanca-Sarah);
  • A good 3-player core is a good start to building rotations that are meant to win games (“Winning Time Rotations” + “Pocket Bridge Rotations”) and are also easy mnemonic devices;
  • &A: The 3-player stats framework in the way CBAnalytics uses it gives across-D1 deeper-dive &1 insight (as does RAPM and BPM etc.) that is not available in box scores or individual player stats.
&B Is it useful to Geno and CD?
  • Geno and CD concentrate on producing “plug-and-play” “basketball players” playing “perfect possessions” all capable of producing uniform separation from opponents. Best-in-class (team) metrics affirm their best-in-class approach.
&1 Why I think Sarah and Azzi are first team AA material.

Example 1: KK + Azzi + Sarah
View attachment 116904

Example 2: Azzi + Blanca + Sarah
View attachment 116905
Expand ...
Can I PM you with a few questions today or tomorrow? I think I'd bore everyone on here. No problem if you are too busy. Super Bowl is over.
 
hoophuskee said:
Thne why was there a reference of Sarah?
Expand ...
Skim and most sites thrive and earn on controversy and discussion.

Sarah favorite for player of the year, but this stat shows others on UConn team may be more valuable.

This nugget sure to be controversial and generate responses.

That is only reason she is mentioned.

The truly interesting takeaway is UConn has 7 of top10 3 player combos in ncaa.
 
.-.
The stats would be more meaningful to me if they were derived from playing common opponents.
Otherwise the "glass of stats" only seems to be 1/2 full & somewhat lacking a degree of relevant predictibility.
While retrospectively interesting, how can they be interpreted as being applicable to being predictable indicators of future game performance?
IMO the short answer is that they can't be reliably applied as being predictible of future game performance.
 
Last edited:
packwrap said:
Skim and most sites thrive and earn on controversy and discussion.
Sarah favorite for player of the year, but this stat shows others on UConn team may be more valuable.
This nugget sure to be controversial and generate responses.
That is only reason she is mentioned.

The truly interesting takeaway is UConn has 7 of top10 3 player combos in ncaa.
Expand ...
The lineups are mathematically derived, not opinions. Skim is reporting numbers calculated by CBB Analytics . And she doesn’t thrive on controversy. Her entire reputation is built on reporting facts, especially on recruiting

I agree with your final sentence - that was the point of my posting the info

I don’t know why people are upset about some statistics that show how great the Huskies are. I’ll limit the presentation of more stats and facts going forward because they seem to make people angry
 
HuskyNan said:
The lineups are mathematically derived, not opinions. Skim is reporting numbers calculated by CBB Analytics . And she doesn’t thrive on controversy. Her entire reputation is built on reporting facts, especially on recruiting

I agree with your final sentence - that was the point of my posting the info

I don’t know why people are upset about some statistics that show how great the Huskies are. I’ll limit the presentation of more stats and facts going forward because they seem to make people angry
Expand ...
For the record, at least for me, I wasn't upset. I don't know why she bothered to report these numbers though. That's why I posted to nyc - what purpose do these stats show? If Sarah has such spectacular numbers cumulative, imo that has a ton more meaning. And the person didn't mention the top player from Texas not being on the list. I just found it wierd. She is right there for NPOY. And I have no idea who his person is. But okay - I'm done with this. Matter of fact, I treid further to learn from them. That's why I sent a dm to nyc.
 
The “53.7” minutes against Quad 1 opponents does filter in only &1 11 3-player lineups from UConn; I suspect that similar numbers of lineups were also filtered-in from other teams.
  • CBBAnalytics uses OffRtg and DefRtg (and derivatively NetRtg) derived using a regression based on Dean Oliver’s Four Factors (with data from play-by-play info at the rotation stint level) for its “D1 Leaderboard”.
  • For a comparable statistic to NetRtg (with much less data requirements) see Per 40 +/-.
It’s very noteworthy that 7 (of the 11) UConn 3-Player Lineups are in the Top 10 “D1 Leaderboard”. This is perhaps due to UConn’s top rankings in the Four Factors vs Quad 1 opponents.

&1 Blanca being out for two Quad 1 games (Louisville and Tennessee) filtered out her combinations. This also affected Sarah in this statistic.

IMG_9202.jpeg
 
HuskyNan said:
The lineups are mathematically derived, not opinions. Skim is reporting numbers calculated by CBB Analytics . And she doesn’t thrive on controversy. Her entire reputation is built on reporting facts, especially on recruiting

I agree with your final sentence - that was the point of my posting the info

I don’t know why people are upset about some statistics that show how great the Huskies are. I’ll limit the presentation of more stats and facts going forward because they seem to make people angry
Expand ...
Skim didn't make me angry, and I wasn't questioning accuracy of Skim's stats.

The other poster asked why Skim singled out Sarah as the player mentioned. Only her comment about Sarah was an opinion, not the stats. Her opinion "Sarah Strong only in 1 of UConn's best 5 line ups".

This will of course generate discussion, as it implies Sarah not really important, could be on the bench, and maybe not really best player of the year candidate. etc.

It's just always important to know that Skim and other sites are trying to generate clicks, discussion and thereby revenue. Knowing their goal, it is then easy to ignore.
 
.-.
HuskyNan said:
The lineups are mathematically derived, not opinions. Skim is reporting numbers calculated by CBB Analytics . And she doesn’t thrive on controversy. Her entire reputation is built on reporting facts, especially on recruiting

I agree with your final sentence - that was the point of my posting the info

I don’t know why people are upset about some statistics that show how great the Huskies are. I’ll limit the presentation of more stats and facts going forward because they seem to make people angry
Expand ...
Please don’t limit these types of posts as they are informative and thought provoking and better than most of the solely opinion based regurgitation. While unknown, we can’t be afraid of new concepts and ideas. Don’t let the stubborn get you down…
 
You must log in or register to reply here.

Forum statistics

Threads
167,433
Messages
4,522,705
Members
10,399
Latest member
southcampus


Top Bottom