Tracking Duke’s Defense

[cato]

From COYS’ Phase I post:


No matter what defense Duke employs, keep a close eye on opponent 2pt% and turnover rates as indicators for which direction this Duke team is trending. The Michigan State game will be an excellent early test of this young Duke team’s defense.

Excellent advice. Why don’t we do exactly that here.

[Kedsy]

Duke defensive stats vs. Elon:

2-point%: .367
3-point%: .424
%threes: 52.4%
eFG%: 50.8%

TO%: 20.2%
DR%: 82.5%
FTRate: 11.1%
oRating: 0.91

Everything good except letting them hit 42+% of their three-point shots.

[pfrduke]

Duke defensive stats vs. Elon:

2-point%: .367
3-point%: .424
%threes: 52.4%
eFG%: 50.8%

TO%: 20.2%
DR%: 82.5%
FTRate: 11.1%
oRating: 0.91

Everything good except letting them hit 42+% of their three-point shots.

Couple other things to add:

A/TO: 1:1.5
Assist rate: 40%
Block %: 4.8% (10% of 2s)

[sagegrouse]

Duke defensive stats vs. Elon:

2-point%: .367
3-point%: .424
%threes: 52.4%
eFG%: 50.8%

TO%: 20.2%
DR%: 82.5%
FTRate: 11.1%
oRating: 0.91

Everything good except letting them hit 42+% of their three-point shots.

Just a question. Is there a "good" and "not so good" on percent of shots taken from two-point range, or is the stat just a weighting mechanism to determine effective field goal percentage?

[Mtn.Devil.91.92.01.10.15]

Duke defensive stats vs. Elon:

2-point%: .367
3-point%: .424
%threes: 52.4%
eFG%: 50.8%

TO%: 20.2%
DR%: 82.5%
FTRate: 11.1%
oRating: 0.91

Everything good except letting them hit 42+% of their three-point shots.

It does seem that given our size lots of opponents will resort to distance shots. I assume K will be happy to have teams make that adjustment if it means they aren't making easy baskets at the rim.

[COYS]

As Troublemaker pointed out in the Elon post-game thread, Duke has allowed the opposing team to shoot a lot of threes relative to past Duke teams in all three of the exhibitions and now also the first real game. This is such a drastic departure from past Duke teams. If this trend continues, it’s possible my focus on opponent 2pt% and turnover rate could be a little misguided . . . Although if we’re allowing more threes then it is perhaps even more imperative that we limit opponents from taking and making two point shots.

[Kedsy]

Just a question. Is there a "good" and "not so good" on percent of shots taken from two-point range, or is the stat just a weighting mechanism to determine effective field goal percentage?

I don't know that it's a matter of good or not good, but as COYS points out, it has always been a main goal of our defense to limit three-point attempts, since (as we all know) a made three is worth 1.5 of a made two. Here's 30 years worth of data:



Year % 3s %2s 3 pt success 2 pt success 1.5*3 succ
2017 28.8% 71.2% 0.293 0.489 0.440
2016 24.1% 75.9% 0.307 0.503 0.461
2015 27.8% 72.2% 0.314 0.463 0.471
2014 24.1% 75.9% 0.307 0.503 0.461
2013 26.9% 73.1% 0.29 0.462 0.435
2012 24.1% 75.9% 0.317 0.470 0.476
2011 24.6% 75.4% 0.324 0.432 0.486
2010 25.4% 74.6% 0.282 0.441 0.423
2009 26.0% 74.0% 0.338 0.468 0.507
2008 25.0% 75.0% 0.329 0.470 0.494
2007 24.3% 75.7% 0.315 0.457 0.473
2006 21.3% 78.7% 0.304 0.464 0.456
2005 20.9% 79.1% 0.305 0.413 0.458
2004 25.2% 74.8% 0.324 0.435 0.486
2003 25.1% 74.9% 0.346 0.477 0.519
2002 26.3% 73.7% 0.303 0.464 0.455
2001 24.4% 75.6% 0.344 0.439 0.516
2000 24.6% 75.4% 0.358 0.437 0.537
1999 25.6% 74.4% 0.301 0.422 0.452
1998 25.0% 75.0% 0.305 0.447 0.458
1997 23.4% 76.6% 0.340 0.447 0.510
1996 27.0% 73.0% 0.335 0.475 0.503
1995 24.2% 75.8% 0.390 0.454 0.585
1994 23.1% 76.9% 0.297 0.450 0.446
1993 19.9% 80.1% 0.289 0.481 0.434
1992 20.4% 79.6% 0.377 0.490 0.566
1991 19.4% 80.6% 0.347 0.468 0.521
1990 15.4% 84.6% 0.337 0.472 0.506
1989 18.2% 81.8% 0.320 0.446 0.480
1988 16.3% 83.7% 0.351 0.471 0.527
1987 13.9% 86.1% 0.335 0.472 0.503


Obviously we've only seen one game and two exhibitions so far, but our three opponents' aggregate info looks like this:



Year % 3s %2s 3 pt success 2 pt success 1.5*3 succ
2018 49.2% 50.8% 0.336 0.317 0.504


Since our previous team "record" for highest percentage of opponents' threes was 28.8% (set last season), the fact that almost half our opponents shots so far have been threes is a big change.

The imbalance between opponents' two-point efficiency vs. three-point efficiency should also be something to watch this season. In four of the past five seasons, two-pointers have been more efficient against us than three-pointers (the exception being 2015 when they were close to even). But before 2013, in only five of the 26 seasons since the three-point shot was adopted were two-point shots more efficient against us than three-point shots (though 2010 was one of the five). Perhaps our defensive strategy this season is trying to deal with the recent trend?

Presumably, the two efficiencies should be close to even (because if they're not then you have an exploitable flaw). In any event, we've never had anything close to a difference in efficiencies of .187 (.504 - .317). In only six of the 31 seasons since they adopted the three was our absolute difference larger than .055 (1988, 1992, 1995, 1997, 2000, 2001), and in only two seasons was the absolute difference larger than .080 (2000: .100 and 1995: .131).

Put another way, if our opponents' efficiency from two and three stay the same, then if we can lower the number of three-attempts they make our defense will get that much more effective. Even if attempting to limit opposing threes would increase opposing two-point efficiency, it would still seem to be a good strategy until the relative efficiencies were at least close to even.

[Kedsy]

Duke defensive stats vs. Elon:

2-point%: .367
3-point%: .424
%threes: 52.4%
eFG%: 50.8%

TO%: 20.2%
DR%: 82.5%
FTRate: 11.1%
dRating: 0.91


Duke defensive stats vs. UVU:

2-point%: .452
3-point%: .240
%threes: 37.3%
eFG%: 41.8%

TO%: 24.5%
DR%: 62.8%
FTRate: 23.9%
dRating: 0.89

A/to: 1:1.46
Asst Rate: 52.0%
Block %: 10.4% (16.7% of twos)

Efficiency difference: -.092 (opposing two efficiency much better than opposing three efficiency)

So, overall good but defensive rebounding and allowing easy cuts for twos less good.

[Kedsy]

Duke defensive stats vs. UVU:

2-point%: .452
3-point%: .240
%threes: 37.3%
eFG%: 41.8%

TO%: 24.5%
DR%: 62.8%


These are UK's corresponding defensive stats against UVU:

2-point: .474
3-point: .278
%threes: 32.1%
eFG%: 45.5%

TO%: 29.2%
DR%: 55.6%

So, we rebounded against Utah Valley better than Kentucky did and held UVU to a lower shooting percentage, but UK turned them over a lot more. And UVU attempted a higher percentage of threes against Duke than against UK. Our overall defensive efficiency (0.89) was a little better than Kentucky's (0.92).

[Kedsy]

Duke defensive stats vs. Michigan State:

2-point%: .611
3-point%: .360
%threes: 41.0%
eFG%: 58.2%

TO%: 22.5%
DR%: 65.6%
FT Rate: 29.5%
dRating: 1.07

A/to: 1.47:1
Asst Rate: 80.6%
Block %: 8.2% (13.9% of twos)

Efficiency difference: -.071 (opposing two efficiency much better than opposing three efficiency)

Considering that Marvin only played 10 minutes, this isn't bad against the #2 team in the country. But their 2-pt efficiency was really good (for them, bad for us). Lucky for us (or maybe it had little to do with luck), they took a lot of threes.

[Mtn.Devil.91.92.01.10.15]

Duke defensive stats vs. Michigan State:

2-point%: .611
3-point%: .360
%threes: 41.0%
eFG%: 58.2%

TO%: 22.5%
DR%: 65.6%
FT Rate: 29.5%
dRating: 1.07

A/to: 1.47:1
Asst Rate: 80.6%
Block %: 8.2% (13.9% of twos)

Efficiency difference: -.071 (opposing two efficiency much better than opposing three efficiency)

Considering that Marvin only played 10 minutes, this isn't bad against the #2 team in the country. But their 2-pt efficiency was really good (for them, bad for us). Lucky for us (or maybe it had little to do with luck), they took a lot of threes.

It sure felt like MSU hit more than 36% of threes when watching in real time.

[TKG]

Getting back on D and preventing uncontested layups - or, "rim runs" as the kids today like to say - might help our defensive efficiency.

[Mtn.Devil.91.92.01.10.15]

Getting back on D and preventing uncontested layups - or, "rim runs" as the kids today like to say - might help our defensive efficiency.

Seriously. Felt like 60% of MSU points were on breaks or threes. They sure didn't get to the rim on sets like teams in past years.

I was really pleased with our half court D against a solid, experienced, and disciplined team.

[COYS]

Duke defensive stats vs. Michigan State:

2-point%: .611
3-point%: .360
%threes: 41.0%
eFG%: 58.2%

TO%: 22.5%
DR%: 65.6%
FT Rate: 29.5%
dRating: 1.07

A/to: 1.47:1
Asst Rate: 80.6%
Block %: 8.2% (13.9% of twos)

Efficiency difference: -.071 (opposing two efficiency much better than opposing three efficiency)

Considering that Marvin only played 10 minutes, this isn't bad against the #2 team in the country. But their 2-pt efficiency was really good (for them, bad for us). Lucky for us (or maybe it had little to do with luck), they took a lot of threes.

There was a moment in the first half where our 2-point defense was really, really good after we gave up those first three easy fast-break buckets. We didn't actually block any shots during this span, but our bigs altered a few and we forced some weird jumpers/floaters instead of letting MSU players get to the rim. Then Marvin got hurt and fouls started to mount for some of our guys, which seemed to limit their ability to contest shots. Also, because of Marvin's absence and foul trouble, we used a few less-than-idea lineups. Of course, in the second half, our bigs struggled with foul trouble for the entire half, so those same problems persisted.

I'm hoping that the high 2pt% is related to the combination of foul trouble and playing without Marvin. Getting Marvin back should improve both (we'll have an extra five fouls to give in the post and we won't have to play a lineup with JGold, Vrank, AND Alex on the court at the same time in the first half of a competitive game). We might not have been excellent at preventing a high percentage from two, but I bet it would have been at least a little bit better.

[Kedsy]

Duke defensive stats vs. Southern:

2-point%: .300
3-point%: .412
%threes: 25.4%
eFG%: 38.1%

TO%: 11.6%
DR%: 71.1%
FT Rate: 22.4%
dRating: 0.88

A/to: 1:1
Asst Rate: 36.4%
Block %: 14.9% (20.0% of twos)

Efficiency difference: 23.7 (opposing three efficiency WAAAAAY better than opposing two efficiency)

We stifled their inside game but they hit their threes. Good news is they didn't take so many threes, on a percentage basis. Also, we only forced 8 turnovers, which is pretty bad against a team like Southern.

[Troublemaker]

Our two best turnover-producing games so far have been the two games we used zone -- Utah Valley and MSU. (Although for Utah Valley, we only used it for the last 60% of the game or so).

As I wrote elsewhere, I have a strong gut feeling this pattern will continue where zone produces more opponent turnovers than man.

[MChambers]

Our two best turnover-producing games so far have been the two games we used zone -- Utah Valley and MSU. (Although for Utah Valley, we only used it for the last 60% of the game or so).

As I wrote elsewhere, I have a strong gut feeling this pattern will continue where zone produces more opponent turnovers than man.
It's certainly the case that Duke's man defense hasn't produced as many turnovers in recent years as in the past. You'd think this roster might be able to produce more, with long, quick players, but maybe not.

[Kedsy]

Duke defensive stats vs. Furman:

2-point%: .457
3-point%: .310
%threes: 45.3%
eFG%: 46.1%

TO%: 19.4%
DR%: 66.7%
FT Rate: 9.4%
dRating: 0.94

A/to: 1.08:1
Asst Rate: 52.0%
Block %: 12.5% (22.9% of twos)

Efficiency difference: 0.8, basically a wash between twos and threes.

Considering how good our offense was tonight, the defense was more than adequate.

[Skitzle]

Duke defensive stats vs. Furman:

2-point%: .457
3-point%: .310
%threes: 45.3%
eFG%: 46.1%

TO%: 19.4%
DR%: 66.7%
FT Rate: 9.4%
dRating: 0.94

A/to: 1.08:1
Asst Rate: 52.0%
Block %: 12.5% (22.9% of twos)

Efficiency difference: 0.8, basically a wash between twos and threes.

Considering how good our offense was tonight, the defense was more than adequate.

Kedsy Knowing youre a stats geek. Do you have a chart of all the games, or a whole season average?

[Kedsy]

Kedsy Knowing youre a stats geek. Do you have a chart of all the games, or a whole season average?

Right now I have both. As the season advances, I'll probably stop keeping the game-by-game.

[DavidBenAkiva]

Duke defensive stats vs. Furman:

2-point%: .457
3-point%: .310
%threes: 45.3%
eFG%: 46.1%

TO%: 19.4%
DR%: 66.7%
FT Rate: 9.4%
dRating: 0.94

A/to: 1.08:1
Asst Rate: 52.0%
Block %: 12.5% (22.9% of twos)

Efficiency difference: 0.8, basically a wash between twos and threes.

Considering how good our offense was tonight, the defense was more than adequate.

What percentage of the game against Furman was played in zone D? I missed the first 10 minutes of the game. There couldn't have been more than 5 minutes of zone in the second half. That Duke got the TO% up near 20% in primarily man-to-man was a good sign. Furman is very turnover prone, but at least it was better than against Southern.

[Kedsy]

What percentage of the game against Furman was played in zone D? I missed the first 10 minutes of the game. There couldn't have been more than 5 minutes of zone in the second half. That Duke got the TO% up near 20% in primarily man-to-man was a good sign. Furman is very turnover prone, but at least it was better than against Southern.

I believe we played man pretty much the entire first half.

[Kedsy]

Duke defensive stats vs. Portland State:

2-point%: .409
3-point%: .364
%threes: 42.9%
eFG%: 46.8%

TO%: 9.4%
DR%: 69.4%
FT Rate: 15.6%
dRating: 1.08

A/to: 3:1
Asst Rate: 70.0%
Block %: 11.7% (20.5% of twos)

Efficiency difference: 13.7 (opposing three efficiency WAY better than opposing two efficiency).

A somewhat subpar defensive performance, mostly because of how few turnovers we were able to force.

EDIT: Just looked at the box score again. The fact that Wendell played only 21 minutes, and perhaps more importantly that Javin played just 8 minutes, probably had a lot to do with our mediocre defensive performance. Not sure why either of them played so little; hopefully for our defense they'll both get more minutes going forward.

[uh_no]

Duke defensive stats vs. Portland State:

2-point%: .409
3-point%: .364
%threes: 42.9%
eFG%: 46.8%

TO%: 9.4%
DR%: 69.4%
FT Rate: 15.6%
dRating: 1.08

A/to: 3:1
Asst Rate: 70.0%
Block %: 11.7% (20.5% of twos)

Efficiency difference: 13.7 (opposing three efficiency WAY better than opposing two efficiency).

A somewhat subpar defensive performance, mostly because of how few turnovers we were able to force.

EDIT: Just looked at the box score again. The fact that Wendell played only 21 minutes, and perhaps more importantly that Javin played just 8 minutes, probably had a lot to do with our mediocre defensive performance. Not sure why either of them played so little; hopefully for our defense they'll both get more minutes going forward.

trying to keep them fresh for tomorrow? best I can come up with.

[juise]

trying to keep them fresh for tomorrow? best I can come up with.

Javin took a really hard fall in the first half and didn’t come back for a long time. It looked like he fell on his hip.

At the game, I couldn’t understand why Carter wasn’t playing in the first half. I was checking DBR to see if they had said anything on the broadcast (thinking maybe he was hurt/ill).

[El_Diablo]

Javin took a really hard fall in the first half and didn’t come back for a long time. It looked like he fell on his hip.

At the game, I couldn’t understand why Carter wasn’t playing in the first half. I was checking DBR to see if they had said anything on the broadcast (thinking maybe he was hurt/ill).

I don’t think it was all that complicated; Bolden was outplaying him.

[Kedsy]

Duke defensive stats vs. Texas:

2-point%: .553
3-point%: .174
%threes: 32.9%
eFG%: 45.7%

TO%: 14.3%
DR%: 69.6%
FT Rate: 30.0%
dRating: 1.01

A/to: 1.27:1
Asst Rate: 46.7%
Block %: 4.3% (6.4% of twos)

Efficiency difference: -29.2 (opposing two efficiency WAAAAAAY better than opposing three efficiency).

The fact that Javin only played 5 minutes almost certainly had a negative impact on our defense.

The good news is, in the last 15 minutes (last 10 of regulation plus OT), our defensive efficiency was an outstanding 0.61, so at least we were able to hunker down and defend at the end.

[juise]

To add some balance, since efficiency is kind of depressing, I thought I would give a quick update on blocked shots. Someone brought it up in a thread a few weeks ago and I have been curious how we're tracking since. As a quick refresher, the most blocked shots Duke has had in a season are:

1. 1998-99 245 (6.28/game - 39 games)
2. 2003-04 240
3. 2004-05 220
4. 1993-94 204
5. 2005-06 202


This year, Duke has 42 blocks in 7 games (6 per game).
If Duke plays an extremely conservative 33 total games, they will need to average 7.84 for the remaining 26 to top the 98-99 squad, who also owns the ACC record.
If Duke can stem the defensive efficiency issues and gets to play 39 games like the record holders, they will need to average 6.38 the rest of the way.

It will be fun to see if these young devils can give Brand/Battier/Williams/Randolph/Parks/Hill a run for their money.

[arnie]

I don’t think it was all that complicated; Bolden was outplaying him.

I didn’t see Bolden outplaying anyone last night. IMHO he appeared slower to react than normal and still uncomfortable on O. Of course, i wasn’t watching Javin closely, so maybe they both were off.

[Troublemaker]

To add some balance, since efficiency is kind of depressing, I thought I would give a quick update on blocked shots. Someone brought it up in a thread a few weeks ago and I have been curious how we're tracking since. As a quick refresher, the most blocked shots Duke has had in a season are:

<snip>

This year, Duke has 42 blocks in 7 games (6 per game).
If Duke plays an extremely conservative 33 total games, they will need to average 7.84 for the remaining 26 to top the 98-99 squad, who also owns the ACC record.
If Duke can stem the defensive efficiency issues and gets to play 39 games like the record holders, they will need to average 6.38 the rest of the way.

It will be fun to see if these young devils can give Brand/Battier/Williams/Randolph/Parks/Hill a run for their money.

Technically, Duke hasn't been all that inefficient on D this season. We're giving up about 0.96 points per possession (both raw and adjusted, according to kenpom), which is good for around 35th in the country. Interestingly, kenpom's preseason projection for our D was 35th in the country, and for those that follow his numbers closely, you know that he's repeatedly nailed or come close to nailing our game to game projections for points per possession. It's been sort of eerie, actually. All of which is a long-winded way to say that we were projected to be 35th in the country, and our D has played like it IS 35th in the country.

Not horrible, but only "pretty good." But I can see why someone would think it's been horrible because certain individual possessions have been terrible indeed.

The question is where does it all go from here? With all this athletic talent and length, I'm betting up. I see us eventually performing like a top 20 D, maybe a top 10 D. Defensive upside is the difference between this team and some of Duke's teams in the recent past that have struggled with defense.

[Newton_14]

I didn’t see Bolden outplaying anyone last night. IMHO he appeared slower to react than normal and still uncomfortable on O. Of course, i wasn’t watching Javin closely, so maybe they both were off.
I believe the OP was referring to the Thursday game against PSU arnie. Bolden did play well in that game. I agree with you regarding his overall play against Texas, but he did have one brief stretch where he played well, blocking two shots. The ref made a terrible foul call on the second block thought.

However, Bagley and Carter were also really bad on defense against Texas, mainly their help defense. Once the team can get some practices in (they've only had 2 practices since the Mich St game, some of our defensive woes can get corrected). I feel like it will. Too much talent and length to not improve the defense a lot imo.

[DukieTiger]

I believe the OP was referring to the Thursday game against PSU arnie. Bolden did play well in that game. I agree with you regarding his overall play against Texas, but he did have one brief stretch where he played well, blocking two shots. The ref made a terrible foul call on the second block thought.

However, Bagley and Carter were also really bad on defense against Texas, mainly their help defense. Once the team can get some practices in (they've only had 2 practices since the Mich St game, some of our defensive woes can get corrected). I feel like it will. Too much talent and length to not improve the defense a lot imo.

Usually,I hate the big gaps between games in December, because I have felt that it’s thrown several hot-starting Duke teams off their rhythm. This year is a different story. They might still be sloppy in late Dec and early Jan, but I’m excited about the prospects of some rest and extended practice time for this particular Duke team. Have to agree that it will help the defense.

[Kedsy]

Duke defensive stats vs. Florida:

2-point%: .431
3-point%: .400
%threes: 28.2%
eFG%: 47.9%

TO%: 11.6%
DR%: 72.1%
FT Rate: 28.2%
dRating: 1.08

A/to: 1.22:1
Asst Rate: 15.5%
Block %: 5.6% (7.8% of twos)

Efficiency difference: 16.9 (opposing three efficiency WAY better than opposing two efficiency).

108 isn't bad against a top 10 offense like Florida's. And nice to know that if we really try to keep an opponent off the three-point line, we can be relatively successful (Florida only took 28.2% of their shots from distance, the lowest percentage of any of our opponents so far this season, and well below the 38.1% of their shots they'd taken from three in their first five games). But we just don't turn teams over (this was the fourth game out of eight that our opponent's TO% was under 15%). Also, while it seemed to me Florida dominated the offensive boards, our DR% of 72.1% was our 2nd-best defensive rebounding performance of the year.

[pfrduke]

Duke defensive stats vs. Florida:

2-point%: .431
3-point%: .400
%threes: 28.2%
eFG%: 47.9%

TO%: 11.6%
DR%: 72.1%
FT Rate: 28.2%
dRating: 1.08

A/to: 1.22:1
Asst Rate: 15.5%
Block %: 5.6% (7.8% of twos)

Efficiency difference: 16.9 (opposing three efficiency WAY better than opposing two efficiency).

108 isn't bad against a top 10 offense like Florida's. And nice to know that if we really try to keep an opponent off the three-point line, we can be relatively successful (Florida only took 28.2% of their shots from distance, the lowest percentage of any of our opponents so far this season, and well below the 38.1% of their shots they'd taken from three in their first five games). But we just don't turn teams over (this was the fourth game out of eight that our opponent's TO% was under 15%). Also, while it seemed to me Florida dominated the offensive boards, our DR% of 72.1% was our 2nd-best defensive rebounding performance of the year.

I see UF as having a 36.7% assist rate (11 assists on 30 made FGs) - still very low (and lower than it seemed like, particularly given how well the Gators seemed to move the ball), but not the extreme level of one-on-one play that 15.5% suggests we forced.

[proelitedota]

Didn't budge on KP after the win. Played to his prediction exactly.

[Kedsy]

I see UF as having a 36.7% assist rate (11 assists on 30 made FGs) - still very low (and lower than it seemed like, particularly given how well the Gators seemed to move the ball), but not the extreme level of one-on-one play that 15.5% suggests we forced.

Yes, of course you're right. It was late and I accidentally applied the 11 assists to all UF FG attempts.

[Jeffrey]

However, Bagley and Carter were also really bad on defense against Texas, mainly their help defense.

Very true! IMO, Carter played poor D many times during the tournament. IMO, Bagley's amazing hops offset some of his positioning mistakes.

[mbwalker]

Technically, Duke hasn't been all that inefficient on D this season. We're giving up about 0.96 points per possession (both raw and adjusted, according to kenpom), which is good for around 35th in the country. Interestingly, kenpom's preseason projection for our D was 35th in the country, and for those that follow his numbers closely, you know that he's repeatedly nailed or come close to nailing our game to game projections for points per possession. It's been sort of eerie, actually. All of which is a long-winded way to say that we were projected to be 35th in the country, and our D has played like it IS 35th in the country.

What data does KenPom use to make that prediction? This year's team and last year's team are drastically different.

[uh_no]

What data does KenPom use to make that prediction? This year's team and last year's team are drastically different.

he uses a myriad of values....the point is that regardless of his method, he seems to be spot on so far WRT duke's defense.

[flyingdutchdevil]

What data does KenPom use to make that prediction? This year's team and last year's team are drastically different.

Their effectiveness on defense (or lack thereof) is not drastically different.

[Jeffrey]

Technically, Duke hasn't been all that inefficient on D this season. We're giving up about 0.96 points per possession (both raw and adjusted, according to kenpom), which is good for around 35th in the country. Interestingly, kenpom's preseason projection for our D was 35th in the country, and for those that follow his numbers closely, you know that he's repeatedly nailed or come close to nailing our game to game projections for points per possession. It's been sort of eerie, actually. All of which is a long-winded way to say that we were projected to be 35th in the country, and our D has played like it IS 35th in the country.

Not horrible, but only "pretty good." But I can see why someone would think it's been horrible because certain individual possessions have been terrible indeed.

The question is where does it all go from here? With all this athletic talent and length, I'm betting up. I see us eventually performing like a top 20 D, maybe a top 10 D. Defensive upside is the difference between this team and some of Duke's teams in the recent past that have struggled with defense.

I think K deserves a lot of the credit for us currently being around 35th. K's defensive substitutions and play calling have repeatedly changed negative momentum. IMO, this season, so far, may be the best in-game defensive coaching of K's career.

I'm very impressed, given K is not (yet) a great X's and O's coach.

[mbwalker]

he uses a myriad of values...the point is that regardless of his method, he seems to be spot on so far WRT duke's defense.

A myriad of values? Anything more specific? If he's basing his projections on last year, they're pretty close to meaningless. And if they're spot on based on last year, it's luck, not anything scientific.

And if he isn't basing it on last year, what is the preseason projection based on? Just curious.

[uh_no]

A myriad of values? Anything more specific? If he's basing his projections on last year, they're pretty close to meaningless. And if they're spot on based on last year, it's luck, not anything scientific.

And if he isn't basing it on last year, what is the preseason projection based on? Just curious.

https://kenpom.com/blog/preseason-ratings-now-with-transfers/

[El_Diablo]

I believe the OP was referring to the Thursday game against PSU arnie. Bolden did play well in that game. I agree with you regarding his overall play against Texas, but he did have one brief stretch where he played well, blocking two shots. The ref made a terrible foul call on the second block thought.

However, Bagley and Carter were also really bad on defense against Texas, mainly their help defense. Once the team can get some practices in (they've only had 2 practices since the Mich St game, some of our defensive woes can get corrected). I feel like it will. Too much talent and length to not improve the defense a lot imo.

Yes, I was saying that in the first half of the PSU game Carter was relatively ineffective (zero points until the last ~30 seconds, zero rebounds, zero blocks) compared to Bolden, who came in and had six points and eight rebounds by halftime, along with a couple blocks. Carter definitely stepped it up in the second half though.

[-jk]

Yes, I was saying that in the first half of the PSU game Carter was relatively ineffective (zero points until the last ~30 seconds, zero rebounds, zero blocks) compared to Bolden, who came in and had six points and eight rebounds by halftime, along with a couple blocks. Carter definitely stepped it up in the second half though.

Freshmen. We can hope they'll be Sophs some day...

-jk

[flyingdutchdevil]

Freshmen. We can hope they'll be Sophs some day...

-jk

Not at Duke, my friend

[Kedsy]

Duke defensive stats vs. Indiana:

2-point%: .658
3-point%: .238
%threes: 35.6%
eFG%: 55.1%

TO%: 13.4%
DR%: 61.8%
FT Rate: 44.1%
dRating: 1.20

A/to: 1.33:1
Asst Rate: 40.0%
Block %: 5.1% (7.9% of twos)

Efficiency difference: -30.1 (opposing two efficiency WAAAAAAAAAAAAAAAAAAAAAY better than opposing three efficiency).

By far our worst defensive performance of the young season, featuring our worst defensive rebounding performance of the season. Still not turning anybody over, and fouling a lot to boot. Lucky for us Indiana shot 23.8% from three and 61.5% from the line. Duke seemed to hunker down at crunch time again, but not a lot of good things to say about our D tonight.

[COYS]

Duke defensive stats vs. Indiana:

2-point%: .658
3-point%: .238
%threes: 35.6%
eFG%: 55.1%

TO%: 13.4%
DR%: 61.8%
FT Rate: 44.1%
dRating: 1.20

A/to: 1.33:1
Asst Rate: 40.0%
Block %: 5.1% (7.9% of twos)

Efficiency difference: -30.1 (opposing two efficiency WAAAAAAAAAAAAAAAAAAAAAY better than opposing three efficiency).

By far our worst defensive performance of the young season, featuring our worst defensive rebounding performance of the season. Still not turning anybody over, and fouling a lot to boot. Lucky for us Indiana shot 23.8% from three and 61.5% from the line. Duke seemed to hunker down at crunch time again, but not a lot of good things to say about our D tonight.

I won't lie, the fact that we just aren't forcing any turnovers this year is very surprising to me. I did not expect us to ever return to the heights of the early 2000's in terms of turnovers forced, but I did not expect this team to be so anemic at it. It makes it so much harder to have strong efficiency numbers when the other team is taking a shot every single time down the floor.

Hopefully we'll be able to look back and chalk the Indiana game up to fatigue and a crazy travel schedule. Also, hopefully the extra practice time the team has over the next month will show itself on the defensive end.

[jv001]

I'm making this statement without looking back at the stats from Michigan State game. I seem to remember we forced several turnovers in our zone, which seems out of the norm. I always thought our man2man was the defense that forced turnovers. Just my 2 cents on the lack of turnovers. GoDuke!

[hallcity]

I wonder what would happened if Duke had played zone defense the whole game. I'm guessing it would have been an easy 20 point win. Was K playing M2M because he wanted to force the Duke team to get better at M2M for the future rather than because it was the best defense for the team to play in this game? If so, either he didn't care that much about the final result against Indiana or he was confident that Duke would win even though they weren't playing their best defense.

[uh_no]

I wonder what would happened if Duke had played zone defense the whole game. I'm guessing it would have been an easy 20 point win. Was K playing M2M because he wanted to force the Duke team to get better at M2M for the future rather than because it was the best defense for the team to play in this game? If so, either he didn't care that much about the final result against Indiana or he was confident that Duke would win even though they weren't playing their best defense.

yeah, if someone had time, they'd compute our efficiency in either set....

[Hingeknocker]

I won't lie, the fact that we just aren't forcing any turnovers this year is very surprising to me. I did not expect us to ever return to the heights of the early 2000's in terms of turnovers forced, but I did not expect this team to be so anemic at it. It makes it so much harder to have strong efficiency numbers when the other team is taking a shot every single time down the floor.

Hopefully we'll be able to look back and chalk the Indiana game up to fatigue and a crazy travel schedule. Also, hopefully the extra practice time the team has over the next month will show itself on the defensive end.

There are enough possessions where we've shown lockdown defense for me to be optimistic and hope that we can elevate the defense to somewhere around an acceptable level. Getting some rest and practice time will be an enormous help, so I'm particularly looking for those post-finals games to see how much we've improved on that end of the floor.

[Billy Dat]

I wonder what would happened if Duke had played zone defense the whole game. I'm guessing it would have been an easy 20 point win. Was K playing M2M because he wanted to force the Duke team to get better at M2M for the future rather than because it was the best defense for the team to play in this game? If so, either he didn't care that much about the final result against Indiana or he was confident that Duke would win even though they weren't playing their best defense.

My interpretation of the K quote below is, basically, like you said.

On Indiana’s hot start to the second half:
“Our defense in the last 10 minutes of games or eight minutes has been better than before. We have four freshmen out there all the time. They are
just learning how to manage a game as far as the effort, the intensity. And when they learn how to play in four minute stretches, we will improve. We
kind of pace because they know they are going to play minutes. And they have played a lot of games. So they are a little bit tired, but at the end, they
know there is not another four minutes. Those are the last four minutes. In our games, we've been the better team in the last for minutes of our
games and it's a good four minutes to be the better team in.”

“Like Gary Trent, he could not hit a shot. Then he got that steal, then he got fouled and hit two free throws. And then when we were going to Marvin
and Grayson, Marvin had the ball in the low post, opposite of our bench, they were not watching. Gary just flashed into the middle and got a big threepoint
play. I'm proud of him because freshman if they not hitting right, they can go that's it, it's not a good day. That's not him. We have a really good
group.”

[Troublemaker]

I wonder what would happened if Duke had played zone defense the whole game. I'm guessing it would have been an easy 20 point win. Was K playing M2M because he wanted to force the Duke team to get better at M2M for the future rather than because it was the best defense for the team to play in this game? If so, either he didn't care that much about the final result against Indiana or he was confident that Duke would win even though they weren't playing their best defense.

For some reason, Duke had scouted Indiana as being a very dangerous 3-pt shooting team. In the postgame radio interviews, Bolden mentioned how hard they had stressed taking away IU's 3-pt shots.

So, I think when Duke shifted to zone, it was purely for foul trouble purposes; I believe both Gary and Trevon had two fouls at that point. But Duke was always going to shift back to M2M.

It's a questionable scout of the Hoosiers, to be frank. I don't think they're that dangerous from long range. But the alternative theory that Coach was willing to take a loss to practice M2M is much less plausible to me. I can believe we scouted a team wrong and thought that M2M would be better against them. (It's also possible we scouted correctly, and the Hoosiers would've won if we stuck with zone).

[jv001]

For some reason, Duke had scouted Indiana as being a very dangerous 3-pt shooting team. In the postgame radio interviews, Bolden mentioned how hard they had stressed taking away IU's 3-pt shots.

So, I think when Duke shifted to zone, it was purely for foul trouble purposes; I believe both Gary and Trevon had two fouls at that point. But Duke was always going to shift back to M2M.

It's a questionable scout of the Hoosiers, to be frank. I don't think they're that dangerous from long range. But the alternative theory that Coach was willing to take a loss to practice M2M is much less plausible to me. I can believe we scouted a team wrong and thought that M2M would be better against them. (It's also possible we scouted correctly, and the Hoosiers would've won if we stuck with zone).

I'm with you because Coach K hates to lose but I do believe he might have stuck to the man for the reason you gave. Anyway, we got a win in a tough venue and got more experience playing Duke man2man as well. GoDuke!

[Kedsy]

Duke defensive stats vs. South Dakota:

2-point%: .435
3-point%: .455
%threes: 32.4%
eFG%: 51.5%

TO%: 13.1%
DR%: 78.0%
FT Rate: 22.1%
dRating: 1.05

A/to: 1.80:1
Asst Rate: 60.0%
Block %: 8.8% (13.0% of twos)

Efficiency difference: 24.8 (opposing three efficiency WAAAAY better than opposing two efficiency).

Better defense than against Indiana, but still pretty mediocre. I think our inability to force turnovers can be characterized as a trend. At least our rebounding came back this game. And during the first half we put up a defensive rating of 0.86, so maybe the overall performance is tainted by an inability to focus on defense when you have a 26-point halftime lead.

[Kedsy]

Duke defensive stats vs. St. Francis (PA):

2-point%: .415
3-point%: .233
%threes: 42.3%
eFG%: 38.7%

TO%: 8.7%
DR%: 89.8%
FT Rate: 21.1%
dRating: 0.84

A/to: 1.57:1
Asst Rate: 45.8%
Block %: 5.6% (9.8% of twos)

Efficiency difference: -6.6 (opposing two efficiency a little better than opposing three efficiency).

Not sure how much you can take away from a total beatdown like this, but a pretty good performance all around, especially defensive rebounding which was a fraction under 90%.

Bad news for COYS, though. Even in a dominating performance like this, Duke didn't turn its opponent over. In fact, this was our worst defensive turnover percentage performance of the season.

[proelitedota]

Are TOs that important in today's landscape? I think rebounding and generating more possessions is much reliable and effective.

[Kedsy]

Are TOs that important in today's landscape? I think rebounding and generating more possessions is much reliable and effective.

I disagree. In possessions that don't end in turnovers, our opponents have scored 1.18 points per possession against us this season, which would be the worst defense in the nation if we didn't force any turnovers (because, obviously, the opponent doesn't score at all in possessions that do end in turnovers). Counting what few turnovers we've managed to force, our overall defensive ppp is a tad shy of 1.00 (813 points in 815 possessions).

We'd like that overall ppp to get down to 0.90, or at worst 0.95. Other than turnovers, there are only three ways to get that points per possession number lower:

(a) Get more defensive rebounds, except our defensive rebounding has been pretty strong, currently at 72.1%. That would be an all-time Duke record for DR% (dating back to 1987, when they first started keeping track of offensive and defensive rebounds), if the percentage holds all season. Meaning we're not likely to get more defensive rebounds than we're getting now.

(b) Foul less, except our opposing free throw rate of 20.9% would also be an all-time Duke record, by a lot (again, dating back to 1987). Last season, for example, our opposing free throw rate was 31.2%. Again, it's not likely we can improve on this as the season progresses. Much more likely that number goes up.

(c) Make opponents shoot worse. Currently, opponents have an eFG% of 47.1% against us, which is more or less in line with five of our last six seasons and pretty much in the middle of Duke performances, historically. Clearly this could be improved, but is hoping to bother enough shots to significantly decrease our opponents' eFG% more "reliable and effective" than turning them over? I can't imagine that it is.

Our opposing turnover percentage is just 15.2% this season, which would be (by far) the worst such Duke performance ever (again going back to 1987). Seems to me there's plenty of room for improvement, which is good because improving in this area is not only important, it's probably the only way to significantly improve our defensive efficiency.

[cato]

In possessions that don't end in turnovers, our opponents have scored 1.18 points per possession against us this season, which would be the worst defense in the nation if we didn't force any turnovers . . .

Is that worst in the nation compared to other teams in possessions that don’t end in turnovers?

[Ultrarunner]

Our opposing turnover percentage is just 15.2% this season, which would be (by far) the worst such Duke performance ever (again going back to 1987). Seems to me there's plenty of room for improvement, which is good because improving in this area is not only important, it's probably the only way to significantly improve our defensive efficiency.

I wonder if the coaches are preaching a "stay home, lock in your defensive rotations" message right now to solidify the MTM defense with an eye to opening up the defense to challenge passes and harass dribblers more later. Certainly we have had very opportune thefts this season to win ball games so the base talent is there.

[jv001]

I wonder if the coaches are preaching a "stay home, lock in your defensive rotations" message right now to solidify the MTM defense with an eye to opening up the defense to challenge passes and harass dribblers more later. Certainly we have had very opportune thefts this season to win ball games so the base talent is there.

Very good point. Coach K could be preaching "a stay home" approach to our man to man defense this season. Duke's man2man is hard enough to learn even if you are not a freshman. I've seen too many instances where one of our perimeter players gambles for a steal and help defense never comes like it should. I believe that could be one reason there's less takeaways this season. GoDuke!

[BandAlum83]

Very good point. Coach K could be preaching "a stay home" approach to our man to man defense this season. Duke's man2man is hard enough to learn even if you are not a freshman. I've seen too many instances where one of our perimeter players gambles for a steal and help defense never comes like it should. I believe that could be one reason there's less takeaways this season. GoDuke!

I saw it last night. Grayson went for a steal, missed it and his man got the ball for an open 3 that he hit.

[Kedsy]

Is that worst in the nation compared to other teams in possessions that don’t end in turnovers?

I can't say for sure, but almost certainly not. There are teams that are almost that bad including turnovers.

I was just responding to the question of how important are turnovers to efficient defense.

[gam7]

One defensive thing I am seeing this year that I love is our having some success (so far) taking away other teams' #1 option on offense. We used to do that with consistent success almost every single game (up until the late 2000s), but have not been able to do it in recent years.

[Troublemaker]

Our opposing turnover percentage is just 15.2% this season, which would be (by far) the worst such Duke performance ever (again going back to 1987). Seems to me there's plenty of room for improvement, which is good because improving in this area is not only important, it's probably the only way to significantly improve our defensive efficiency.

I'm not sure the relationship between the Four Factors and Defensive Efficiency is as strong as you're making it out to be here. At least as measured by kenpom.

As a long-time observer of kenpom, the way to improve our DefEff is to play a good defensive game against a good offense, which will bump up our defensive rank a bit. Then follow that up by playing another good defensive game against a good offense, which will bump up our defensive rank some more. Two good defensive games won't change any of our Four Factors (including turnover pct) all that much, but all of a sudden we could find ourselves ranked in the 30s instead of the 50s on defense. And in those two games, we might not turn them over well at all; we might just hold them to a low shooting percentage.

Hopefully I'm making my point well here. Put another way, the 2015 team had the 11th-best defensive efficiency in the country, as measured by kenpom. But its Four Factors were sort of middling: 204th in turnover pct, 125th in defensive rebounding, 70th in effective FG%, and 4th in free throw rate. Except for avoiding fouls, we weren't really elite in anything defensively. But what we did do was play great defense those final 6 games of the season against great competition to raise our defensive efficiency.

[budwom]

^ along those lines, I'll feel a lot better about the validity of these advanced stats once we get into some league play. With much of the country loading up on weak teams early on, I just don't take the data too seriously yet. By the end of January I think it'll mean vastly more.

[DukieTiger]

I'm not sure the relationship between the Four Factors and Defensive Efficiency is as strong as you're making it out to be here. At least as measured by kenpom.

As a long-time observer of kenpom, the way to improve our DefEff is to play a good defensive game against a good offense, which will bump up our defensive rank a bit. Then follow that up by playing another good defensive game against a good offense, which will bump up our defensive rank some more. Two good defensive games won't change any of our Four Factors (including turnover pct) all that much, but all of a sudden we could find ourselves ranked in the 30s instead of the 50s on defense. And in those two games, we might not turn them over well at all; we might just hold them to a low shooting percentage.

Hopefully I'm making my point well here. Put another way, the 2015 team had the 11th-best defensive efficiency in the country, as measured by kenpom. But its Four Factors were sort of middling: 204th in turnover pct, 125th in defensive rebounding, 70th in effective FG%, and 4th in free throw rate. Except for avoiding fouls, we weren't really elite in anything defensively. But what we did do was play great defense those final 6 games of the season against great competition to raise our defensive efficiency.

One might argue that part of a “good defensive game” is forcing turnovers.

In the national championship season you mentioned, KenPom gave a fairly strong negative correlation between Duke’s defensive efficiency and defensive turnover percentage (-.59). Meaning, there was a decent relationship between Duke forcing turnovers at an increased rate and the Duke limiting opposing teams’ offensive efficiency. That relationship was actually mathematically stronger than Duke avoiding fouls, even if opponent FT% was a more elite stat for the team that year.

[Troublemaker]

One might argue that part of a “good defensive game” is forcing turnovers.

Yes, no doubt. This is why I wrote that I hope I'm making my point well; I certainly am not saying forcing more turnovers won't help.



In the national championship season you mentioned, KenPom gave a fairly strong negative correlation between Duke’s defensive efficiency and defensive turnover percentage (-.59).

Hmmm, I wonder how the stat gurus here feel about you describing -.59 as "fairly strong negative correlation." Regardless, I'm sure you noticed that the correlation between turnover pct and DefEff this season is -.08 (i.e. basically no relationship so far).

[Ian]

Hopefully I'm making my point well here. Put another way, the 2015 team had the 11th-best defensive efficiency in the country, as measured by kenpom. But its Four Factors were sort of middling: 204th in turnover pct, 125th in defensive rebounding, 70th in effective FG%, and 4th in free throw rate. Except for avoiding fouls, we weren't really elite in anything defensively. But what we did do was play great defense those final 6 games of the season against great competition to raise our defensive efficiency.

That's because those four factor ratings are raw numbers and not adjusted for strength of competition. If you look at the 2015 SOS Duke played the 3rd most difficult offensive schedule in the country, so the numbers you're looking at are skewed by that, and looks off when compared with the 11th best defensive efficiency which is SOS adjusted.

That's why on those unadjusted numbers it is instructive to compare with historical Duke numbers to gauge how well we are doing, because we consistently play a top 20-30 SOS.

[Kedsy]

One might argue that part of a “good defensive game” is forcing turnovers.

I'm with DukieTiger on this one. It's worth noting that Duke had strong turnover percentages against Utah (21.8%), Gonzaga (21.9%), and Michigan State (20.7%) among those six games you mention.

Adding SD State (an average 18.0% performance), those four NCAAT opponents had an aggregate of 28 assists against 52 turnovers. In contrast, our last four opponents (including two "lesser" teams) had an aggregate of 52 assists against 35 turnovers.

And while you're right, Troublemaker, that any great defensive performance will help us in KenPom's ratings, if in subsequent games we go back to average/below average, our rating will sputter again. Only two ways to keep it down: (a) play consistently good defense, which requires forcing at least a decent percentage of turnovers; or (b) end the season on a great defensive run so the rank doesn't have a chance to spike again after.

To illustrate, here's Duke's TO% and final KenPom defensive rank since he's published the rankings:



Year TO% Pom dRank
2017 17.2 47
2016 17.2 86
2015 18.5 12
2014 18.5 87
2013 20.4 26
2012 18.5 78
2011 21.0 11
2010 21.4 5
2009 23.5 31
2008 24.7 7
2007 22.1 5
2006 22.5 18
2005 21.8 2
2004 24.4 3
2003 24.4 16
2002 25.7 1


The only year where our TO% was subpar but our defensive rank was good is the 2015 season you're talking about. And that team's defensive rank was #57 going into the Tournament.

Plus, as I said before, our current TO% of 15.2% is quite a bit worse than Duke's worst season ever for this stat (an "honor" held by our previous two seasons, in a tie for worst). The fact that we're as strong as we are in KenPom dRank (#54) is a testament that we're doing something right on D. But if you want that to get into the 20s, we need to force more turnovers. Although I suppose becoming a top 5 team in opposing eFG% might also do the trick, but I don't see that happening.

To budwom, I say it's hard to imagine our TO% getting better as our competition improves. This is not a matter of comparing our performance to other teams who have feasted on cupcakes. We're comparing our current inability to turn our opponents over to past Duke teams who played (presumably) similar schedules.

[Troublemaker]

Plus, as I said before, our current TO% of 15.2% is quite a bit worse than Duke's worst season ever for this stat (an "honor" held by our previous two seasons, in a tie for worst). The fact that we're as strong as we are in KenPom dRank (#54) is a testament that we're doing something right on D. But if you want that to get into the 20s, we need to force more turnovers. Although I suppose becoming a top 5 team in opposing eFG% might also do the trick, but I don't see that happening.

You know, we probably don't disagree all that much. I'll definitely agree that any defensive improvement we make will be accompanied by an uptick in turnovers. It almost has to because right now our turnover pct is so low. But I'll predict that (1) the turnover gains will be small (unless we actually do become a zone-first team) and (2) we'll also find noticeable and comparable gains in defensive shooting percentage and maybe defensive rebounding. Now, measuring all that and figuring out what was the biggest factor to improvement might become tricky, but I'm guessing it'll be something we'll be happy to do if the defense does indeed improve. So, for now, let's wait for the hypothetical big defensive improvement to come, and then we'll do the full diagnosis of what went right.

[Kedsy]

You know, we probably don't disagree all that much. I'll definitely agree that any defensive improvement we make will be accompanied by an uptick in turnovers. It almost has to because right now our turnover pct is so low. But I'll predict that (1) the turnover gains will be small (unless we actually do become a zone-first team) and (2) we'll also find noticeable and comparable gains in defensive shooting percentage and maybe defensive rebounding. Now, measuring all that and figuring out what was the biggest factor to improvement might become tricky, but I'm guessing it'll be something we'll be happy to do if the defense does indeed improve. So, for now, let's wait for the hypothetical big defensive improvement to come, and then we'll do the full diagnosis of what went right.

OK, I'm fine to wait and see. I agree there's plenty of room for improvement in defensive shooting percentage. Not sure how much room for improvement in defensive rebounding. Maybe some, but not a lot.

[ipatent]

Down to #71 on KenPom.

[Tappan Zee Devil]

Down to #71 on KenPom.

and rightfully so

[Kedsy]

Duke defensive stats vs. BC:

2-point%: .457
3-point%: .577
%threes: 42.6%
eFG%: 63.1%

TO%: 22.6%
DR%: 67.7%
FT Rate: 24.6%
dRating: 1.18

A/to: 1.06:1
Asst Rate: 58.1%
Block %: 4.9% (8.6% of twos)

Efficiency difference: 40.9 (opposing three efficiency OMG SO MUCH better than opposing two efficiency.

Other than opposing 3-point shooting, our D really wasn't so bad. Our defensive rebounding was not great but was better than it seemed while watching the game. We even turned them over a decent rate, for the first time in a while. But, let's face it, you can't give up a 63.1% eFG% and hope to win. It seemed to me that a lot (but not all) of BC's threes were contested, they just made them at an outrageous clip. I'll be surprised if another team hits threes at that rate against us this season.

That said, for the season (including the BC game) we're allowing opponents to hit three-pointers at a 34.9% clip. The last time Duke's opponents shot better than that from three was in 1999-2000 (35.8%), though we came close in 2002-03 (34.6%) and 2000-01 (34.4%, and that season still ended up ok).

After their otherworldly performance against us, BC is hitting 34.8% of their three-pointers. If they'd only done really well in this game and, say, hit 42% of their threes, we win the game by at least 7. In other words, there may be reason for concern, but this also could have just been a fluke.

[Kedsy]

I copied this from the game thread, because I think the conversation makes more sense here.

This chart shows Duke's opponents' three-point attempts as a percentage of total opponents' shots, since the three-point shot was adopted, from highest to lowest:



Year % threes
2018 0.380
2017 0.288
2015 0.278
2014 0.270
2013 0.269
2012 0.263
2011 0.260
2010 0.256
2009 0.254
2008 0.252
2007 0.251
2006 0.250
2005 0.250
2004 0.246
2003 0.246
2002 0.244
2001 0.243
2000 0.242
2016 0.241
2014 0.241
2012 0.241
2011 0.234
2010 0.231
2009 0.213
2008 0.209
2007 0.204
2006 0.199
2005 0.194
2004 0.182
2003 0.163
2002 0.154
2001 0.139


So far Duke's opponents have taken 38% of their shots from downtown this season, while the 2nd highest Duke full-season performance was less than 29%. That's a huge difference.

I also did a game-by-game analysis to see if the percentage of shots taken from three-ville is related to our success at stopping the two-point shot. In other words, when we allow opponents to take more threes, are we shutting down their inside game in compensation?



Opponent %three 2pt%
Elon 52.4 36.7
Furman 45.3 45.7
Portland St 42.9 40.9
BC 42.6 45.7
St Francis 42.3 41.5
Mich St 41.0 61.1
UVU 37.3 45.2
Indiana 35.6 65.8
Texas 32.9 55.3
S Dakota 32.4 43.5
Florida 28.2 43.1
Southern 25.4 30.0


The answer looks like sort of, but not really. We seemed to make an effort to limit three-attempts against Indiana and Texas, and they responded by making a high percentage of their two-point shots, and the top five on the chart were all worse than 46% on their two-point shots, especially the jackiest team, Elon, which took more than half their shots from the three-o-sphere but made only 36.7% of their two-attempts. But Michigan State took a bunch of threes and still clobbered us inside, while Southern eschewed the three but still couldn't score inside. Not sure there's a legitimate pattern here.

[Billy Dat]

ESPN college basketball quant, John Gasaway, takes a look at Duke's defense:

http://www.espn.com/mens-college-basketball/story/_/id/21751975/duke-blue-devils-defense-struggling-season

His conclusion, basically, is that the defense is bad but the offense is so good it may not matter.

[Rich]

ESPN college basketball quant, John Gasaway, takes a look at Duke's defense:

http://www.espn.com/mens-college-basketball/story/_/id/21751975/duke-blue-devils-defense-struggling-season

His conclusion, basically, is that the defense is bad but the offense is so good it may not matter.

In the interesting, yet miserable, department, comes this statistic:

Prior to 2017-18, Duke had allowed an opponent to make more than 11 3s in a game just once in the past 16 seasons and 572 games. This season, however, Portland State (12 made 3s), Elon (14) and, of course, Boston College (15) have all eclipsed that number.

[pfrduke]

In the interesting, yet miserable, department, comes this statistic:

Prior to 2017-18, Duke had allowed an opponent to make more than 11 3s in a game just once in the past 16 seasons and 572 games. This season, however, Portland State (12 made 3s), Elon (14) and, of course, Boston College (15) have all eclipsed that number.

This has been remarked on before, but the one foundational hallmark of Coach K's defense, year over year, has been a steadfast refusal to let opponents shoot 3s. From 2002-2017, Duke never let opponents shoot 30% of their shots from 3 (we ranged from 20.8%-29.6%) and never finished out of the top 25 in terms of lowest percentage of opponent shots from beyond the arc. This season, opponents have shot 38% of their attempts from 3, which puts us at 190th (!!) in the country in avoiding 3 point attempts. I think because of the shift in team makeup, Coach concluded that we can't run teams off the 3 point line as well as we have in years past*, which has opened up all of these opportunities for teams to hit shots.

Probably relatedly (because 3 point shots are the most heavily assisted shots) the other traditional hallmark of our defense - denying assists - is also at a historic low. Only one Duke team in the past 16 years has allowed opponents to get assists on more than half their baskets (2016), and this year's team is currently even worse - 52.6% assisted baskets, ranking 184th in the country.

Maybe all this will equalize as the season goes on, but if it doesn't we're going to be vulnerable to teams getting hot from 3 and springing upsets.

*As an aside, much of the criticism in past years about Duke allowing too much penetration has been a byproduct of this defensive choice. We opted to play tighter on the perimeter to avoid 3s, opening up driving lanes as a trade-off. This season is showing that trading 2s to defend 3s isn't necessarily such a bad idea.

[uh_no]

In the interesting, yet miserable, department, comes this statistic:

Prior to 2017-18, Duke had allowed an opponent to make more than 11 3s in a game just once in the past 16 seasons and 572 games. This season, however, Portland State (12 made 3s), Elon (14) and, of course, Boston College (15) have all eclipsed that number.

staggering....that's a lot for "teams won't ever shoot as well as BC did"...didn't realize it was that bad

One of KPs observations is that good three point defense doesn't change the opponents %. Just the number that they take. If they're taking a lot, it's because they're likely less guarded....and so an unusually good shooting night is going to result in a lot of made threes. of course the effect is somewhat mitigated by the increased number of chances....but while trying to avoid sounding cheesy: you can't make a lot of threes if you don't take a lot of threes....and duke is letting them take a lot of threes

[pfrduke]

staggering...that's a lot for "teams won't ever shoot as well as BC did"...didn't realize it was that bad

One of KPs observations is that good three point defense doesn't change the opponents %. Just the number that they take. If they're taking a lot, it's because they're likely less guarded...and so an unusually good shooting night is going to result in a lot of made threes. of course the effect is somewhat mitigated by the increased number of chances...but while trying to avoid sounding cheesy: you can't make a lot of threes if you don't take a lot of threes...and duke is letting them take a lot of threes

This is borne out in our numbers over time. We have had substantial variation in how well opponents shoot 3s - everywhere from 28.3%-34% - that has not really been correlated with how many 3s the opponents took. This year we're both letting teams take and make a lot of 3s.

[Kedsy]

staggering...that's a lot for "teams won't ever shoot as well as BC did"...didn't realize it was that bad

Well, neither Elon nor Portland State shot nearly as well as BC. Elon was 14 for 33 (42.4%) and Portland State was 12 for 33 (36.4%). BC was 15 for 26 (57.7%). If BC had shot 42.3% (similar to Elon) they'd have hit four fewer three-pointers and we would likely have won the game.

So, the idea that "teams won't ever shoot as well as BC did" is still valid, in my mind. Though of course it might happen again, the odds are against it.

I agree with the rest of your post. As I said elsewhere, if we're going to allow them to attempt so many threes, and we're not going to turn them over, we're going to have to smother our opponents at the rim. On two-pointers, we held Elon to 36.7%, Portland State to 40.9%, and BC to 45.7%. Which may be one reason why we crushed Elon, struggled against Portland State, and lost to BC.

[Mtn.Devil.91.92.01.10.15]

Well, neither Elon nor Portland State shot nearly as well as BC. Elon was 14 for 33 (42.4%) and Portland State was 12 for 33 (36.4%). BC was 15 for 26 (57.7%). If BC had shot 42.3% (similar to Elon) they'd have hit four fewer three-pointers and we would likely have won the game.

So, the idea that "teams won't ever shoot as well as BC did" is still valid, in my mind. Though of course it might happen again, the odds are against it.

I agree with the rest of your post. As I said elsewhere, if we're going to allow them to attempt so many threes, and we're not going to turn them over, we're going to have to smother our opponents at the rim. On two-pointers, we held Elon to 36.7%, Portland State to 40.9%, and BC to 45.7%. Which may be one reason why we crushed Elon, struggled against Portland State, and lost to BC.

Our talented big men encourage outside shooting at a higher rate. That isn't mysterious.

[Troublemaker]

Our talented big men encourage outside shooting at a higher rate. That isn't mysterious.

We're helping off shooters more than we have in the past. And recklessly I might add. I'll try to show this in a GIF at some point.

For example, we played two bigs together in 2010, 2012, and 2013, and in all those seasons, we still depressed opposition attempts from three.

We definitely have changed our defense aside from there being atypical personnel this season.

[Mtn.Devil.91.92.01.10.15]

We're helping off shooters more than we have in the past. And recklessly I might add. I'll try to show this in a GIF at some point.

For example, we played two bigs together in 2010, 2012, and 2013, and in all those seasons, we still depressed opposition attempts from three.

We definitely have changed our defense aside from there being atypical personnel this season.

I would also wonder whether flopping from MtM and zone over and over might have a negative effect.

Or a positive effect.

But, I bet there's an effect.

[uh_no]

Well, neither Elon nor Portland State shot nearly as well as BC. Elon was 14 for 33 (42.4%) and Portland State was 12 for 33 (36.4%). BC was 15 for 26 (57.7%). If BC had shot 42.3% (similar to Elon) they'd have hit four fewer three-pointers and we would likely have won the game.

So, the idea that "teams won't ever shoot as well as BC did" is still valid, in my mind. Though of course it might happen again, the odds are against it.


Yes and no. while the probability of shooting the lights out decreases, the amount that even a slightly better than average night has increases their output is higher.

Consider (pulling numbers outta nowhere):

team shoots 40% from 3, 40% from 2. In a game against a "normal" team, lets say they take 60 shots, 40 from 2 and 20 from 3. In such a game, they'd be expected to score 56 points. If they have a "good" shooting night, and shoot, say, 60% from 3, they're expected to score 68 points.

Now lets take a look at a game against duke

because of our defense, they take 30 shots each from 3 and 2. Lets say they shoot 40% for three, but because we're big, they shoot 33% from 2. They'd be expected to score the same 56 points (conveniently chosen %s :) ). Now, though, when they have their "good" shooting night, they shoot 60%, and they score 74 points. Uh oh.

Now, this effect is tempered by the fact that they are LESS likely to have such a good shooting night because of the fact they took more shots. But we can calculate what this effect would be expected to be.

In the first case, the odds of hitting 12+ shots on 20 attempts with 40% is 5.65%

The top 5% on 30 shots is somewhere around 17+, or shooting around 57% If someone shoots 57% with the increased attempts, then they're expected to score 71 points.



So anyway, the point is, there are two effects:
1) the increased number of chances to get more points (3 vs 2)
2) decreased likelihood of having a stellar night % wise

Turns out 1 is the bigger effect.

If we play like team A, BC takes fewer 3's, even if they shoot an increased percentage (which is equally likely given fewer shots) from 3, they still end up with fewer points. One can see the effect is even more pronounced when the other team's 2 point % isn't markedly worse...as Kedsy pointed out, was the case with BC.

Bottom line: gotta prevent 3's...or as I agree with the twitter guy now that I've done this analysis, we're really susceptible to marginally good shooting nights by opponents.

[Kedsy]

Bottom line: gotta prevent 3's...or as I agree with the twitter guy now that I've done this analysis, we're really susceptible to marginally good shooting nights by opponents.

What you're saying makes sense, but I think we're just as susceptible (or at least almost as susceptible) if we're bad at two-point defense and the opponent gets hot from three in fewer attempts.

Last season, for example, we had only four games (out of 37 total games) in which our opponents shot 40% or better from three:



Opponent Shots made 3s 3 att 3% 2% %threes result
Va Tech 58 8 13 0.615 0.533 22.4% L (14 pts)
Syracuse 47 6 11 0.545 0.500 23.4% L (3 pts)
Florida 60 6 11 0.545 0.408 18.3% W (10 pts)
Miami 55 7 16 0.438 0.385 29.1% W (12 pts)


So, first of all, it doesn't happen often and, as you point out, it's less likely with more attempts. Though having said that, this season we've already had five opponents shoot better than 40% from three (Elon, .424 on 33 attempts; Southern, .412 on 17 attempts; Florida, .400 on 20 attempts; South Dakota, .455 on 22 attempts; and BC, .577 on 26 attempts). Not coincidentally, we struggled (to a lesser or greater degree) in four of those five games. It's also worth noting that we kept Southern (W by 17) and Elon (W by 31) in the 30s on two-point shots (percentage-wise); Florida (W by 3) and South Dakota (W by 16) in the low-40s on twos; and BC (L by 5) in the mid-40s on twos.

My second lesson from the table above is, even if opponents only take a quarter of their shots from three, if they're hot from outside and we also can't stop the two, we lose. (Obvious, I know.)

In last year's VT game, to make up for their hot three-point shooting, we would have had to hold them to 37.8% on two-point shooting to make up the 14 point margin, which was 11+% below our opponents' season average of 48.9% on two-pointers. In this year's BC game, to make up for their hot three-point shooting in more attempts, we would have had to hold them to 37.1% on two-point shooting to make up the 5 point margin, which is a little less than 9% below our opponents' current season average of 45.9% on two-pointers.

I get that it's not apples-to-apples, since VT beat us by a lot more points than BC did, but my point is that we're susceptible to hot outside shooting (both in more attempts and fewer attempts), depending on how good or bad we are at stopping easy twos.

Assuming that reducing the number of three-attempts would enable our opponents to shoot a higher percentage on twos, then if we can clamp down near the rim I don't think we're significantly more susceptible to a hot shooting night from outside, even with the increased attempts. But no matter whether we limit three-attempts or not, if an opponent gets hot from three and we can't stop the two, we have a good chance of losing. Hopefully that doesn't happen too often, but with our current roster, I'm OK with attempting to reduce opposing two-point shooting percentage, even if it allows our opponents to attempt more threes.

[proelitedota]

We should be actively decreasing opposing 3PA. We can't even defensively rebound the ball as well on 3pt miss. I say we chase shooters off the lane and take chances in forcing opponents into 2pt jump shots and bad angle drives.

Maybe K doesn't want our bigs to get into foul trouble no matter what.

[Mtn.Devil.91.92.01.10.15]

We should be actively decreasing opposing 3PA.

Sorry, but with the size of our line up and talent of our bigs, teams are going to pop threes against us.

[jv001]

Sorry, but with the size of our line up and talent of our bigs, teams are going to pop threes against us.

Exactly what I told a friend at church last night. Teams seem to be reluctant to get the ball down low on a consistent basis. Wonder why? :cool: GoDuke!

[porkpa]

One of the major problems associated with landing "can't miss" prospects to come to play at a school like Duke is that very few of these kids have ever been been told or taught that defense is a priority.
The reason, in most cases, that they are so greatly desired is because of their offensive ability.
When was the last time you've heard that an incoming recruit is the best defensive player available? For me, that answer is never.
Its not because these kids don't have the physical ability to play defense. Its because its never been a priority and has likely never heretofore been emphasized to them.

[Saratoga2]

One of the major problems associated with landing "can't miss" prospects to come to play at a school like Duke is that very few of these kids have ever been been told or taught that defense is a priority.
The reason, in most cases, that they are so greatly desired is because of their offensive ability.
When was the last time you've heard that an incoming recruit is the best defensive player available? For me, that answer is never.
Its not because these kids don't have the physical ability to play defense. Its because its never been a priority and has likely never heretofore been emphasized to them.

We have a Sr player and captain on the floor all game. He is athletic and plays hard. We have Trent who is a solid defensive player who will only get better. Javin is a very athletic and mobile player who needs to avoid fouling as much but has exceptional defensive potential. Trevon is decent sized and very athletic and we have two excellent and large players on the back like.

I believe communication is what is lacking and what is hard to get going without sufficient practice time. The coaching staff needs to find the most appropriate defense based upon the great talent that we have. Lets hope that the practice time we get before the first of the year will sharpen the defense as another BC event could occur playing against quality ACC opponents.

[jv001]

We have a Sr player and captain on the floor all game. He is athletic and plays hard. We have Trent who is a solid defensive player who will only get better. Javin is a very athletic and mobile player who needs to avoid fouling as much but has exceptional defensive potential. Trevon is decent sized and very athletic and we have two excellent and large players on the back like.

I believe communication is what is lacking and what is hard to get going without sufficient practice time. The coaching staff needs to find the most appropriate defense based upon the great talent that we have. Lets hope that the practice time we get before the first of the year will sharpen the defense as another BC event could occur playing against quality ACC opponents.

I wonder if the hightly rated freshmen have bought into defense like they have offense. For sure they got their high ranking because of their offense. I don't think most high school coaches require the super stars to play good defense in order to get playing time. Practice time does matter at Duke and it's not Coach K blowing smoke when he says Duke has not had the practice time these young men need to get better. I'm hoping we see a vast improvement on defense by the first of January. GoDuke!

[English]

I wonder if the hightly rated freshmen have bought into defense like they have offense. For sure they got their high ranking because of their offense. I don't think most high school coaches require the super stars to play good defense in order to get playing time. Practice time does matter at Duke and it's not Coach K blowing smoke when he says Duke has not had the practice time these young men need to get better. I'm hoping we see a vast improvement on defense by the first of January. GoDuke!

Your point completely stands, but let’s be clear, HS basketball means nearly nothing in terms of recruitment rankings and prep. AAU, and traveling teams in other high level sports, is king. HS hoops may mean an eyebrow raise when talking about basketball factories like Findlay Prep and Monteverde (not coincidentally, where our guy Barrett matriculates), but the guys make their hay in AAU.

That said, AAU is woefully inadequate preparing these young men for college team defensive principles. Soooo, nit-picky way of agreeing with you

[Kedsy]

Duke defensive stats vs. E-ville:

2-point%: .278
3-point%: .357
%threes: 28.0%
eFG%: 35.0%

TO%: 31.4%
DR%: 79.4%
FT Rate: 12.0%
dRating: 0.60

A/to: 0.52:1
Asst Rate: 73.3%
Block %: 18.0% (25.0% of twos)

Efficiency difference: 25.8 (opposing three efficiency much better than opposing two efficiency, not that it mattered too much).


Well.

[Kedsy]

According to Pomeroy, that dominating defensive performance moved us from #76 before the game to #68 after. Funny how a bad defensive performance can move you down (e.g., BC game moved us from #53 to #71) much further than a great one can move you up.

[uh_no]

According to Pomeroy, that dominating defensive performance moved us from #76 before the game to #68 after. Funny how a bad defensive performance can move you down (e.g., BC game moved us from #53 to #71) much further than a great one can move you up.

not really. evansville had a terrible offense vs duke's okay defense. that means that assuming the result was a good defensive performance by duke the result will have a low weight. boston college's offense was closely matched to duke's defense, relatively....so the game would have much higher weight.

[Kedsy]

Duke defensive stats vs. FSU:

2-point%: .432
3-point%: .469
%threes: 46.4%
eFG%: 55.8%

TO%: 13.0%
DR%: 76.9%
FT Rate: 31.9%
dRating: 1.14

A/to: 2:1
Asst Rate: 71.0%
Block 7.2%: (13.5%of twos)

Efficiency difference: 26.4 (opposing three efficiency much better than opposing two efficiency).

OK, I'm officially worried about our 3-point defense. The combination of letting our opponent shoot 46% of their shots from three, and successfully making 47% of those threes, while also not turning them over and letting them hit 43% of their two-point shots adds up to bad defense. I admit it didn't look so bad, but it was. I'm not sure what we can do about it.

[curtis325]

Duke defensive stats vs. FSU:

2-point%: .432
3-point%: .469
%threes: 46.4%
eFG%: 55.8%

TO%: 13.0%
DR%: 76.9%
FT Rate: 31.9%
dRating: 1.14

A/to: 2:1
Asst Rate: 71.0%
Block 7.2%: (13.5%of twos)

Efficiency difference: 26.4 (opposing three efficiency much better than opposing two efficiency).

OK, I'm officially worried about our 3-point defense. The combination of letting our opponent shoot 46% of their shots from three, and successfully making 47% of those threes, while also not turning them over and letting them hit 43% of their two-point shots adds up to bad defense. I admit it didn't look so bad, but it was. I'm not sure what we can do about it.

There may be nothing Duke can do about it. Except maybe score 100 points. Yeah, that's the ticket!

[superdave]

OK, I'm officially worried about our 3-point defense. The combination of letting our opponent shoot 46% of their shots from three, and successfully making 47% of those threes, while also not turning them over and letting them hit 43% of their two-point shots adds up to bad defense. I admit it didn't look so bad, but it was. I'm not sure what we can do about it.

Coach K went into spitball defense to start the game - full court press the first couple of minutes, then some mix of zone and man. It slowed FSU down (they always seems to thrive in a helter skelter game) until their 3's started falling. He definitely game-planned this defensive scheme and I would not say it worked well, but provides for some good insight into the team. It seems to me that if they want to run, Duke is going to have to play a lot more man defense. But if they decide to slow the game down some, zone and press will be used more. This team should be running with finishers at all 5 positions.

Within the halfcourt defense, we are switching a lot of screens. We just arent rotating when the offense makes multiple quick passes. That to me is where the improvement needs to come - backside rotations, closing out on shooters. If Duval can continue to be a big disruptor, that helps everything. But Allen and Trent are going to need to avoid losing their guy on the weak side. O'Connell can maybe help here too. Use him as an energy guy when the defense seems lackluster. That's the role Allen played in 2015.

I have to say I am surprised Javin didnt come in a play better yesterday. Seemed like his kind of game but he was lost. We are going to need him and O'Connell to contribute, particularly on defense, because we are just going to be flat at times.

[Kedsy]

I'm copying this from the FSU game thread, because the subject matter matches the discussion we've been having here:


Lost in all of the worrying about our three point defense is the fact that we actually defended the two point shots really well, holding them to 16-37 inside the arc. And I don't have the numbers in front of me but I suspect our two point defense on the season is pretty decent. We've had a number of seasons where our fans complained about constantly overplaying the three point line, defending the three really well while giving up easy buckets inside. Dozens of threads have been started in recent years begging Coach K to play a pack-line defense, shutting down the interior at the expense of giving up open threes. Which is essentially what we're now doing (not by playing pack-line, but by overplaying and helping on the interior, leaving the perimeter open).

In an ideal world we would perfectly defend the interior and also perfectly defend the three. But such a system doesn't really exist, at least not when you're starting four freshmen and having them play the bulk of the minutes. I think our defense right now fits our personnel, the guys do seem committed to playing D and they have the physical tools so I think we will continue to improve. And as Mr. Sumner so eloquently put it, there were plenty of positive intangibles to take away from this game.

Our two-point percentage for the season is 44.5%. That's good but I'm not sure it's good enough to make up for all the threes.

Here are Duke's top ten opposing two-point percentages since the three-pointer was introduced:

2005: 41.3%
1999: 42.2%
2011: 43.2%
2004: 43.5%
2000: 43.7%
2001: 43.9%
2010: 44.1%
1989: 44.6%
1998: 44.7%
1997: 44.7%

So, it's a solid top ten performance. Unfortunately, our opponents' three-point percentage, 36.0%, is the third-highest Duke has allowed since the three-pointer was legal, behind only 1992 (37.7%) and 1995 (39.0%). It's worth noting the the high percentages in those two seasons were somewhat ameliorated by the fact that our 1995 opponents only took threes 24.2% of the time and our 1992 opponents only 20.4% of the time.

This year's team's opponents take 38.1% of their shots from the three-o-verse. That's the most ever allowed by a Duke team, by far -- the second highest such season percentage by Duke's opponents is 28.8% (which happened last season). This is the real issue, along with our anemic turnover percentage of 16.3%, the lowest turnover percentage since they started keeping that stat in 1987.

Seems to me, if we're going to let opponents take three-pointers at a historically high rate, and they're making them at a very high clip, and we're not forcing many turnovers, we need to stop two-pointers a lot better than we have so far, if we want to have a really good defense.

[Kedsy]

Duke defensive stats vs. NCSU:

2-point%: .547
3-point%: .357
%threes: 20.9%
eFG%: 54.5%

TO%: 13.4%
DR%: 58.8%
FT Rate: 37.3%
dRating: 1.28

A/to: 1.7:1
Asst Rate: 50.0%
Block 7.4%: (9.4%of twos)

Efficiency difference: 1.2 (opposing two efficiency a bit better than opposing three efficiency).

Well, the good news was our opponent didn't take so many threes this game. The bad news was everything else. Even our defensive rebounding was off.

Worst defensive performance of the season, and that's saying something.

[curtis325]

Need to change the title of the thread to "Tracking uke's efense"

[wavedukefan70s]

Maybe Zefense.

[Kedsy]

Duke defensive stats vs. Pitt:

2-point%: .448
3-point%: .182
%threes: 43.1%
eFG%: 37.3%

TO%: 23.6%
DR%: 63.9%
FT Rate: 43.1%
dRating: 0.82

A/to: 0.53:1
Asst Rate: 47.0%
Block 9.8%: (17.2%of twos)

Efficiency difference: 17.5 (opposing two efficiency a lot better than opposing three efficiency).

Our defensive rebounding was subpar, and their free throw rate was way too high. But they couldn't hit a three and we forced a good amount of turnovers, so overall our D was strong.

Unfortunately, we still gave up a huge number of three-point attempts. If they'd hit their season average on threes and we'd turned them over at our season average, this game would have looked like another poor defensive performance.

That said, the end of our bench played 54 minutes tonight (non-starters played 70 min). On the other hand, I'll be surprised if Pitt wins an ACC game this season. I'm not sure we can reasonably take anything away from this game.

[Wander]

If.... we'd turned them over at our season average, this game would have looked like another poor defensive performance.


So, if our defense was worse, then our defense would have been worse?

That said, I'm not sure the aggressive going for steals thing we did in the first 10 minutes wouldn't have backfired against a decent team, so I agree with you there's not much to take from the Pitt game (other than some positive thoughts on post depth).

[Kedsy]

So, if our defense was worse, then our defense would have been worse?

Depends on whether you attribute Pitt's poor three-point shooting to our defense. Not sure we did anything special to make them miss, and you can't really count on a 34% shooting team to shoot 18% on 22 attempts. And, like you say, I'm also not sure the increased forcing of turnovers is sustainable. I was just saying our defense wasn't nearly as good as the 0.82 suggests. Though I suppose it's the other side of the BC/FSU coin where our D wasn't as bad as it looked due to those teams shooting threes way better than they should have.

[CDu]

Depends on whether you attribute Pitt's poor three-point shooting to our defense. Not sure we did anything special to make them miss, and you can't really count on a 34% shooting team to shoot 18% on 22 attempts. And, like you say, I'm also not sure the increased forcing of turnovers is sustainable. I was just saying our defense wasn't nearly as good as the 0.82 suggests. Though I suppose it's the other side of the BC/FSU coin where our D wasn't as bad as it looked due to those teams shooting threes way better than they should have.

To be fair, I don't think FSU shot threes better than they should have. I did a breakdown of their 32 attempts, and they reasonably broke down to be as you'd expect a team of good shooters to shoot. I haven't done the same with BC though.

[Kedsy]

Duke defensive stats vs. Wake:

2-point%: .327
3-point%: .435
%threes: 30.7%
eFG%: 42.7%

TO%: 15.1%
DR%: 75.0%
FT Rate: 12.0%
dRating: 0.90

A/to: 1.25:1
Asst Rate: 55.6%
Block 6.7%: (9.6% of twos)

Efficiency difference: -32.6 (opposing three efficiency a LOT better than opposing two efficiency).

Even though our dRtg today was a little worse than it was against Pitt, this game made me more comfortable with our defensive improvement. For one thing, Wake's offense (#52 in the nation according to KenPom) is WAY better than Pitt's (#234). For another, our defensive performance in this game seems more sustainable. Despite it seeming (at least to me) that Wake got every offensive rebound, in actual fact Duke grabbed 75% of available defensive boards, which is pretty good, especially in comparison to the 64% we grabbed against Pitt. And our opposing free throw rate tonight (12%) was outstanding, again especially compared to the somewhat atrocious 43% we allowed against Pitt. Unlike Pitt (18% on threes), Wake shot really well from three (43.5%), but we made up for that by allowing fewer three attempts (only 31%) and a lower two-point shooting percentage (33%). Also, we got a lot lower turnover percentage tonight (15%) than we did against Pitt (24%), but since that's more in line with our season performance to date, to me that makes today's above average defensive performance less fluky than our D against Pitt.

Baby steps, I guess. If we allow 0.90 ppp for the rest of the season, I'll be very happy. Although, duh, since an adjusted 0.90 rate would be 6th in the nation according to KenPom. For what it's worth, Wake's offense is 52nd in the nation, but 9th in the ACC.

[JasonEvans]

Let me start by saying that Kedsy’s posts here are truly great and I appreciate his efforts.

Secondly, I wonder what the defensive experts around here have to say about the way Capel ran the team yesterday versus K’s tendencies. I am not able to assess some stat on this, but it sure seemed that we played a lot lot lot more zone yesterday than we usually do. I wonder if that may have helped our efficiency. Does anyone have zone vs man defensive stat breakdowns?

Jason “I think that when you rebound as well as we do, zone is ok because you don’t need to force turnovers quite as much” Evans

[uh_no]

Let me start by saying that Kedsy’s posts here are truly great and I appreciate his efforts.

Secondly, I wonder what the defensive experts around here have to say about the way Capel ran the team yesterday versus K’s tendencies. I am not able to assess some stat on this, but it sure seemed that we played a lot lot lot more zone yesterday than we usually do. I wonder if that may have helped our efficiency. Does anyone have zone vs man defensive stat breakdowns?

Jason “I think that when you rebound as well as we do, zone is ok because you don’t need to force turnovers quite as much” Evans

I'm not sure I saw that. I did, however, see more fighting thru screens instead of switching between guards/big men....something that I think is at the crux of a lot of our defensive issues the past several years, as it both complicates things and puts inexperienced defenders in big mismatches.

[CDu]

I'm not sure I saw that. I did, however, see more fighting thru screens instead of switching between guards/big men...something that I think is at the crux of a lot of our defensive issues the past several years, as it both complicates things and puts inexperienced defenders in big mismatches.

No, we did play a ton of zone yesterday. Possibly the entire second half, and a fair amount in the first. We did start the game in man though.

[Rich]

Let me start by saying that Kedsy’s posts here are truly great and I appreciate his efforts.

Secondly, I wonder what the defensive experts around here have to say about the way Capel ran the team yesterday versus K’s tendencies. I am not able to assess some stat on this, but it sure seemed that we played a lot lot lot more zone yesterday than we usually do. I wonder if that may have helped our efficiency. Does anyone have zone vs man defensive stat breakdowns?

Jason “I think that when you rebound as well as we do, zone is ok because you don’t need to force turnovers quite as much” Evans


No, we did play a ton of zone yesterday. Possibly the entire second half, and a fair amount in the first. We did start the game in man though.

I echo Jason's comments - I'm learning a lot from this thread. I usually just focus on the man with the ball but am trying more to watch the whole floor so I can see how we're defending. I was proud of myself for recognizing on the first play of the game there was a high PNR and Carter went into drop coverage. WF missed a mid-range jumper, if I recall correctly. I'm pretty sure we were in zone the entire second half. I thought at some points it looked more like a 1-3-1 (or a 3-2) rather than the standard 2-3, but it could just have been that we were playing the 2-3 more aggressively than I've seen in the past.

[JasonEvans]

I echo Jason's comments - I'm learning a lot from this thread. I usually just focus on the man with the ball but am trying more to watch the whole floor so I can see how we're defending. I was proud of myself for recognizing on the first play of the game there was a high PNR and Carter went into drop coverage. WF missed a mid-range jumper, if I recall correctly. I'm pretty sure we were in zone the entire second half. I thought at some points it looked more like a 1-3-1 (or a 3-2) rather than the standard 2-3, but it could just have been that we were playing the 2-3 more aggressively than I've seen in the past.

Wake was playing 4 guys on the perimeter a good bit so we were pulling a man out from underneath which made the 2-3 look different than you might think.

[CDu]

Yep 2-3 zone. But when nobody was in the corner, the forward would shift up. This gives the appearance of a 3-2.

[stedge]

We started m2m but went zone less than halfway through first half. At that point we pulled away.

Was a tie game w m2m

[Stray Gator]

We started m2m but went zone less than halfway through first half. At that point we pulled away.

Was a tie game w m2m

Actually, I believe that except for switching from man-to-man for one possession shortly before the under-12 TV timeout in the first half, Duke didn't go to the zone exclusively until right after Crawford drove the lane for that dunk at the 5:48 mark of the first half. My son and I noticed that after adding a free throw on their next possession to get to 28, Wake was stuck there, unable to score for 5 minutes after Duke went to the zone. During that span, Duke went on an 11-0 run to build a double-digit cushion that it never surrendered.

[DukieInBrasil]

We started m2m but went zone less than halfway through first half. At that point we pulled away.

Was a tie game w m2m

yeah, i don't see what the point of playing m2m is anymore, unless it's just to change things up for a bit if the zone starts leaking points. The zone got us the win vs MSU, now whether that was surprise or execution is up for debate, but the zone generated turnovers and fast break points, something our m2m hasn't done this year.
One of the dangers of starting out in m2m is that the other team can get hot and build confidence. The m2m D leaves too many open shooters, too many driving lanes and not enough turnovers to be worthwhile. Sometimes teams miss those open shots and don't get hot, but when they do it can be nearly impossible to get them to cool down. Witness BC and to a lesser extent NC State, and if not for a choke factor that FSU seems to apply to itself in nearly every game, them too.
When Duke does go to m2m, using the drop by Carter (usually) has been much more effective than what was happening before, so that's a positive. The major benefit of that change is the opponent taking less efficient shots, which is good. The major benefit of the zone appears to be that we get more live ball turnovers, which is great, and allows our offensive efficiency to soar.
I don't think this team needs to even become an elite defensive team, though that would be welcome, just a slightly better than average defense would be fine, because the offense is so good. Allowing the guys to use their length to disrupt passing lanes both disrupts the flow of the opponents' offense which might lower their overall efficiency and gives us live ball turnovers, which is pure gold for this team. For that reason, it seems that zone defense is a better fit for this team.

[WakeDevil]

It's effectiveness, not efficiency, no matter what the NFL, Pomeroy and some of you say.

[Kedsy]

It's effectiveness, not efficiency, no matter what the NFL, Pomeroy and some of you say.

What's the difference?

[Kedsy]

Let me start by saying that Kedsy’s posts here are truly great and I appreciate his efforts.

Secondly, I wonder what the defensive experts around here have to say about the way Capel ran the team yesterday versus K’s tendencies. I am not able to assess some stat on this, but it sure seemed that we played a lot lot lot more zone yesterday than we usually do. I wonder if that may have helped our efficiency. Does anyone have zone vs man defensive stat breakdowns?

Jason “I think that when you rebound as well as we do, zone is ok because you don’t need to force turnovers quite as much” Evans

I just spent the last hour attempting to find this information. Sadly, I could find nothing definitive.

This 247 article (https://247sports.com/college/duke/Bolt/ESPN-analyst-discusses-Dukes-defensive-issues-113354232) from January 7 (i.e., doesn't include stats from Pitt or Wake games) says Duke is playing zone on 28% of opposing halfcourt possessions and that Duke's opposing points per possession in zone this season is 0.856 while Duke's opposing ppp in man was 0.876.

Since Duke's overall opposing points per possession on January 7 was 0.993, I'm not sure how accurate the above numbers are.

Trying to parse it out, according to Hoop Math (http://hoop-math.com/Duke2018.php), 20.8% of Duke's opposing possessions are in transition, and in those transition possessions, opponents have a .545 eFG% on the initial transition shot attempt. That translates to 1.09 ppp on the initial transition shot attempt, but how many additional points are attributed to transition (e.g., free throws, quick rebound putbacks, etc.)?

In order for 247's numbers to be correct, opponents in our first 15 games must have scored 87 points in our opponents' 237 transition opportunities, in addition to the 258 points scored on the initial transition shot attempt. Is that possible? Sure. Is it likely? I have no idea, but if those additional points didn't happen, I'm not sure how to reconcile it with 247's numbers.

[CDu]

What's the difference?

The argument is that efficiency is the ease (or lack of resources needed) to produce a thing, whereas effectiveness is the probability of success one generates from a measure of attempts.

Practically speaking, though, I agree. They get at the same concept. We are very effective in scoring per possession. Therefore, we are very efficiently able to reach a high point total.

[Kedsy]

While I was researching the above answer, it occurred to me that if transition scoring is so much more efficient than halfcourt scoring, it's possible we're not entirely focusing on the right things when we evaluate Duke's D. I went back to Hoop-Math and assembled the following table with some Duke transition defense information:



Year %transition trans eFG% trans %3att trans 3pt% Pom D Rank
2018 20.8% 54.5% 39.2% 34.6% 86
2017 23.8% 58.4% 28.0% 32.3% 47
2016 24.4% 55.7% 30.6% 30.2% 86
2015 23.5% 52.2% 30.6% 25.9% 11
2014 22.1% 54.5% 30.7% 24.1% 86
2013 23.9% 51.7% 32.7% 33.3% 26
2012 22.1% 58.4% 27.3% 33.3% 79


I'm not entirely sure what to make of this data, but I have a few semi-random observations:

- note that the eFG% is on initial transition shot attempt. I couldn't find any data on transition putbacks or transition free throws.

- Our two best defensive teams in the period had the two best opposing transition eFG%. That said, if this year's team let up an initial transition shot attempt eFG% of 52.2% like the 2015 team did, it would only be a 0.73 points per game improvement in opposing scoring (or about 1 point per 100 possessions). Which isn't close to enough to explain the difference in the 2013 or 2015 defenses and this year's defense. Though a one point improvement in opposing scoring per 100 possessions would move our current KenPom defensive rank from #86 to #67.

- this year's team is giving up by far the highest percentage of opposing transition three-point attempts, and also the highest transition three-point shooting percentage. This is in transition, though, not in the halfcourt. Are there not enough guys getting back in transition D? Are they not getting back fast enough? Are they running to the wrong spot? Or is it just bad luck which could even out over time?

- this year's team gives up the lowest percentage of transition opportunities of any team in the period. Possibly because we're so good at offensive rebounding. Whatever the reason, since transition scoring is so much more efficient than halfcourt scoring, this can only be a good thing, right? But does it mean that we have more potential to end up with a good defense because we can limit transition opportunities? Or, since our overall dRtg is so mediocre, does the fact that we're already limiting transition opportunities mean our halfcourt defense is even worse than we think it is? Or both?

I don't really know the answers to any of these questions, but I thought it was interesting to think about.

[devildeac]

I echo Jason's comments - I'm learning a lot from this thread. I usually just focus on the man with the ball but am trying more to watch the whole floor so I can see how we're defending. I was proud of myself for recognizing on the first play of the game there was a high PNR and Carter went into drop coverage. WF missed a mid-range jumper, if I recall correctly. I'm pretty sure we were in zone the entire second half. I thought at some points it looked more like a 1-3-1 (or a 3-2) rather than the standard 2-3, but it could just have been that we were playing the 2-3 more aggressively than I've seen in the past.

I noted something similar to what you observed, too. At times, it *almost* looked like a 2-2-1 zone. Syracuse did something very similar yesterday vs FSU, bringing the "wings" on the baseline up closer to the "top" of the zone. I'm guessing it's to stop penetration and/or open looks from the seam in the defense. Not that Jimmy B has ever coached anything like that before ;):rolleyes:.

[Stray Gator]

On a few possessions, it appeared to me that Duke morphed into an active zone scheme that looked suspiciously like Tark's old "amoeba" defense -- for those who remember back that far.

[CDu]

I noted something similar to what you observed, too. At times, it *almost* looked like a 2-2-1 zone. Syracuse did something very similar yesterday vs FSU, bringing the "wings" on the baseline up closer to the "top" of the zone. I'm guessing it's to stop penetration and/or open looks from the seam in the defense. Not that Jimmy B has ever coached anything like that before ;):rolleyes:.

The team was playing a 2-3. However, they appear to have instructed the baseline forwards in the 2-3 (i.e., the two guys on either side of the center in the "3" of the 2-3) to cheat up if there is nobody lurking near the corner 3 area. This effectively gives a 3-2 look in a lot of cases, and takes pressure off the rotations for the 2 guards up front. If a ball swings across the top of the key, the baseline forward picks up at the 3pt line until the guard recovers over from the middle of the court. Once the guard gets there, the forward drops back to the natural forward spot. At the same time, most likely the OTHER forward (formerly on the ball side, now on the weak side) will likely cheat up as the former strongside/now weakside guard slides to the middle of the court.

So it will look a lot like a 3-2 in a lot of cases, but is a 2-3 with aggressive rotations from the weakside forwards.

And yes, we played roughly the last 25+ minutes in the 2-3 zone. The last possession of man-to-man was the possession in which Crawford drove from the scorer's table to the rim for an uncontested dunk. That made the score 31-27 with under 6 minutes to go. We would then go on a 57-36 stretch over the next 22 minutes of play in the 2-3 zone.

We certainly weren't terrible in man-to-man. But until the final minutes, we were a bit better in zone. The biggest difference though was our offense, which really took off late in the first half and through the majority of the second half.

[TruBlu]

I just spent the last hour attempting to find this information. Sadly, I could find nothing definitive.

This 247 article (https://247sports.com/college/duke/Bolt/ESPN-analyst-discusses-Dukes-defensive-issues-113354232) from January 7 (i.e., doesn't include stats from Pitt or Wake games) says Duke is playing zone on 28% of opposing halfcourt possessions and that Duke's opposing points per possession in zone this season is 0.856 while Duke's opposing ppp in man was 0.876.

Since Duke's overall opposing points per possession on January 7 was 0.993, I'm not sure how accurate the above numbers are.

Trying to parse it out, according to Hoop Math (http://hoop-math.com/Duke2018.php), 20.8% of Duke's opposing possessions are in transition, and in those transition possessions, opponents have a .545 eFG% on the initial transition shot attempt. That translates to 1.09 ppp on the initial transition shot attempt, but how many additional points are attributed to transition (e.g., free throws, quick rebound putbacks, etc.)?

In order for 247's numbers to be correct, opponents in our first 15 games must have scored 87 points in our opponents' 237 transition opportunities, in addition to the 258 points scored on the initial transition shot attempt. Is that possible? Sure. Is it likely? I have no idea, but if those additional points didn't happen, I'm not sure how to reconcile it with 247's numbers.

Sadly, one possible reason for the disparity may be this: For quite a few of the opponents' possessions, they (like some of us) couldn't tell what the heck Duke was doing on defense.

[Kedsy]

Duke defensive stats vs. Miami:

2-point%: .377
3-point%: .360
%threes: 32.1%
eFG%: 42.9%

TO%: 13.6%
DR%: 73.5%
FT Rate: 12.8%
dRating: 0.93

A/to: 0.91:1
Asst Rate: 34.5%
Block 6.4%: (9.4% of twos)

Efficiency difference: -16.3 (opposing three efficiency a LOT better than opposing two efficiency).

Tonight's defensive performance was almost exactly like our performance against Wake. To me, that's a good thing, because it means whatever we're doing might be reproducible. Or it might be coincidence, but I'm going with reproducible. Again it seemed like Miami got every offensive rebound, but we snagged 73.5% of available defensive boards, which is pretty good. Again, our opposing free throw rate (12.8%) was outstanding. Again, we allowed fewer three attempts (only 32%) and a lower two-point shooting percentage (38%). Alas, we also once again didn't turn them over very well (13.6%), but I think we kind of have to get used to that.

Only real difference is Miami (#107 per Pomeroy) doesn't have as good an offense as Wake, but personally, I'll take it.

I'd also note that Miami had 31 fast break points tonight, which was 41% of their points, and probably makes the overall defensive performance look worse than it was.

[Kedsy]

While writing my previous post, I realized the official box score has totals for "fast break points."

Since transition offense generally has a much higher points per possession, I wondered if the number of fast break points we have given up has skewed our view of our defense, so I created the following table to show how many fast break points we've given up in each game.



Opponent Points Fast Break Pts % ppp
Miami 75 31 41.3% 0.93
FSU 93 19 20.4% 1.13
Mich St 81 16 19.8% 1.04
Wake 71 13 18.3% 0.89
Port St 81 14 17.3% 1.03
Florida 84 12 14.3% 1.05
Pitt 52 6 11.5% 0.81
Indiana 81 8 9.9% 1.14
UVU 69 6 8.7% 0.86
Southern 61 5 8.2% 0.86
Texas 78 6 7.7% 0.95
Eville 40 3 7.5% 0.58
NC State 96 7 7.3% 1.23
S Dakota 80 4 5.0% 1.01
BC 89 0 0.0% 1.17
Elon 68 0 0.0% 0.91
St. Fran 67 0 0.0% 0.80
Furman 63 0 0.0% 0.94


Observations and possible conclusions:

- Miami scored over 40% of their points on fast breaks? OMG!

- Since lots of fast breaks have nothing to do with halfcourt defense, our halfcourt D wasn't as bad as it seemed against Michigan State, Florida, and Portland State. Same with FSU, though it was still bad in that one.

- On the other side of the coin, our halfcourt defense was even worse than it appeared in the NC State and BC games, and it already looked really bad in those games.

- For the same reason, our halfcourt defense in our last three games has been even better than it looked (and it has looked pretty good). Maybe our halfcourt D really is turning a corner?

- Four of our highest seven opponents' fast break point performances (by percentage) have come in our last five games -- what's up with that? Looks like a clear area for potential improvement.

- I'm really glad we don't have any more games this year in the state of Florida.

[-jk]

And we moved from 88 to 72 in KenPom AdjD! (OK, while moving from allowing 99.9 to 99.1 - it's really bunched up...)

-jk

[English]

And we moved from 88 to 72 in KenPom AdjD! (OK, while moving from allowing 99.9 to 99.1 - it's really bunched up...)

-jk

Something I've often wondered reading posts about this season's team, in complete sincerity: folks (rightly, IMO) posit that this Duke team doesn't need to play great defense to have tremendous success, because our offense is so good. Rather, the team only needs to play average-to-slightly-above-average defense.

Well, what does an average or above average defense look like, per KP? Is that like top-50? Top-35? I imagine it's a moving scale as more data come into the model, but still, a snapshot at this point in the season would help me wrap my mind around what I should hope for (analytically speaking, of course; I root for stops on every defensive possession and tend to notice when our defense is a disaster).

[Ian]

Something I've often wondered reading posts about this season's team, in complete sincerity: folks (rightly, IMO) posit that this Duke team doesn't need to play great defense to have tremendous success, because our offense is so good. Rather, the team only needs to play average-to-slightly-above-average defense.

Well, what does an average or above average defense look like, per KP? Is that like top-50? Top-35? I imagine it's a moving scale as more data come into the model, but still, a snapshot at this point in the season would help me wrap my mind around what I should hope for (analytically speaking, of course; I root for stops on every defensive possession and tend to notice when our defense is a disaster).

By average generally people mean the average NCAA tournament team, which would be around 30.

[CDu]

By average generally people mean the average NCAA tournament team, which would be around 30.

I don't think there is a hard and fast rule here, as it also depends on just how good the offense is. Ultimately, what you want is a team that has a high margin. Whether that is more defense than offense probably matters less than the actual difference.

Last year, the average tourney team seeded 11 or higher (i.e., the average at large quality team) had an adjusted defensive rating of 95.47, or a defensive ranking of around 43-44. The median at-large quality tourney team had a 95.5 rating, and was 31-32 in defense.

Note that these are based on post-tourney numbers, which obviously are skewed some by the tournament. But since they include most of the teams, it probably doesn't skew too wildly. I'd guess that it skews the average slightly downward though, as the teams that played better played more games.

Still, somewhere in the 95-96 range (which would put them in the 30-35 range) nationally would feel pretty good as it relates to this team's chances of winning it all.

Though again, there's really no hard and fast rule here. The tourney is largely about matchups. People talk about how the 2015 team turned it on in the tournament. Well, it's worth noting that they faced an almost perfect stretch of opponents for their strengths in the later rounds of that tournament. Gonzaga, Utah, MSU, and Wisconsin all played slow paces and lacked elite athleticism at the PG spot. 3 of those 4 also lacked sufficient skill at the center spot to take advantage of Okafor's limited mobility and defensive instincts (Wisconsin clearly had that, and got Okafor in foul trouble; Utah had Poetl, but he was less of a polished offensive player away from the basket; MSU and Gonzaga played slow and/or low-caliber centers). If we'd faced a team like Notre Dame, who knows if we win it? Similar story with Kentucky. But we faced the right matchups, and plowed through to the title.

With this team, I think our concerns are teams that can spread us out, drive, and also hit 3s. If we avoid those teams in the tournament, I like our chances. If we don't, well, we'll need to play defense more like the last few games than we have for much of the season.

And to be fair, I definitely think we can win the title if we play defense at the level we've played the last 3 games. I wasn't as sure after the Wake game just because it was a home game and a one-game sample. But we played very well defensively last night too, and on the road. That bodes well. Had our offense not struggled for 3/4 of the game, we'd have won by a landslide.

There is still a lot of "prove it" left to do. But I really like the direction the team is going right now.

[Troublemaker]

Something I've often wondered reading posts about this season's team, in complete sincerity: folks (rightly, IMO) posit that this Duke team doesn't need to play great defense to have tremendous success, because our offense is so good. Rather, the team only needs to play average-to-slightly-above-average defense.

I actually don't agree with that supposition, personally. Most likely, if Duke ends up winning 6 NCAAT games to become national champs, we'll have done it by playing great defense over the course of those 6 games. (If I'm wrong and we win it all while playing only average defense, feel free for anyone who remembers to call me out in April, haha).

The other argument I've seen is that Duke can enter the NCAA tournament with merely a top-60 (or so) defense and go on to play 6 great defensive games to raise our ranking a la 2015. While I would normally agree with that, imo it's no longer relevant for this season's team because we "made our move" early, in mid-January. These past 3 games, we've already raised our ranking from 100+ to the 90s following Pitt, to the 80s following Wake, and now we sit at 72 following Miami. There are a LOT of games left in the regular season. Something would have to go horribly wrong for us to still only be a top-60 defense entering the tournament. Most likely, that something would involve either a major injury or Coach K abandoning zone defense and abandoning icing + drop in the m2m. On the other hand, if Duke stays healthy and Coach K continues to use zone and icing + drop in m2m, Duke should enter the NCAA tournament with a top-20 defense to go along with our top-3 offense.

With the defensive adjustments we've made, Duke is back as a serious contender for the national championship, the ACC regular season crown, and a 1 seed.

[House P]

Tonight's defensive performance was almost exactly like our performance against Wake. To me, that's a good thing, because it means whatever we're doing might be reproducible. Or it might be coincidence, but I'm going with reproducible.

First of all, a big “thanks” to Kedsy for all the excellent info in this thread.

As for whether Duke’s improved defense in the past couple games is reproducible, I don’t know. However, it definitely seems like things are trending in the right direction since the NC State game. The graph below estimates* Duke’s adjusted defensive efficiency (AdjD) for every game this year. This graph indicates that the last 3 games have been among Duke’s best defensive games this season.

Duke’s defense over the past 3 games (average AdjD = 86.6) is better than the season-long AdjD of every team in the county except Virginia (AdjD = 85.5).



http://forums.dukebasketballreport.com/forums/attachment.php?attachmentid=7992&stc=1

Warning: I would be cautious interpreting Duke’s best 3 game stretch this year as the “new normal”. Duke’s defensively-challenged 2014 team finished the season ranked #86 defensively, but still had a 3 game stretch between Jan 18 and 25 where they played like the #1 defense in the country. Over the next 3 games, the 2014 team put up an average AdjD of 105.4, which would have been on par with the #174 ranked defense of Nebraska-Omaha. The 2016 team had a similar 3 game stretch against ACC teams where they played like a top 10 defense, even though they ended the season ranked #86 by KenPom. Of course, Duke’s current three game trend of very good defense is certainly better than continuing to play like they did against NC State.



Finally, a trivia question. By my calculation, Duke’s single game adjusted defensive efficiency vs Evansville was 64.2. As best as I can tell, this was Duke’s 7th best defensive game (in terms of AdjD) since the 2001-02 season, the first year KenPom has stats available.

Sporks to the first person who can guess Duke’s lowest single game AdjD since the 2001-02 season. Hint: Duke’s opponent for that game ended the season in KenPom’s top 100.


*If anyone is interested in exactly how I estimated the adjusted efficiency for each game, I can dig up a previous post explaining the calculation.

[uh_no]

Sporks to the first person who can guess Duke’s lowest single game AdjD since the 2001-02 season. Hint: Duke’s opponent for that game ended the season in KenPom’s top 100.



all I got is 82-50 v UNC or 60-20 v md

[Kedsy]

*If anyone is interested in exactly how I estimated the adjusted efficiency for each game, I can dig up a previous post explaining the calculation.

Thanks for all this info, House P. I am interested in how you calculated the adjusted efficiency.

[sagegrouse]

By average generally people mean the average NCAA tournament team, which would be around 30.

Yeah, maybe, but I reference the post above, which pointed out that a minor improvement in efficiency had us jump 20 places. Efficiency ratings, I suppose, have their purposes, but ordinal measures (rankings) need to be watched carefully.

[sagegrouse]

Thanks for all this info, House P. I am interested in how you calculated the adjusted efficiency.

House P, this is Admiral Ackbar -- it's a trap!

[House P]

all I got is 82-50 v UNC or 60-20 v md

82-50 is always a good guess, but the 2009-10 UNC team was barely a top 100 offense per KenPom*. As such holding them to 50 points in 64 possessions was lots of fun to watch, but only ranks as Duke's 79th best defensive game since 2002.

Duke's adjusted offensive efficiency that game ranks as Duke's 53rd best offensive game. So, when the 79th best defensive game and 53rd best offensive game are combined, the result is a fantastic blowout, but not the answer to my trivia question.



*The exact ranking of UNC's offense per KenPom depending on where you get the numbers from his site. When looking up the specifics of the 2010 UNC game, I realized that there are differences between the ratings listed on KenPom's main page for each season and what you will get when you download the "data" spreadsheets from his site. I suspect that this has to do with KenPom tweaking his formula a couple years ago. I assume that the ratings listed in the "data" spreadsheets are his original ratings, while those listed on the front page use the updated method. For clarity, I used the results in KenPom "data" spreadsheet for my trivia question because this was easier to work with.

[Matches]

82-50 is always a good guess, but the 2009-10 UNC team was barely a top 100 offense per KenPom*. As such holding them to 50 points in 64 possessions was lots of fun to watch, but only ranks as Duke's 79th best defensive game since 2002.

Duke's adjusted offensive efficiency that game ranks as Duke's 53rd best offensive game. So, when the 79th best defensive game and 53rd best offensive game are combined, the result is a fantastic blowout, but not the answer to my trivia question.



I'll play. The 2013 Duke-UNC game in Chapel Hill?

[House P]

I'll play. The 2013 Duke-UNC game in Chapel Hill?

Another good guess, but not what I was looking for. By my calculations, the 2013 UNC game was Duke's 72nd best AdjD game since 2002. Unfortunately, the top 2 defensive games vs UNC since 2002 have been guessed, so the answer is not a Duke-UNC game.

[CDu]

Another good guess, but not what I was looking for. By my calculations, the 2013 UNC game was Duke's 72nd best AdjD game since 2002. Unfortunately, the top 2 defensive games vs UNC since 2002 have been guessed, so the answer is not a Duke-UNC game.

My initial thought was Duke/Texas early in the 2005-06 season. But then I realize that the big reason for that blowout was the offense. But that clued me in to another game from that season. Duke beat Seton Hall 83-40, on about 72 possessions. Either that or when we beat Maryland 85-44 in 2009, in a 73-possession game.

[Wander]

My initial thought was Duke/Texas early in the 2005-06 season. But then I realize that the big reason for that blowout was the offense. But that clued me in to another game from that season. Duke beat Seton Hall 83-40, on about 72 possessions.

I was trying to think of this game, but couldn't remember the opponent. Wasn't JJ Redick "winning" for most of the game?

How about the beatdown Duke's 2010 title team put on Gonzaga?

[tbyers11]

I was trying to think of this game, but couldn't remember the opponent. Wasn't JJ Redick "winning" for most of the game?

How about the beatdown Duke's 2010 title team put on Gonzaga?

I believe the Gonzaga game in Dec 2009 is the answer. But because I cheated and looked at KenPom I didn't want to spoil the answer.

Our raw Deff was 57.1. Against a Gonzaga team that averaged 112.3 over the 2010 season.

[House P]

I was trying to think of this game, but couldn't remember the opponent. Wasn't JJ Redick "winning" for most of the game?

How about the beatdown Duke's 2010 title team put on Gonzaga?

This is the game I was looking for. Congratulations!


I believe the Gonzaga game in Dec 2009 is the answer. But because I cheated and looked at KenPom I didn't want to spoil the answer.

Our raw Deff was 57.1. Against a Gonzaga team that averaged 112.3 over the 2010 season.

As you point out, Duke held a Gonzaga team which was solid offensively (#38 in the final rankings) to 41 points in 72 possessions. That is about half of the points they would be expected to score against an average D1 opponent. Not bad for a team which was "alarmingly unathletic"!

One other factor which helps make the Gonzaga game look better from an adjusted efficiency perspective is that it was held on a neutral court. Most of Duke's other dominate games took place at Cameron where the visiting team is expected to be less efficient than they would be otherwise.

[House P]

My initial thought was Duke/Texas early in the 2005-06 season. But then I realize that the big reason for that blowout was the offense. But that clued me in to another game from that season. Duke beat Seton Hall 83-40, on about 72 possessions. Either that or when we beat Maryland 85-44 in 2009, in a 73-possession game.

Excellent guesses. By my calculations, these are the 3rd and 5th best single game AdjD performances for Duke since 2002.

Here is the full Top 25 defensive performances for Duke since 2001-02.






Date
Year
Opponent
Result

Adj DE
Final Rank


Sat Dec 19
2009
Gonzaga
W, 76-41
N
54.0
45


Thu Mar 14
2002
Winthrop
W, 84-37
N
59.1
219


Wed Nov 16
2005
Seton Hall
W, 93-40
H
60.3
83


Thu Feb 7
2002
Florida St.
W, 80-49
H
61.0
122


Sat Jan 24
2009
Maryland
W, 85-44
H
62.7
55


Wed Dec 20
2017
Evansville
W, 104-40
H
64.0
149


Thu Jan 18
2007
Wake Forest
W, 62-40
H
64.1
101


Wed Feb 16
2011
Virginia
W, 56-41
A
64.5
103


Sun Dec 14
2003
Portland
W, 84-43
H
64.6
162


Sat Nov 27
2004
UNC Greensboro
W, 98-44
H
64.8
209


Sat Dec 10
2016
UNLV
W, 94-45
a
65.1
249


Fri Nov 11
2016
Marist
W, 94-49
H
65.6
309


Tue Dec 31
2013
Elon
W, 86-48
N
66.8
201


Sat Dec 1
2012
Delaware
W, 88-50
H
67.0
167


Sat Nov 20
2004
Tennessee Martin
W, 88-46
H
68.0
299


Sun Mar 24
2013
Creighton
W, 66-50
N
68.3
17


Sun Nov 12
2006
Columbia
W, 86-43
H
68.4
205


Wed Dec 19
2012
Cornell
W, 88-47
H
68.4
260


Sun Jan 4
2009
Virginia Tech
W, 69-44
H
68.6
73


Wed Jan 11
2006
Maryland
W, 76-52
H
68.6
55


Wed Dec 31
2008
Loyola MD
W, 92-51
H
68.8
222


Thu Feb 21
2002
Wake Forest
W, 90-61
A
68.9
35


Tue Jan 8
2013
Clemson
W, 68-40
H
68.9
110


Sun Nov 25
2007
Eastern Kentucky
W, 78-43
H
69.1
278


Sat Nov 25
2006
Davidson
W, 75-47
H
69.3
61





Note: The Final Rank column indicates the teams final overall KenPom rank for the season.

One thing that stands out to me is that most of these games are against fairly bad teams. I think most of us would agree that holding Doug McDermott's Creighton team to 50 points in the NCAA tournament is more impressive than holding Marist to 49 points in Cameron. Unfortunately, the formula doesn't work out that way. This is likely a big reason why KenPom tweaked his formula in 2015 to diminish the impact of blowouts vs overmatched teams. I assume that when KenPom calculates Duke's overall rating for 2018, the Saint Francis game this year has almost no influence.

[Kedsy]

Obliquely regarding the discussion about man vs. zone defense in the Miami game, I went back to the official box score and checked transition possessions. I counted as a "transition opportunity" any time Miami did something (shot attempt, free throw, turnover) within 8 seconds of a possession change (e.g, Duke missed shot, made shot, turnover; but I didn't count it if there was a stoppage of play in the middle). This is different from "fast break" points, but I figured 8 seconds probably isn't enough time to get your defense set, man or zone.

Here are the results:

transition possessions: 25
transition points: 41
transition ppp: 1.64

non-transition possessions: 54
non-transition points: 34
non-transition ppp: 0.63

I don't know which non-transition possessions were man-to-man and which were zone, but here's the breakdown by half:

FIRST HALF
transition: 24 points in 13 possessions (1.85 ppp);
non-transition: 16 points in 28 possessions (0.57 ppp);

SECOND HALF
transition: 17 points in 12 possessions (1.42 ppp);
non-transition: 18 points in 26 possessions (0.69 ppp);

So (assuming first half was mostly man and second half was mostly zone) I conclude that our halfcourt defense was pretty solid the whole game. Despite how it looked, the difference in the game wasn't really zone vs. man (and if it was, our D was slightly better in the first half). It was really that we finally cut down on Miami transition opportunities.

We allowed zero transition opportunities after the 8-minute mark of the second half, until there was just 41 seconds to play. The score after Miami's transition basket with 8:00 to play was 66-53 Miami. The score when they next had a transition opportunity (at 0:41) was 78-70 Duke. And that was the ballgame.

[Ian]

Obliquely regarding the discussion about man vs. zone defense in the Miami game, I went back to the official box score and checked transition possessions. I counted as a "transition opportunity" any time Miami did something (shot attempt, free throw, turnover) within 8 seconds of a possession change (e.g, Duke missed shot, made shot, turnover; but I didn't count it if there was a stoppage of play in the middle). This is different from "fast break" points, but I figured 8 seconds probably isn't enough time to get your defense set, man or zone.

Here are the results:

transition possessions: 25
transition points: 41
transition ppp: 1.64

non-transition possessions: 54
non-transition points: 34
non-transition ppp: 0.63

I don't know which non-transition possessions were man-to-man and which were zone, but here's the breakdown by half:

FIRST HALF
transition: 24 points in 13 possessions (1.85 ppp);
non-transition: 16 points in 28 possessions (0.57 ppp);

SECOND HALF
transition: 17 points in 12 possessions (1.42 ppp);
non-transition: 18 points in 26 possessions (0.69 ppp);

So (assuming first half was mostly man and second half was mostly zone) I conclude that our halfcourt defense was pretty solid the whole game. Despite how it looked, the difference in the game wasn't really zone vs. man (and if it was, our D was slightly better in the first half). It was really that we finally cut down on Miami transition opportunities.

We allowed zero transition opportunities after the 8-minute mark of the second half, until there was just 41 seconds to play. The score after Miami's transition basket with 8:00 to play was 66-53 Miami. The score when they next had a transition opportunity (at 0:41) was 78-70 Duke. And that was the ballgame.

What about possessions where they had a transition FG attempt, missed it, got the OR and then ran their regular half court D. I'm not sure if it happened last night but if it did do you count the entire possession as transition, or do you count them in both.?

[Kedsy]

What about possessions where they had a transition FG attempt, missed it, got the OR and then ran their regular half court D. I'm not sure if it happened last night but if it did do you count the entire possession as transition, or do you count them in both.?

I counted it as transition if they missed a transition FG attempt, got an offensive rebound and put it right back up again. That happened on two possessions. I don't remember seeing a transition miss, offensive rebound and then more than a couple seconds run off (probably because Miami is really bad at halfcourt offense, so why would they do that?), but if I had I would not have counted it.

[CDu]

What about possessions where they had a transition FG attempt, missed it, got the OR and then ran their regular half court D. I'm not sure if it happened last night but if it did do you count the entire possession as transition, or do you count them in both.?

There were no such possessions in the game. For one thing, all of Miami’s offensive rebounds came on possessions longer than 8 seconds. For another, they all resulted in immediate baskets or turnovers. So there were no “offensive rebound, let’s slow it down” possessions for them.

[Kedsy]

This 247 article (https://247sports.com/college/duke/Bolt/ESPN-analyst-discusses-Dukes-defensive-issues-113354232) from January 7 (i.e., doesn't include stats from Pitt or Wake games) says Duke is playing zone on 28% of opposing halfcourt possessions and that Duke's opposing points per possession in zone this season is 0.856 while Duke's opposing ppp in man was 0.876.

After going through the Miami play-by-play (in which 8 points were scored from transition possessions that wouldn't have been caught by Hoop-Math's first-shot analysis), I believe 247's numbers (which were calculated before the Miami game) might be accurate. Or, at least, they are credible.

If they're right, then my calculations rate our zone D approximately 1.5 points per 100 possessions better than our M2M D. If that's correct, then if we'd played zone 100% of the time instead of 28% of the time, it extrapolates to around #55 on KenPom. If we'd played 100% m2m, it extrapolates to around #84. With our current combination, we are actually ranked #72.

However, the past three games I think both our zone and our m2m have been better than they were earlier in the season. Our next two games are against Pitt and Wake Forest (again), and then Virginia. None of these teams are offensive powerhouses, but Wake (#52 KenPom offense on the road) and UVa (#30) are decent tests. Hopefully both our zone and our m2m will continue to improve.

That said, potentially the biggest additional jump we can make in defensive efficiency is to reverse the recent seeming trend of allowing such a large number of transition opportunities. Opposing fast break points have been more than 10% of opponent's total points in four of our last five games, after that phenomenon occurring in only three of our first 13 games. If we can keep that down to 7% or 8%, while maintaining our halfcourt defense of the past three games, I think we'll be on the right track.

[uh_no]

Obliquely regarding the discussion about man vs. zone defense in the Miami game, I went back to the official box score and checked transition possessions. I counted as a "transition opportunity" any time Miami did something (shot attempt, free throw, turnover) within 8 seconds of a possession change (e.g, Duke missed shot, made shot, turnover; but I didn't count it if there was a stoppage of play in the middle). This is different from "fast break" points, but I figured 8 seconds probably isn't enough time to get your defense set, man or zone.

Here are the results:

transition possessions: 25
transition points: 41
transition ppp: 1.64

non-transition possessions: 54
non-transition points: 34
non-transition ppp: 0.63

I don't know which non-transition possessions were man-to-man and which were zone, but here's the breakdown by half:

FIRST HALF
transition: 24 points in 13 possessions (1.85 ppp);
non-transition: 16 points in 28 possessions (0.57 ppp);

SECOND HALF
transition: 17 points in 12 possessions (1.42 ppp);
non-transition: 18 points in 26 possessions (0.69 ppp);

So (assuming first half was mostly man and second half was mostly zone) I conclude that our halfcourt defense was pretty solid the whole game. Despite how it looked, the difference in the game wasn't really zone vs. man (and if it was, our D was slightly better in the first half). It was really that we finally cut down on Miami transition opportunities.

We allowed zero transition opportunities after the 8-minute mark of the second half, until there was just 41 seconds to play. The score after Miami's transition basket with 8:00 to play was 66-53 Miami. The score when they next had a transition opportunity (at 0:41) was 78-70 Duke. And that was the ballgame.


the 1 ppp matches what I came up with earlier in the year when I did it for....maybe florida? Hard to know whether it's intrinsic to fast break points, or duke is especially bad at defending it? I'm guessing it's a bit of both. but the medicine is the same: more cowbell...wait no...don't turn the ball over, and don't take inefficient shots. For this team, the offense really can feed the defense.

Also of note, your calculation will make the transition defense look worse than it might actually be since it has a selection bias. It won't count possessions where nothing happened BECAUSE duke played good d, AKA, "pull it back out." Watching the game, that was probably scant few, but an important thing to consider. When I did the calculation, I swung the opposite way and took all possessions which could have been transition (defensive rebound and live ball TO) and counted whatever happened, including offensive rebounds and put backs. The theory is that even if it wasn't a transition possession in reality, it should be as efficient as a normal possession, which serves to make our transition defense look better than it might actually be, which is the side you want to lean on if you're trying to prove that there is a difference between the two. And it was still bad. In any case, i doubt it's hugely significant, and as I said, the fix is the same.

[Kedsy]

the 1 ppp matches what I came up with earlier in the year when I did it for...maybe florida? Hard to know whether it's intrinsic to fast break points, or duke is especially bad at defending it? I'm guessing it's a bit of both. but the medicine is the same: more cowbell...wait no...don't turn the ball over, and don't take inefficient shots. For this team, the offense really can feed the defense.

I'm not sure it matters whether it's 1 ppp difference or 0.8 ppp difference or whatever. Since transition opportunities are so much more efficient than halfcourt, an important aspect of defense (or of overall play, since as you note turnovers and long rebounds tend to turn into transition opportunities fairly often), it behooves us to try to limit those opportunities.

You're saying I undercounted transition opportunities, but even using my count, transition opportunities were 32% of Miami's possessions. Cut that down to 24%, and you've turned 6 possessions from high-efficiency transition to low-efficiency halfcourt. If it's a 1ppp difference, that's a 6-point swing. Even if it's 0.75 ppp difference, that's a 4-point swing. Either way, it would have a much bigger impact than any other defensive tweak I can think of.

Also, it's not just limiting turnovers and bad shots. It's also hustling back on D. Maybe not crashing the o-boards quite so hard, though that's a two-edged sword since offensive rebounding is such an important part of our offense. There might be other ways to limit transition as well, though I can't think of any others off the top of my head, other than just paying attention to the game situation and not lollygagging or pouting when you miss a shot or make an offensive mistake.

[uh_no]

I'm not sure it matters whether it's 1 ppp difference or 0.8 ppp difference or whatever. Since transition opportunities are so much more efficient than halfcourt, an important aspect of defense (or of overall play, since as you note turnovers and long rebounds tend to turn into transition opportunities fairly often), it behooves us to try to limit those opportunities.

You're saying I undercounted transition opportunities, but even using my count, transition opportunities were 32% of Miami's possessions. Cut that down to 24%, and you've turned 6 possessions from high-efficiency transition to low-efficiency halfcourt. If it's a 1ppp difference, that's a 6-point swing. Even if it's 0.75 ppp difference, that's a 4-point swing. Either way, it would have a much bigger impact than any other defensive tweak I can think of.

yep. we're 100% on the same page. It's even bigger than that, though, since you limiting those high-efficiency miami possessions, you're likely also increasing your own offensive efficiency (since you didn't turn it over, at least in some cases).

Fortunately, we KNOW this team can play good offense....and so many of those turnovers were frustration...trying to force things when the offense wasn't flowing. I expect that there will be work on how to avoid falling into that kind of trap of letting the turnovers build on each other.

Ultimately, that's what you expect from freshman, though. If the non-transition numbers are any indication, they've done a lot of growing up already (time will tell!), and i thus I would fully expect them to overcome this as well.

To answer your question you asked the other day, if this team did finish at the top of the standings, it would significantly decrease the correlation between both freshman minutes and average age and defense...whether it's enough to chalk it up wholely to random noise is up to interpretation. I think there is still enough to say "old duke teams don't play bad defense, but young duke teams might." Which is pretty much just saying "with one and dones and duke defense, it's a crapshoot."

[Kedsy]

Which is pretty much just saying "with one and dones and duke defense, it's a crapshoot."

Thanks. Hard to imagine anyone arguing with that statement.

[House P]

Thanks for all this info, House P. I am interested in how you calculated the adjusted efficiency.

My understanding (which could be wrong and I would be happy to be corrected) is that the formula KenPom uses for Adjusted Defensive Efficiency is something like the following:

Adjusted Defensive Efficiency = DE/OppAdjO * HCA * AvgD1Efficiency

Where,

DE = a team’s unadjusted defensive efficiency for a game

OppAdjO = the opponent’s adjusted offensive efficiency for the season

HCA = the home court advantage (1.028 for a home game, 1/1.028 for a road game, and 1 for a neutral site. I am not sure where I got these specific values, but they translate to a home court advantage of about 3.8 points per game for the average D1 team.)

AvgD1Efficiency = the average efficiency across D1 for the season (currently 104.4 for the 2017-18 season)


Here’s is an example of Duke’s AdjD vs Miami

DE = 95.0 (75 points allowed in 79 possessions)
OppAdjO = 108.2 (Miami’s season-long AdjO)
HCA = 0.973 (road game for Duke)
AvgD1Efficiency = 104.4

Therefore, Duke’s AdjD vs Miami = (95.0/108.2) * 0.971 * 104.4 = 89.2





When calculating his overall ratings, my understanding is that KenPom applies three additional adjustments:

1. Early in the season, the overall ratings are influenced by his preseason ratings.
2. Late in the season, there is an adjustment to reduce the influence of early season games.
3. Throughout the year, there is an adjustment to reduce the influence of blowouts against overmatched teams.

I think the “blowout adjustment” is has the biggest impact for his overall final ratings. In his blog, KenPom provide a vague description of the adjustment, but I don’t know exactly how he adjusts for blowouts. For Duke’s 2018 team, I find that I can replicate KenPom’s overall ratings within a couple tenths of a point if I ignore 20+ point wins vs teams ranked above 125 in his overall ratings.

[sammy3469]

I'm not sure it matters whether it's 1 ppp difference or 0.8 ppp difference or whatever. Since transition opportunities are so much more efficient than halfcourt, an important aspect of defense (or of overall play, since as you note turnovers and long rebounds tend to turn into transition opportunities fairly often), it behooves us to try to limit those opportunities.

You're saying I undercounted transition opportunities, but even using my count, transition opportunities were 32% of Miami's possessions. Cut that down to 24%, and you've turned 6 possessions from high-efficiency transition to low-efficiency halfcourt. If it's a 1ppp difference, that's a 6-point swing. Even if it's 0.75 ppp difference, that's a 4-point swing. Either way, it would have a much bigger impact than any other defensive tweak I can think of.

Also, it's not just limiting turnovers and bad shots. It's also hustling back on D. Maybe not crashing the o-boards quite so hard, though that's a two-edged sword since offensive rebounding is such an important part of our offense. There might be other ways to limit transition as well, though I can't think of any others off the top of my head, other than just paying attention to the game situation and not lollygagging or pouting when you miss a shot or make an offensive mistake.

It should be noted that the past two games were the worst two O-rebounding games by far this year. It also corresponds to both Javin and Bolden being out (which makes sense since they are good offensive rebounders).

Correspondingly, I'd also say the offensive rebounding has been limiting transition opportunities for the opposition based off of the relatively poor the floor balance of Duval, Allen, and Trent on offense. That floor balance has been improving (it's been really bad at times earlier in the year), but anytime Duval drives with Allen or Trent in the corner/wing with the 2 bigs underneath, you're opening up the possibility of a transition opportunity if the 2 bigs don't get the O-reb. Same basic idea happens if Allen or Trent misses from the wing.

You're essentially setting yourself up to only have 2 guys back on defense with one of those two running back hard from the wing. If we play two bigs they both have to offensive rebound and/or stop the initial outlet pass if they don't get that rebound (UNC does an excellent job of this).

[uh_no]

My understanding (which could be wrong and I would be happy to be corrected) is that the formula KenPom uses for Adjusted Defensive Efficiency is something like the following:

Adjusted Defensive Efficiency = DE/OppAdjO * HCA * AvgD1Efficiency



i believe it's now additive, though. meaning without HCA, it's DE - opp O + average

fwiw, his goal is to make the average 100, even though it tracks slightly....but that's a safe number to use.

[Neals384]

So (assuming first half was mostly man and second half was mostly zone) I conclude that our halfcourt defense was pretty solid the whole game. Despite how it looked, the difference in the game wasn't really zone vs. man (and if it was, our D was slightly better in the first half). It was really that we finally cut down on Miami transition opportunities .

Except that the second half was not mostly zone. There were two possessions of zone between the 15 and 14 minute mark, then all zone starting at 9:00. The rest was man.

[CDu]

Except that the second half was not mostly zone. There were two possessions of zone between the 15 and 14 minute mark, then all zone starting at 9:00. The rest was man.

Per Kedsy’s point, it doesn’t matter. Man-to-man and zone defenses were both very effective Monday. What wasn’t effective was our transition defense. And that was spurred more by turnovers.

We happened to be in zone for the last 8 minutes, and our defense looked fabulous over that time. But it was largely because we stopped giving them easy transition baskets off turnovers.

[Wander]

Per Kedsy’s point, it doesn’t matter. Man-to-man and zone defenses were both very effective Monday. What wasn’t effective was our transition defense. And that was spurred more by turnovers.

We happened to be in zone for the last 8 minutes, and our defense looked fabulous over that time. But it was largely because we stopped giving them easy transition baskets off turnovers.

It's like in fantasy football when the quarterback of the team you're playing at D/ST throws a pick six. It doesn't really have anything to do with your defense, but it gets counted against you anyway. Not exactly the same because some turnover buckets have a realistic chance of running back and making a play, but similar.

The defense was really good Monday (but Miami's offense sucks, so we'll see if it holds up).

[mkirsh]

While I was researching the above answer, it occurred to me that if transition scoring is so much more efficient than halfcourt scoring, it's possible we're not entirely focusing on the right things when we evaluate Duke's D. I went back to Hoop-Math and assembled the following table with some Duke transition defense information:



Year %transition trans eFG% trans %3att trans 3pt% Pom D Rank
2018 20.8% 54.5% 39.2% 34.6% 86
2017 23.8% 58.4% 28.0% 32.3% 47
2016 24.4% 55.7% 30.6% 30.2% 86
2015 23.5% 52.2% 30.6% 25.9% 11
2014 22.1% 54.5% 30.7% 24.1% 86
2013 23.9% 51.7% 32.7% 33.3% 26
2012 22.1% 58.4% 27.3% 33.3% 79


I'm not entirely sure what to make of this data, but I have a few semi-random observations:

- note that the eFG% is on initial transition shot attempt. I couldn't find any data on transition putbacks or transition free throws.

- Our two best defensive teams in the period had the two best opposing transition eFG%. That said, if this year's team let up an initial transition shot attempt eFG% of 52.2% like the 2015 team did, it would only be a 0.73 points per game improvement in opposing scoring (or about 1 point per 100 possessions). Which isn't close to enough to explain the difference in the 2013 or 2015 defenses and this year's defense. Though a one point improvement in opposing scoring per 100 possessions would move our current KenPom defensive rank from #86 to #67.

- this year's team is giving up by far the highest percentage of opposing transition three-point attempts, and also the highest transition three-point shooting percentage. This is in transition, though, not in the halfcourt. Are there not enough guys getting back in transition D? Are they not getting back fast enough? Are they running to the wrong spot? Or is it just bad luck which could even out over time?

- this year's team gives up the lowest percentage of transition opportunities of any team in the period. Possibly because we're so good at offensive rebounding. Whatever the reason, since transition scoring is so much more efficient than halfcourt scoring, this can only be a good thing, right? But does it mean that we have more potential to end up with a good defense because we can limit transition opportunities? Or, since our overall dRtg is so mediocre, does the fact that we're already limiting transition opportunities mean our halfcourt defense is even worse than we think it is? Or both?

I don't really know the answers to any of these questions, but I thought it was interesting to think about.

This is fascinating data, thanks for doing so much legwork!

Is there an even more micro cut available? It would be interesting to see % and points per possession for transition offense that initiates from turnovers vs transition that initiates off of rebounds. I would assume turnovers is a much higher % of transition offense and much more effective in scoring and that few transition opportunities come from missed shots, which would lead to not changing strategy of crashing offensive boards, but would be interested to see if the data bears that out.

Also, do KenPom or other sources have aggregates stats? Per your point above that more teams are shooting transition threes against DUke, I would posit that is universal - that all teams are shooting more threes in all situations following the NBA's lead in increasing utilization of the three point shot. Is there data available to see whether that is actually true?

[JasonEvans]

It's like in fantasy football when the quarterback of the team you're playing at D/ST throws a pick six. It doesn't really have anything to do with your defense, but it gets counted against you anyway. Not exactly the same because some turnover buckets have a realistic chance of running back and making a play, but similar.

Actually, most fantasy football leagues do not count defensive or special teams touchdowns (pick six, returned kick, returned fumble, blocked kick) as points against the defense. If you have the Ravens D and they give up 21 points in the game due to 1 offensive TD plus 1 pick six and a returned kickoff, the Ravens D has only given up 9 points in the game (TD + 3 point after kicks) according to fantasy scoring.

At least that is the way both Yahoo and ESPN score their leagues (I think).

-Jason "you may now return to the truly excellent conversation about Duke's defense" Evans

[ChillinDuke]

Actually, most fantasy football leagues do not count defensive or special teams touchdowns (pick six, returned kick, returned fumble, blocked kick) as points against the defense. If you have the Ravens D and they give up 21 points in the game due to 1 offensive TD plus 1 pick six and a returned kickoff, the Ravens D has only given up 9 points in the game (TD + 3 point after kicks) according to fantasy scoring.

At least that is the way both Yahoo and ESPN score their leagues (I think).

-Jason "you may now return to the truly excellent conversation about Duke's defense" Evans

Is that true? If so, I haven't noticed it that way.

- Chillin

[Kedsy]

This is fascinating data, thanks for doing so much legwork!

Is there an even more micro cut available? It would be interesting to see % and points per possession for transition offense that initiates from turnovers vs transition that initiates off of rebounds. I would assume turnovers is a much higher % of transition offense and much more effective in scoring and that few transition opportunities come from missed shots, which would lead to not changing strategy of crashing offensive boards, but would be interested to see if the data bears that out.

Also, do KenPom or other sources have aggregates stats? Per your point above that more teams are shooting transition threes against DUke, I would posit that is universal - that all teams are shooting more threes in all situations following the NBA's lead in increasing utilization of the three point shot. Is there data available to see whether that is actually true?

It's possible sites like Synergy Sports Tech has info like this, but it costs money (I'm not sure how much, but my guess is it's significant) and I'm not willing to go that far to fuel my obsession.

For what it's worth, in the Miami game, here's my breakdown based on what initiated the transition opportunity:

live ball turnovers: 10 opportunities, 18 points (1.80 ppp);
missed shots: 12 opportunities, 20 points (1.67 ppp);
made shots (all in last minute of game): 3 opportunities, 3 points (1.00 ppp).

So, at least in this game, turnovers (40%) were not a much higher percentage of transition offense, compared to missed shots (48%), though turnover-initiated opportunities were slightly better in efficiency.

[kAzE]

Is that true? If so, I haven't noticed it that way.

- Chillin

ESPN made that change to their scoring system either last year or this year I believe. It was fairly recent.

[CDu]

ESPN made that change to their scoring system either last year or this year I believe. It was fairly recent.

This year, yes.

[Ian]

It's possible sites like Synergy Sports Tech has info like this, but it costs money (I'm not sure how much, but my guess is it's significant) and I'm not willing to go that far to fuel my obsession.

For what it's worth, in the Miami game, here's my breakdown based on what initiated the transition opportunity:

live ball turnovers: 10 opportunities, 18 points (1.80 ppp);
missed shots: 12 opportunities, 20 points (1.67 ppp);
made shots (all in last minute of game): 3 opportunities, 3 points (1.00 ppp).

So, at least in this game, turnovers (40%) were not a much higher percentage of transition offense, compared to missed shots (48%), though turnover-initiated opportunities were slightly better in efficiency.

Interesting, for the live ball turnovers, there is not much we can do defensivesly, just have to cut the number of TOs down.
The missed shots though, is a concern because they only grabbed 25 DREB for the entire game, so they are converting 48% of their defensive rebounds into fastbreak opportunities, I suspect that number is very high and shows our poor transition defense. Without looking up any stats I suspect the ncaa average is somewhere between 25-35%. And that's ignoring that some of the defensive rebounds were off missed FTs which are difficult to initiate a transition opportunity from, so the real number off a missed FG attempt were probably in the mid to high 50%s.

[Kedsy]

Interesting, for the live ball turnovers, there is not much we can do defensivesly, just have to cut the number of TOs down.
The missed shots though, is a concern because they only grabbed 25 DREB for the entire game, so they are converting 48% of their defensive rebounds into fastbreak opportunities, I suspect that number is very high and shows our poor transition defense. Without looking up any stats I suspect the ncaa average is somewhere between 25-35%. And that's ignoring that some of the defensive rebounds were off missed FTs which are difficult to initiate a transition opportunity from, so the real number off a missed FG attempt were probably in the mid to high 50%s.

Miami had 5 rebounds off of missed free throws (and none of those resulted in transition opportunities). Meaning they turned 12 of 20 defensive rebounds off of missed field goals into transition opportunities (60%). Though I suspect this season Miami is better at running off of missed shots than most teams.

[Kedsy]

Except that the second half was not mostly zone. There were two possessions of zone between the 15 and 14 minute mark, then all zone starting at 9:00. The rest was man.

Thanks, Neals. So the rest of the 2nd half was all m2m? Do you know in which (if any) possessions we employed zone in the first half?

[Neals384]

Per Kedsy’s point, it doesn’t matter. Man-to-man and zone defenses were both very effective Monday. What wasn’t effective was our transition defense. And that was spurred more by turnovers.

We happened to be in zone for the last 8 minutes, and our defense looked fabulous over that time. But it was largely because we stopped giving them easy transition baskets off turnovers.

Understand your poiint, but I felt that "played zone entire 2nd half" was so far off it should be corrected.

[Acymetric]

Actually, most fantasy football leagues do not count defensive or special teams touchdowns (pick six, returned kick, returned fumble, blocked kick) as points against the defense. If you have the Ravens D and they give up 21 points in the game due to 1 offensive TD plus 1 pick six and a returned kickoff, the Ravens D has only given up 9 points in the game (TD + 3 point after kicks) according to fantasy scoring.

At least that is the way both Yahoo and ESPN score their leagues (I think).

-Jason "you may now return to the truly excellent conversation about Duke's defense" Evans

I believe you are right about defensive points (fumble recoveries, interceptions, and safeties), but not special teams TDs (so the points allowed in your example would be 15 for the offensive TD and the kickoff return plus the XP from the interception return). This makes sense since you actually have the Ravens D/ST not just the Ravens D so you are penalized for allowing special teams touchdowns but also get points for Ravens special teams touchdowns. This is at least true for Yahoo/ESPN, obviously some sites allow more customization or have different default rules for this. Curiously, in Yahoo a blocked punt returned for a TD counts against you but a blocked field goal returned for a TD does not (ESPN does not make this distinction and counts both against your defensive points allowed as far as I know).

To bring this back on topic...this recent back and forth seems like a long, overly technical conversation of advanced metrics that boils down to "allowing transition opportunities makes your defense look bad and we have been allowing more of those than we would like" which I'm not sure we needed Kenpom or any other statistics service to tell us although I do really like them for other types of analysis.

[English]

I believe you are right about defensive points (fumble recoveries, interceptions, and safeties), but not special teams TDs (so the points allowed in your example would be 15 for the offensive TD and the kickoff return plus the XP from the interception return). This makes sense since you actually have the Ravens D/ST not just the Ravens D so you are penalized for allowing special teams touchdowns but also get points for Ravens special teams touchdowns. This is at least true for Yahoo/ESPN, obviously some sites allow more customization or have different default rules for this. Curiously, in Yahoo a blocked punt returned for a TD counts against you but a blocked field goal returned for a TD does not (ESPN does not make this distinction and counts both against your defensive points allowed as far as I know).

To bring this back on topic...this recent back and forth seems like a long, overly technical conversation of advanced metrics that boils down to "allowing transition opportunities makes your defense look bad and we have been allowing more of those than we would like" which I'm not sure we needed Kenpom or any other statistics service to tell us although I do really like them for other types of analysis.

You're absolutely right that the bolded is obvious, and we don't need someone with an expensive calculator to tell us that. I think the bigger point, that at least I choose to focus on, is that what we previously took to be abysmal defense all around may actually be a better-than-thought halfcourt defense in both M2M and zone, but a poor transition defense exacerbated by a tendency to have stretches of turnover-prone ball and/or suboptimal shot selection. The problems are different, the fixes are different, but from my view, it's easier to solve the latter than completely rebuild the entire defense from the ground up. YMMV.

[CDu]

You're absolutely right that the bolded is obvious, and we don't need someone with an expensive calculator to tell us that. I think the bigger point, that at least I choose to focus on, is that what we previously took to be abysmal defense all around may actually be a better-than-thought halfcourt defense in both M2M and zone, but a poor transition defense exacerbated by a tendency to have stretches of turnover-prone ball and/or suboptimal shot selection. The problems are different, the fixes are different, but from my view, it's easier to solve the latter than completely rebuild the entire defense from the ground up. YMMV.

I would suggest that our half-court defense was previously bad too. We aren’t typically turnover-prone, so our defensive problems were still half-court problems. It has improved in the three most recent games. For instance, in spite of being awfully sloppy with the ball against Miami, we were still having a decent defensive night. Had we been playing halfcourt defense like earlier in the season, we lose that game.

[Kedsy]

I would suggest that our half-court defense was previously bad too. We aren’t typically turnover-prone, so our defensive problems were still half-court problems. It has improved in the three most recent games. For instance, in spite of being awfully sloppy with the ball against Miami, we were still having a decent defensive night. Had we been playing halfcourt defense like earlier in the season, we lose that game.

I agree with this. Against NC State, we put up an appalling 1.23 dRtg while only giving up 7 "fast break points" (we gave up 31 fast break points against Miami). Against BC, we had a dreadful 1.17 dRtg while allowing 0 (zero) fast break points. A couple other games were also bad without giving up a ton of fast break points.

Our halfcourt D has improved a lot in the past three games, but we've also given up more transition opportunities than we did earlier in the season. I don't know if these are related (hard to see exactly how they would be), but if they're not then good overall defense should theoretically be attainable by sustaining the recent halfcourt D improvements while also cutting down on opposing transition opportunities.

[Kedsy]

Duke defensive stats vs. Pitt (round 2):

2-point%: .485
3-point%: .250
%threes: 32.7%
eFG%: 44.9%

TO%: 23.6%
DR%: 76.7%
FT Rate: 28.6%
dRating: 0.85

A/to: 0.47:1
Asst Rate: 35.0%
Block 14.3%: (21.2% of twos)

Opposing fast break points: 2

Efficiency difference: 11.0 (opposing two efficiency better than opposing three efficiency).

This game was similar defensively to the first Pitt game. Low possession (at least for Duke), bad opposing three-point shooting, mediocre two-point defense by Duke, high-ish free throw rate (not nearly as high as the first Pitt game though), good job forcing turnovers. We did do a better job on the defensive boards this game than we did in the game in Pittsburgh.

It's a shame we can't force turnovers against other ACC teams like we seem to be able to against the Panthers. It would really add a good dimension to our D. We also only gave up 2 fast break points tonight (last time we gave up 6 fb points to Pitt, but that constituted 11.5% of their overall points). I mean, it was Pitt, but it's good to see the improvement in this area.

[Kedsy]

Duke defensive stats vs. Wake Forest (round 2):

2-point%: .524
3-point%: .250
%threes: 32.3%
eFG%: 47.6%

TO%: 28.7%
DR%: 66.0%
FT Rate: 21.0%
dRating: 0.96

A/to: 0.52:1
Asst Rate: 40.7%
Block 3.2%: (4.8% of twos)

Opposing fast break points: 13 (18.6% of WF's points)

Efficiency difference: 14.9 (opposing two efficiency better than opposing three efficiency).

This game was similar defensively to the Pitt games. We didn't give up that many three attempts, and they didn't make many of the ones they did take, but Wake did way too well on two-pointers. We only blocked two shots, our defensive rebounding was not so good, and we once again gave up a bunch of fast break points. What saved us was we forced a ton of turnovers, Duke's second-best TO% of the season, behind only the Evansville rout. Overall, a decent but not great defensive performance.

[Hingeknocker]

I try not to pay too much attention to my emotion-laden analysis while watching a game, but man it seemed like a ton of Wake's ORebs came because they missed shots so badly that there were either long rebounds or our guys were just out of position, through no real fault of their own. Gotta box out, of course, but I'm much more willing to focus on the better on-ball defense that I saw and trust that the rebounding won't be like that again.

[Kedsy]

I try not to pay too much attention to my emotion-laden analysis while watching a game, but man it seemed like a ton of Wake's ORebs came because they missed shots so badly that there were either long rebounds or our guys were just out of position, through no real fault of their own. Gotta box out, of course, but I'm much more willing to focus on the better on-ball defense that I saw and trust that the rebounding won't be like that again.

Our defensive rebounding (66%) wasn't terrible, it just wasn't as good as it usually is. Biggest defensive problem tonight was (once again) giving up too many transition opportunities. Fortunately, we made up for that by turning them over a lot. Our 28.7% forced turnover pct was almost twice as good as the last time we played Wake (15.1% in that game). That was the real takeaway here. :cool:

[TheOldBattleship]

We didn't give up that many three attempts, and they didn't make many of the ones they did take, but Wake did way too well on two-pointers. We only blocked two shots, our defensive rebounding was not so good, and we once again gave up a bunch of fast break points. What saved us was we forced a ton of turnovers, Duke's second-best TO% of the season, behind only the Evansville rout. Overall, a decent but not great defensive performance.

The high opponent FG% on twos in exchange for getting turnovers and running guys off the three as much as possible seems to be the deal with the devil that K is willing to make when playing this higher pressure man to man over the last two or three games. We're pushing way out and really pressuring with the intent of forcing turnovers, so when we get beat, we really get beat, and all the help is spread all over the floor waiting to deny the ball or jump a passing lane, so there's much less traffic/shot blocking at the rim than when we were keeping things a bit more collapsed. It's worked well against some of the lesser teams in the ACC. I'm intrigued to see if K sticks with it against UVA, a team that NEVER turns the ball over, or if he dials it back a bit. I will say that I really like the contrast between the frenetic man to man and the sudden slow-down of the zone. We saw today how effective that change of pace is, especially in the second half when there just aren't that many possessions left in the game, and a team has to basically waste two or three trying to get the switch flipped to get into their zone offense.

[pfrduke]

The high opponent FG% on twos in exchange for getting turnovers and running guys off the three as much as possible seems to be the deal with the devil that K is willing to make when playing this higher pressure man to man over the last two or three games. We're pushing way out and really pressuring with the intent of forcing turnovers, so when we get beat, we really get beat, and all the help is spread all over the floor waiting to deny the ball or jump a passing lane, so there's much less traffic/shot blocking at the rim than when we were keeping things a bit more collapsed. It's worked well against some of the lesser teams in the ACC. I'm intrigued to see if K sticks with it against UVA, a team that NEVER turns the ball over, or if he dials it back a bit. I will say that I really like the contrast between the frenetic man to man and the sudden slow-down of the zone. We saw today how effective that change of pace is, especially in the second half when there just aren't that many possessions left in the game, and a team has to basically waste two or three trying to get the switch flipped to get into their zone offense.

One hundred percent agree with this. It is much more traditional Duke defense to give up a higher percentage shooting from 2 in exchange for limiting 3 point attempts and forcing turnovers. I welcome the return of that from the more passive defense that allowed tons of 3s earlier in the season.

[Ian]

Duke defensive stats vs. Wake Forest (round 2):

2-point%: .524
3-point%: .250
%threes: 32.3%
eFG%: 47.6%

TO%: 28.7%
DR%: 66.0%
FT Rate: 21.0%
dRating: 0.96

A/to: 0.52:1
Asst Rate: 40.7%
Block 3.2%: (4.8% of twos)

Opposing fast break points: 13 (18.6% of WF's points)

Efficiency difference: 14.9 (opposing two efficiency better than opposing three efficiency).

This game was similar defensively to the Pitt games. We didn't give up that many three attempts, and they didn't make many of the ones they did take, but Wake did way too well on two-pointers. We only blocked two shots, our defensive rebounding was not so good, and we once again gave up a bunch of fast break points. What saved us was we forced a ton of turnovers, Duke's second-best TO% of the season, behind only the Evansville rout. Overall, a decent but not great defensive performance.

It was pretty good, Wake is overall 109.6 Oeff, holding them to .96 is a adjusted .87 Deff, which their defensive performance is what you'd expect the #3 defense in the country would do. And that's not even taking into account the game is at Wake.

[Kedsy]

It's worked well against some of the lesser teams in the ACC.

Yeah, but not against anyone else. Not even against Wake, the first time.

I said this after the last (Pitt) game, but if we can force turnovers at a decent rate, our defense will not just be average (which some have persuasively argued is all we need), but will actually be good. If we can do that and also cut down on opposing transition opportunities, our D could be really good.

One thing at a time, I suppose.

[JasonEvans]

One small point... while it is nice to talk about how Duke did forcing turnovers, a heck of a lot of those TOs were not forced but were just giveaways by Wake. I swear, they probably threw at least 3 or 4 balls out of bounds for no reason at all. We had at least 2 steals that were just terribly lazy passes that any high school basketball player could have intercepted. Wake... was... awful in that game.

[Kedsy]

One small point... while it is nice to talk about how Duke did forcing turnovers, a heck of a lot of those TOs were not forced but were just giveaways by Wake. I swear, they probably threw at least 3 or 4 balls out of bounds for no reason at all. We had at least 2 steals that were just terribly lazy passes that any high school basketball player could have intercepted. Wake... was... awful in that game.

Yeah, but that's probably true in most games. Some percentage of turnovers are forced and some unforced. I'm sure that varies from team to team. And from game to game too, but I bet that doesn't vary as much as you might think. In the first Duke/Wake game, Wake turned the ball over in 15.1% of their possessions (12 turnovers). Wake for the season, on average turns it over on 19.5% of their possessions (13.8 topg). Meaning, they held onto the ball better in game 1 than they usually do, which I believe has been a fairly common phenomenon this season against Duke.

In this game, they turned it over in 28.7% of their possessions (21 turnovers). Huge difference from the first game. I'm sure some of that can be credited to Wake reverting to the mean and some maybe can be credited to them having an unusually bad ballhandling/passing performance last night. But not all, or even necessarily most. And in many ways it doesn't matter. If teams turn it over against us, our defensive numbers will be good and they'll score fewer points, and that's all that really matters, right?

I'm still not sure we've turned a corner in this regard. It's one thing to turn over Pitt and Wake, but against FSU and NCSU and Miami, our TO% was 13.5%, 13.4%, and 13.6%, respectively. We'll know more after we play Virginia and Notre Dame and UNC. But if moving forward our good opponents start turning the ball over at a better clip, I'm not going to care whether those turnovers were forced or unforced.

Put another way, we at DBR tend to obsess over our KenPom defensive rank. When BC hit 58% of their threes, our dRank tanked, even though probably a lot of those made shots weren't really our defense's fault. This game, our dRank improved, in large part because of the high number of turnovers, even though probably some of the turnovers couldn't be fairly attributed to better defense on our part. The "luck factor" tends to even out over the course of a season.

[mkirsh]

Duke defensive stats vs. Wake Forest (round 2):

2-point%: .524
3-point%: .250
%threes: 32.3%
eFG%: 47.6%

TO%: 28.7%
DR%: 66.0%
FT Rate: 21.0%
dRating: 0.96

A/to: 0.52:1
Asst Rate: 40.7%
Block 3.2%: (4.8% of twos)

Opposing fast break points: 13 (18.6% of WF's points)

Efficiency difference: 14.9 (opposing two efficiency better than opposing three efficiency).

This game was similar defensively to the Pitt games. We didn't give up that many three attempts, and they didn't make many of the ones they did take, but Wake did way too well on two-pointers. We only blocked two shots, our defensive rebounding was not so good, and we once again gave up a bunch of fast break points. What saved us was we forced a ton of turnovers, Duke's second-best TO% of the season, behind only the Evansville rout. Overall, a decent but not great defensive performance.

I thought the mtm defense last night was really good, on both the numbers and the eye test. The scheme was solid, dropping on middle screens and icing on the side ball screens, and the players executed well, regularly in position, moving well, and communicating. Very few Wake points were from a player getting lost and being totally out of position. Rebounding was a little weaker than normal, but if Bagley wasn't limited with fouls and played six more minutes, his per minute rebounding stats extrapolate to him getting 2 more boards, taking the defensive rebounding % back up to 72%, which is more in line with the rest of the season.

Conversely I didn't think the zone was as effective in preventing points, but where it was used brilliantly was to shorten the game. While Wake was able to score against the zone, it took them a long time to do so, which was really well used late in the game.

Huge game ahead - with a win this weekend Duke would be only 1 game back in the loss column.

[cato]

where [the zone] was used brilliantly was to shorten the game. While Wake was able to score against the zone, it took them a long time to do so, which was really well used late in the game.

Interesting. K went to the zone with, what, 9 minutes left at Miami?

Start in MTM and see how far the team goes with that, and then if time, score and situation dictate, go to the zone to reduce the number of possessions in the game?

[mkirsh]

Interesting. K went to the zone with, what, 9 minutes left at Miami?

Start in MTM and see how far the team goes with that, and then if time, score and situation dictate, go to the zone to reduce the number of possessions in the game?

I wasn't suggesting it was a pre-meditated strategy, but we will have to see if that plays out over the next few games. Against Miami I think Duke went zone as K thought it would be better than MTM rather than to slow the pace since Duke was down and probably wanted more possessions to catch up. Against Wake I think zone was used to 1) slow Wake down as Duke had the lead, 2) protect Bagley from his 4th/5th fouls, and 3) (though K would probably never say this) "rest" players on D since the rotation was tight in the second half.

Also interesting is that Duke in prior years, as well as the beginning of this one, used a 2-2-1 full court token press to change the pace of the game, but we haven't seen that in a while. Probably just a coaching decision to focus on mtm and zone rather than trying to add another piece at the moment. Did try a few mins of full court mtm against Pitt, so will see if that makes a comeback at some point.

[devildeac]

Looks like we're #65 now, tied with Northern Iowa, ND, Idaho and, wait for it, Wagner.

(too soon? :o:rolleyes:)

[duke4ever19]

Looks like we're #65 now, tied with Northern Iowa, ND, Idaho and, wait for it, Wagner.

(too soon? :o:rolleyes:)

#65 by which metric?

Kenpom currently has our D at #73, though earlier today we were at #71.

I'm not a member at kenpom so maybe I'm not seeing the most recent rankings.

[TheOldBattleship]

One small point... while it is nice to talk about how Duke did forcing turnovers, a heck of a lot of those TOs were not forced but were just giveaways by Wake. I swear, they probably threw at least 3 or 4 balls out of bounds for no reason at all. We had at least 2 steals that were just terribly lazy passes that any high school basketball player could have intercepted. Wake... was... awful in that game.

No argument here about the level of Wake's focus offensively throughout most of that game. That said, though, K has always said that this is more or less the point of that classic Duke pressure-heavy defense: if you pressure the ball a lot and cut off passing lanes, the opposing players tend make really bad mistakes offensively. And it's an accumulative effect. They start anticipating the pressure and doing dumb things to avoid it. Wake is a generally high-turnover team to begin with, and this Duke team certainly doesn't have the sort of on-ball hawks that we've had in past years or the level of communication required to make the pressure and overplay consistently possible, but I do think that those bad turnovers from last night can be partly ascribed to the effect that K has always looked for with the pressure D.

[CDu]

Interesting to see the difference between TRank and KenPom with respect to defensive efficiency. Pomeroy has us at an adjusted DEff of 98.0, good for 73rd nationally. TRank has us at 94.4, good for 36th. I haven’t researched the differences in methods to figure out exactly why the two differ so much. Just worth noting the difference.

[devildeac]

#65 by which metric?

Kenpom currently has our D at #73, though earlier today we were at #71.

I'm not a member at kenpom so maybe I'm not seeing the most recent rankings.

NCAA "rankings"

https://www.teamrankings.com/ncaa-basketball/stat/defensive-efficiency

[LasVegas]

Interesting to see the difference between TRank and KenPom with respect to defensive efficiency. Pomeroy has us at an adjusted DEff of 98.0, good for 73rd nationally. TRank has us at 94.4, good for 36th. I haven’t researched the differences in methods to figure out exactly why the two differ so much. Just worth noting the difference.

Based on my quick glance, it looks like the rankings are similar for the other teams on both systems. I wonder why duke has such a huge difference?

[House P]

Interesting to see the difference between TRank and KenPom with respect to defensive efficiency. Pomeroy has us at an adjusted DEff of 98.0, good for 73rd nationally. TRank has us at 94.4, good for 36th. I haven’t researched the differences in methods to figure out exactly why the two differ so much. Just worth noting the difference.

I suspect that this may have to do with the way KenPom adjusts his overall rankings to reduce the impact of blowouts vs overmatched teams.

The chart below shows what I get when I estimate Duke's game-by-game Adjusted Defensive Efficiency. Duke's average Adjusted Defensive Efficiency differs depending on how much weight you give to the 30+ point blowouts vs Evansville, St. Francis, and Utah Valley- Duke's three best game's in terms of Adjusted Defensive Efficiency.

- Duke's average the Adjusted Defensive Efficiency of all 20 games is 84.7. This is pretty close to the number you mention for T-rank (94.4).

- Duke's average Adjusted Defensive Efficiency is 98.0 if you don't include the Evansville, St. Francis, and Utah Valley games. This is the pretty close to KenPom's current number (98.2).

I suspect that KenPom's "blowout adjustment" is more nuanced than simply ignoring games decided by 30+ points, but KenPom doesn't provide details on his exact method and this approach seems to get fairly close to the values KenPom lists on his homepage.



Here is an estimate of Duke's game-by-game adjusted defensive efficiency. The values in the "rank" column represent where Duke would rank if they maintained the single game Adjusted Defensive Efficiency for the entire season.







Result
AdjDE
Rnk


Elon
W, 97-68
94.52
26


Utah Valley
W, 99-69
84.55
3


Michigan St.
W, 88-81
94.60
26


Southern
W, 78-61
100.46
106


Furman
W, 92-63
95.04
29


Portland St.
W, 99-81
96.24
42


Texas
W, 85-78
98.13
69


Florida
W, 87-84
97.83
64


Indiana
W, 91-81
110.66
298


South Dakota
W, 96-80
103.18
146


St. Francis PA
W, 124-67
78.62
1


Boston College
L, 89-84
110.23
288


Evansville
W, 104-40
64.77
1


Florida St.
W, 100-93
106.23
223


North Carolina St.
L, 96-85
114.63
332


Pittsburgh
W, 87-52
86.37
3


Wake Forest
W, 89-71
88.97
4


Miami FL
W, 83-75
87.85
3


Pittsburgh
W, 81-54
90.58
7


Wake Forest
W, 84-70
89.68
6

[House P]

I suspect that this may have to do with the way KenPom adjusts his overall rankings to reduce the impact of blowouts vs overmatched teams.

- Duke's average the Adjusted Defensive Efficiency of all 20 games is 84.7. This is pretty close to the number you mention for T-rank (94.4).

- Duke's average Adjusted Defensive Efficiency is 98.0 if you don't include the Evansville, St. Francis, and Utah Valley games. This is the pretty close to KenPom's current number (98.2).



For anyone who likes graphs more than charts, here is a plot of my estimates for Duke's game by game Adjusted Defensive Efficiency. For this plot, I ignored the blowouts vs Evansville, St. Francis, and Utah Valley.

This graph seems to show three fairly distinct phases.

1. Early in the season (Elon-Florida), Duke's defense was OK (average AdjDE of 96.7).
2. Over the next several games (Indiana - NC State), Duke's defense was terrible in games decided by < 50 points (average AdjDE of 109.0).
3. In the past five games (@Pitt - @Wake), Duke's defense has been outstanding (average AdjDE of 88.7).

That being said, I would still be hesitant to declare that Duke has found a "new normal" until they show they can consistently put up impressive numbers vs top 50 teams.

http://forums.dukebasketballreport.com/forums/attachment.php?attachmentid=8028&stc=1

[Kfanarmy]

For anyone who likes graphs more than charts, here is a plot of my estimates for Duke's game by game Adjusted Defensive Efficiency. For this plot, I ignored the blowouts vs Evansville, St. Francis, and Utah Valley.

This graph seems to show three fairly distinct phases.

1. Early in the season (Elon-Florida), Duke's defense was OK (average AdjDE of 96.7).
2. Over the next several games (Indiana - NC State), Duke's defense was terrible in games decided by < 50 points (average AdjDE of 109.0).
3. In the past five games (@Pitt - @Wake), Duke's defense has been outstanding (average AdjDE of 88.7).

That being said, I would still be hesitant to declare that Duke has found a "new normal" until they show they can consistently put up impressive numbers vs top 50 teams.

http://forums.dukebasketballreport.com/forums/attachment.php?attachmentid=8028&stc=1

awesome chart! It might be a bit more intuitive if the numbers were reversed....what appears as a downward trend, is actually increasingly better defense....at least that's how my brain wants to interpret it.

[Kedsy]

Duke defensive stats vs. Virginia:

2-point%: .455
3-point%: .273
%threes: 33.3%
eFG%: 43.9%

TO%: 7.7%
DR%: 76.2%
FT Rate: 13.6%
dRating: 0.99

A/to: 2.4:1
Asst Rate: 46.2%
Block 7.5%: (11.4% of twos)

Opposing fast break points: 2 (3.1% of UVa's points)

Efficiency difference: 4.6 (opposing two efficiency a bit better than opposing three efficiency).

Other than not turning them over, at all, nothing to be too upset about here from a defensive standpoint. Obviously it would have been nicer if we could have held them to three fewer points.

[Kedsy]

FWIW, Duke has now broken the top 60 (#59) in KenPom's dRank. That's down from #108 after the NC State game.

[CDu]

FWIW, Duke has now broken the top 60 (#59) in KenPom's dRank. That's down from #108 after the NC State game.

Yeah, I think we are past the point of random noise. The defense has been consistently strong over the 6 games since NC St. Let’s keep it up, along with a bit of a rebound from our offense which has slipped slightly over that stretch.

[duke4ever19]

FWIW, Duke has now broken the top 60 (#59) in KenPom's dRank. That's down from #108 after the NC State game.

So, taking for granted that it was a different season, different team, different competition . . . the last Duke championship team/Final Four team had started the tournament ranked #57 on defense and then notably improved to the low 10's by the time they won the title.

Do you think this year's team needs to show a similar improvement to win the whole thing, or (based on several factors) we might be able to win it with a ranking in the 30's or 40's?

I know there is a general KenPom 'profile' of what a championship team looks like (sample size 2002-2017), but I'm wondering if you think this team can either (1) break the mold on what it takes to win it on the defensive side, or (2) make a run at a finding itself in the typical range for a championship team on defense?

[Kedsy]

So, taking for granted that it was a different season, different team, different competition . . . the last Duke championship team/Final Four team had started the tournament ranked #57 on defense and then notably improved to the low 10's by the time they won the title.

Do you think this year's team needs to show a similar improvement to win the whole thing, or (based on several factors) we might be able to win it with a ranking in the 30's or 40's?

I know there is a general KenPom 'profile' of what a championship team looks like (sample size 2002-2017), but I'm wondering if you think this team can either (1) break the mold on what it takes to win it on the defensive side, or (2) make a run at a finding itself in the typical range for a championship team on defense?

I think a large part of that answer lies in who we play. In 2015, we played all (or almost all) slow-tempo, not-super-athletic teams that allowed our defense to play up to its potential (or put another way, teams that weren't equipped to take advantage of our biggest defensive weaknesses).

Also, in the past 6 games, we've probably played like a top 20 defense. For the purposes of winning the championship would it matter whether our final ranking was in the 30's or 40's (or even worse), so long as we played good defense in the 6 games of the tournament? I don't think it would.

There really isn't a reliable "mold" we'd need to break. A fairly high percentage of Final Four teams went into the tournament with non-top 30 defenses. Some of them even ended with non-top 30 defenses (Wisconsin in 2015, for example, had a final KenPom dRank of #35; they went into the tournament at #30, which suggests they didn't drastically improve their D in the tourney, either, but they still came pretty close to winning the trophy).

So to answer your question, sure, we could win the tournament with a dRank in the 30's or 40's. We could also have a dRank in the 10s or 20s but run into a hot team and get eliminated in the 2nd or 3rd round (like our 2011 team, which had a final KenPom dRank of #9 but just couldn't stop Arizona for the 2nd half of our Sweet 16 game).

[Ultrarunner]

Yeah, I think we are past the point of random noise. The defense has been consistently strong over the 6 games since NC St. Let’s keep it up, along with a bit of a rebound from our offense which has slipped slightly over that stretch.

I think we'll see another up-tick in defense when we get Bolden and Javin back, fully healthy. Add in more rest for Carter and Bagley and they should be more effective at both ends.

Offensively, we have streaky shooters on the outside. Hoping they get hot as the same time, at the right time.

[CDu]

I think we'll see another up-tick in defense when we get Bolden and Javin back, fully healthy. Add in more rest for Carter and Bagley and they should be more effective at both ends.

Offensively, we have streaky shooters on the outside. Hoping they get hot as the same time, at the right time.

That and foul trouble/injury are the big concerns. We lost yesterday in large part because all 3 perimeter guys struggled. If one of those guys is on and Carter and Bagley are effective, we should beat anyone. If those three guys are all off, we need Carter and Bagley to play close to perfectly to beat the best teams.

The good news is that I feel like our defense is humming. If that continues, it really lifts our floor.

[Rich]

That and foul trouble/injury are the big concerns. We lost yesterday in large part because all 3 perimeter guys struggled. If one of those guys is on and Carter and Bagley are effective, we should beat anyone. If those three guys are all off, we need Carter and Bagley to play close to perfectly to beat the best teams.

The good news is that I feel like our defense is humming. If that continues, it really lifts our floor.

But is the ceiling the roof?

[Troublemaker]

So, taking for granted that it was a different season, different team, different competition . . . the last Duke championship team/Final Four team had started the tournament ranked #57 on defense and then notably improved to the low 10's by the time they won the title.

Do you think this year's team needs to show a similar improvement to win the whole thing, or (based on several factors) we might be able to win it with a ranking in the 30's or 40's?

I know there is a general KenPom 'profile' of what a championship team looks like (sample size 2002-2017), but I'm wondering if you think this team can either (1) break the mold on what it takes to win it on the defensive side, or (2) make a run at a finding itself in the typical range for a championship team on defense?

The Easter Bunny, the Tooth Fairy, and Duke fans being able to celebrate a national title while saying, "Gosh we played such average defense over the past 6 games, but thank God our offense was so awesome."

We'll need to be top-15 on D by the end of the season, or no national championship. (That, of course, doesn't mean top-15 on D guarantees a national title).

We "made our move" towards the top of the D rankings early this season, starting on Jan 10th with the first game post-NCSU debacle.

In 2015, we "made our move" starting Feb 28th against Syracuse, the first game post-VaTech debacle. People mistakenly believe Duke only played great defense the last 6 games of the season, but we actually started about three weeks earlier. Here's my post in 2015 (Gonzaga postgame thread) explaining all this (http://forums.dukebasketballreport.com/forums/showthread.php?35712-MBB-Duke-66-Gonzaga-52-Postgame-Thread&p=796228#post796228).

Since we made our move so early this season, we probably need to enter the postseason with a top-25 defense (and then continue to climb), not top-60 like we were in 2015. Afterall, what would've slowed the ascent between today and the postseason? Would there be an injury? Did Coach K stop using his adjustments? Or did the adjustments stop working?

In any case, we are trending in the right direction. Let's hope the team stays healthy.

[Troublemaker]

In 2015, we "made our move" starting Feb 28th against Syracuse, the first game post-VaTech debacle. People mistakenly believe Duke only played great defense the last 6 games of the season, but we actually started about three weeks earlier. Here's my post in 2015 (Gonzaga postgame thread) explaining all this (http://forums.dukebasketballreport.com/forums/showthread.php?35712-MBB-Duke-66-Gonzaga-52-Postgame-Thread&p=796228#post796228).

BTW, reading the follow-up posts in that thread is a good idea, too.

[53n206]

UVa 5 turnovers: Duke 16 t.o.
Va shot 39% of 66 shots: duke 48% of 56 shots
Bagley led our 3 point shooters with 2

All add up to losing, esp. 16 turnovers.

[CDu]

UVa 5 turnovers: Duke 16 t.o.
Va shot 39% of 66 shots: duke 48% of 56 shots
Bagley led our 3 point shooters with 2

All add up to losing, esp. 16 turnovers.

The 16 turnovers were incredibly frustrating because so many of them were unforced. Way too many awful post entry passes by Carter and Trent, Duval just losing his dribble on a crossover in transition, the ill-advised outlet attempt into four defenders late, etc. If we protect the ball even just a bit better (say 12 turnovers) we probably win.

The good thing is those should be correctable. And we did play solid defense. Good enough defense to win if our offense was smarter with the ball. I ignore the poor shooting from the perimeter as both teams had bad shooting nights. The other good news is we shouldn’t face a team that good again unless we reach the ACC Final or the Final Four.

[duke4ever19]

The Easter Bunny, the Tooth Fairy, and Duke fans being able to celebrate a national title while saying, "Gosh we played such average defense over the past 6 games, but thank God our offense was so awesome."

We'll need to be top-15 on D by the end of the season, or no national championship. (That, of course, doesn't mean top-15 on D guarantees a national title).

We "made our move" towards the top of the D rankings early this season, starting on Jan 10th with the first game post-NCSU debacle.

In 2015, we "made our move" starting Feb 28th against Syracuse, the first game post-VaTech debacle. People mistakenly believe Duke only played great defense the last 6 games of the season, but we actually started about three weeks earlier. Here's my post in 2015 (Gonzaga postgame thread) explaining all this (http://forums.dukebasketballreport.com/forums/showthread.php?35712-MBB-Duke-66-Gonzaga-52-Postgame-Thread&p=796228#post796228).

Since we made our move so early this season, we probably need to enter the postseason with a top-25 defense (and then continue to climb), not top-60 like we were in 2015. Afterall, what would've slowed the ascent between today and the postseason? Would there be an injury? Did Coach K stop using his adjustments? Or did the adjustments stop working?

In any case, we are trending in the right direction. Let's hope the team stays healthy.

Agreed. We are currently sitting at #50 on KenPom.

[Kedsy]

Duke defensive stats vs. Notre Dame:

2-point%: .417
3-point%: .280
%threes: 41.0%
eFG%: 41.8%

TO%: 14.6%
DR%: 71.8%
FT Rate: 29.5%
dRating: 0.96

A/to: 1.3:1
%assisted: 59.1%
Block%: 9.8% (16.7% of twos)

Opposing fast break points: 6 (9.1% of ND's points)

Efficiency difference: -0.3 (opposing three efficiency approximately the same as opposing three efficiency).

Seventh straight game holding opponents to under one point per possession. Perhaps things to nitpick against a depleted opponent, but overall nothing particular to complain about.

[TheOldBattleship]

Opposing fast break points: 6 (9.1% of ND's points)

Defense is, of course, a combination of a ton of things that you have to do well all at once, but the way that we get back in transition and keep people from scoring easy points seems to me that it might well be the difference between winning and losing down the stretch for us. We just score at such a high rate that the numbers tilt towards us pretty drastically if we can keep the other team trying to go against our set defense in the half-court. Even if your half-court d is a bit below average in and of itself, you can really keep the other team's overall offensive efficiency numbers down if you just make them play in the half-court every single play, just because transition opportunities have such a generally high rate of conversion. Seems like that's been a major, major point of emphasis for this team so far, and it seems to have been working! It's incredibly impressive that this team, with such a heavy emphasis on crashing the offensive boards, has started to defend in transition the way they have the past few games. Really speaks to the discipline of the guards to get back in good order as soon as the shot goes up (particular props to Grayson here, who really has had to tamp down his overaggressive instincts in interest of keeping the team's defensive floor balance) and to the effort of the bigs turning and sprinting back after missing out on o-boards.

As we saw against UVA, there aren't any guarantees when you are playing truly elite teams. They didn't turn our turnovers into transition opportunities; instead, they incredibly smartly ran clock and kept the number of possessions we had as low as possible. But in general, I feel pretty good about our chances in pretty much any game if we get back and keep teams out of transition like we have been recently. Going forward, I'm excited to see us against a team that has a significant running game (oh, hello there, UNC...), just to see whether this trend holds up against someone who wants to push the ball.

[superdave]

Seventh straight game holding opponents to under one point per possession. Perhaps things to nitpick against a depleted opponent, but overall nothing particular to complain about.

Seven straight is pretty good. I take this as a very good trend after the State game.

I also liked the 18-0 run providing the knockout punch. This team is getting more matured.

[Kedsy]

Duke defensive stats vs. St. John's:

2-point%: .458
3-point%: .471
%threes: 26.2%
eFG%: 52.3%

TO%: 16.0%
DR%: 56.8 %
FT Rate: 27.7%
dRating: 1.18

A/to: 1.27:1
%assisted: 46.7%
Block%: 9.2% (12.5% of twos)

Opposing fast break points: (not in box score)

Efficiency difference: -24.9 (opposing three efficiency way better than opposing two efficiency).

Big step backwards defensively this afternoon. Defensive rebounding was atrocious and opposing three-point percentage was very high, though threes weren't a particularly high percentage of their shots (which I guess is a good thing). Overall, a disappointing day on the defensive end.

[superdave]

Duke defensive stats vs. St. John's:

2-point%: .458
3-point%: .471
%threes: 26.2%
eFG%: 52.3%

TO%: 16.0%
DR%: 56.8 %
FT Rate: 27.7%
dRating: 1.18

A/to: 1.27:1
%assisted: 46.7%
Block%: 9.2% (12.5% of twos)

Opposing fast break points: (not in box score)

Efficiency difference: -24.9 (opposing three efficiency way better than opposing two efficiency).

Big step backwards defensively this afternoon. Defensive rebounding was atrocious and opposing three-point percentage was very high, though threes weren't a particularly high percentage of their shots (which I guess is a good thing). Overall, a disappointing day on the defensive end.


You can count the number of really good defensive possessions on one hand, an even some of those ended in points for SJU. Definitely a hustle/energy problem yesterday.

[Kedsy]

Duke defensive stats vs. UNC:

2-point%: .432
3-point%: .333
%threes: 42.9%
eFG%: 46.1%

TO%: 3.1%
DR%: 57.4 %
FT Rate: 16.9%
dRating: 1.26

A/to: 7.5:1
%assisted: 50.0%
Block%: 6.5% (11.4% of twos)

Opposing fast break points: 10 (12.2% of UNC's points)

Efficiency difference: -6.8 (opposing three efficiency slightly better than opposing two efficiency).

Today's terrible defense courtesy of two stats: our second straight pathetic performance on the defensive boards, and our utter inability to force turnovers. I mean, just two UNC TOs, really? Oh, and we also gave up too many fast break points. But other than that, it wasn't so bad...

[Billy Dat]

I keep coming back to the fact that Duke, one of the teams who embraced small ball before most others, is, this season, in a position where teams are using it against us to great effect. I don't have the numbers to back it up, but the eye test tells me that the zone hasn't been great lately in terms of wide open 3s it is allowing and its impact on our rebounding. It also seems like Carter lacks the lateral quickness to effectively guard the perimeter which leaves us really vulnerable when we constantly switch. As I type this, however, I keep trying to find the thing that we DO well, and I can't. I want to say, "Let's stop switching", but that would likely just leave Carter or Bagley guarding a wing who will pull them away from the basket. It's maddening.

[flyingdutchdevil]

I keep coming back to the fact that Duke, one of the teams who embraced small ball before most others, is, this season, in a position where teams are using it against us to great effect. I don't have the numbers to back it up, but the eye test tells me that the zone hasn't been great lately in terms of wide open 3s it is allowing and its impact on our rebounding. It also seems like Carter lacks the lateral quickness to effectively guard the perimeter which leaves us really vulnerable when we constantly switch. As I type this, however, I keep trying to find the thing that we DO well, and I can't. I want to say, "Let's stop switching", but that would likely just leave Carter or Bagley guarding a wing who will pull them away from the basket. It's maddening.

So you must have loved the 3-big line up last night...

[El_Diablo]

I keep coming back to the fact that Duke, one of the teams who embraced small ball before most others, is, this season, in a position where teams are using it against us to great effect. I don't have the numbers to back it up, but the eye test tells me that the zone hasn't been great lately in terms of wide open 3s it is allowing and its impact on our rebounding. It also seems like Carter lacks the lateral quickness to effectively guard the perimeter which leaves us really vulnerable when we constantly switch. As I type this, however, I keep trying to find the thing that we DO well, and I can't. I want to say, "Let's stop switching", but that would likely just leave Carter or Bagley guarding a wing who will pull them away from the basket. It's maddening.

Yeah, we seem to break down in the 2-3 zone too easily (e.g., not bumping our defenders fast enough, not closing out on open shooters, giving easy passes to the free throw area). Even when it seems to work, it's not necessarily due to great defensive execution (e.g., the stretch in the UVA game where they just kept missing wide open 8-10 footers). It's probably too late to try to learn it now, but with the makeup of this team I would have liked to see some 1-3-1 zone (with Carter/Bolden in the center and Bagley/DeLaurier running the baseline). It opens up the corners but puts more pressure on perimeter passes (which our 2-3 zone is unable to do) while still helping us avoid getting gashed repeatedly on pick-and-rolls.

The easier solution though is to get all five guys rebounding on the defensive end. There were many times last night (and in other games) where it seemed 2-3 people were just turning around to watch, expecting Bagley and Carter to handle it.

[Kedsy]

Below is a list of Pomeroy's top 25 offenses. Purdue, Michigan State, and West Virginia are the only top 20 defenses in the bunch. Only 7 of the 25 teams are even top 40 on defense.

Is there something about good offense that somehow precludes good defense?



Team oRtg oRank dRtg dRank
Villanova 129.9 1 98.5 54
Duke 126.9 2 99.9 78
Saint Mary's 123.5 3 101.7 103
Purdue 123.4 4 94.5 13
Xavier 121.1 5 99.4 71
TCU 120.6 6 102.4 117
Auburn 120.2 7 97.3 37
Arizona 119.9 8 102.2 111
Michigan St. 119.7 9 93.7 10
Nevada 119.1 10 100.3 84
Gonzaga 119.1 11 96.2 25
North Carolina 119.1 12 96.6 30
Arizona St. 119 13 101.9 105
Oklahoma 118.8 14 100.6 89
Kansas 118.7 15 96 24
Marquette 118.5 16 104.3 156
Wichita St. 117.9 17 98.1 48
Creighton 117.6 18 99 64
Florida St. 117.6 19 98.7 60
Vermont 117 20 103.5 141
West Virginia 116.9 21 95.2 17
Arkansas 116.8 22 102.4 118
Virginia Tech 116.6 23 101.4 99
Davidson 116.6 24 103.5 142
UCLA 116.6 25 102.6 120

[kAzE]

Below is a list of Pomeroy's top 25 offenses. Purdue, Michigan State, and West Virginia are the only top 20 defenses in the bunch. Only 7 of the 25 teams are even top 40 on defense.

Is there something about good offense that somehow precludes good defense?

I'd need to see that list for at least 10 other years before asking that question. As it pertains to Duke, it seems to me that we recruit players who have great offensive talent, and take the approach of "we'll figure out defense later."

The one time we've had a freshman with elite defensive talent (Winslow, has there been another one in the OAD era?), we were okay on D. I'd say Battier, but that was when we still had NBA guys staying 4 years.

[CDu]

Below is a list of Pomeroy's top 25 offenses. Purdue, Michigan State, and West Virginia are the only top 20 defenses in the bunch. Only 7 of the 25 teams are even top 40 on defense.

Is there something about good offense that somehow precludes good defense?



Team oRtg oRank dRtg dRank
Villanova 129.9 1 98.5 54
Duke 126.9 2 99.9 78
Saint Mary's 123.5 3 101.7 103
Purdue 123.4 4 94.5 13
Xavier 121.1 5 99.4 71
TCU 120.6 6 102.4 117
Auburn 120.2 7 97.3 37
Arizona 119.9 8 102.2 111
Michigan St. 119.7 9 93.7 10
Nevada 119.1 10 100.3 84
Gonzaga 119.1 11 96.2 25
North Carolina 119.1 12 96.6 30
Arizona St. 119 13 101.9 105
Oklahoma 118.8 14 100.6 89
Kansas 118.7 15 96 24
Marquette 118.5 16 104.3 156
Wichita St. 117.9 17 98.1 48
Creighton 117.6 18 99 64
Florida St. 117.6 19 98.7 60
Vermont 117 20 103.5 141
West Virginia 116.9 21 95.2 17
Arkansas 116.8 22 102.4 118
Virginia Tech 116.6 23 101.4 99
Davidson 116.6 24 103.5 142
UCLA 116.6 25 102.6 120


Probably time. It's hard to get good on both ends of the floor with only so much practice time. Furthermore, the most talented players tend to be that way because of their offensive skills. This especially true in the case of a high-turnover one-and-done era. It doesn't seem all that surprising to me that the nation's best offensive teams aren't also the best defensive teams.

[Kedsy]

I'd need to see that list for at least 10 other years before asking that question. As it pertains to Duke, it seems to me that we recruit players who have great offensive talent, and take the approach of "we'll figure out defense later."

The one time we've had a freshman with elite defensive talent (Winslow, has there been another one in the OAD era?), we were okay on D. I'd say Battier, but that was when we still had NBA guys staying 4 years.

As I have noted multiple times, Duke's 2015 defense was #57 going into the Tournament, not significantly better than any other year in the OAD era. Also, I'd say Sulaimon was a very good defender as a freshman (and the 2013 team was also pretty good on defense, though with three senior starters, I'm not sure it's a particularly valid datapoint in this discussion). I think Ingram and Tatum were above-average defenders as well, though not as "elite" as Winslow.

As for another ten years of lists, Pomeroy's post-tournament data is easily available to you.

[Billy Dat]

Is there something about good offense that somehow precludes good defense?


I think it depends on how vigorous and complicated is the scoring celebration.

A three celebrated with the Melo "Three fingers to the dome" is a little more taxing and time consuming than holding up the "OK" three finger salute. A straight-on dunk might result in a stare down of the below basket camera or opposing fan section. A dunk from the wing might necessitate a primal scream. If ones boys are in the stands, a look and point might be called for. All of the above might slow down a sprint back on D.

Maybe that's why Saint Dean insisted on pointing to the teammate who assisted the basket, it was an action that necessitated running back on defense and looking at that end of the floor while en route.

[Kedsy]

Probably time. It's hard to get good on both ends of the floor with only so much practice time. Furthermore, the most talented players tend to be that way because of their offensive skills. This especially true in the case of a high-turnover one-and-done era. It doesn't seem all that surprising to me that the nation's best offensive teams aren't also the best defensive teams.

Makes sense. Also brings up the question, if you have to choose between being a good offensive/not-so-good defensive team or a good defensive/not-so-good offensive team, which would be preferable?

[CDu]

Makes sense. Also brings up the question, if you have to choose between being a good offensive/not-so-good defensive team or a good defensive/not-so-good offensive team, which would be preferable?

That is a great question. My first inclination is defense. But that is probably just because the most obvious recent examples of low-seeded teams winning it or getting to the Final Four were usually better defensive teams like UConn, Butler. I have no logical reason why a great defense would be more valuable than a great offense, so those examples could just be luck.

PS - thank you for saying “brings up” and not “begs”. (Stupid pet peeve of mine) ;)

[El_Diablo]

Is there something about good offense that somehow precludes good defense?

No. Look at literally any other year on kenpom.

[Kedsy]

No. Look at literally any other year on kenpom.

OK, I looked up last year. Only 6 of the top 25 were in the top 20 on offense (4 of the top 20). Though it's true that 15 of the top 25 were in the top 40.

[Kedsy]

Duke defensive stats vs. Georgia Tech:

2-point%: .385
3-point%: .259
%threes: 40.9%
eFG%: 38.6%

TO%: 14.1%
DR%: 65.2 %
FT Rate: 34.8%
dRating: 0.97

A/to: 1.4:1
%assisted: 63.6%
Block%: 10.6% (17.9% of twos)

Opposing fast break points: 8 (11.6% of UNC's points)

Efficiency difference: -0.4 (opposing three efficiency virtually identical to opposing two efficiency).

Defensive rebounding was better than the last two games, but still not good enough. Turnovers back to our average, though still bad. Fast break points still a little high, but not totally terrible. We let them get to the line a lot. But we held Georgia Tech to very bad shooting, so overall the defense was adequate but not great. Baby steps, I guess.

Javin DeLaurier played the most minutes he's had this season since November 14. And I liked what he brought to the defensive end. Statistically, he got a block, two steals, and four defensive rebounds, but more than that he was all over the court, tipping errant passes and diving for loose balls. If only he didn't foul so much, he'd be a defensive force.

[Nugget]

Duke defensive stats vs. Georgia Tech:

2-point%: .385
3-point%: .259
%threes: 40.9%
eFG%: 38.6%

TO%: 14.1%
DR%: 65.2 %
FT Rate: 34.8%
dRating: 0.97

A/to: 1.4:1
%assisted: 63.6%
Block%: 10.6% (17.9% of twos)

Opposing fast break points: 8 (11.6% of UNC's points)

Efficiency difference: -0.4 (opposing three efficiency virtually identical to opposing two efficiency).

Defensive rebounding was better than the last two games, but still not good enough. Turnovers back to our average, though still bad. Fast break points still a little high, but not totally terrible. We let them get to the line a lot. But we held Georgia Tech to very bad shooting, so overall the defense was adequate but not great. Baby steps, I guess.

Javin DeLaurier played the most minutes he's had this season since November 14. And I liked what he brought to the defensive end. Statistically, he got a block, two steals, and four defensive rebounds, but more than that he was all over the court, tipping errant passes and diving for loose balls. If only he didn't foul so much, he'd be a defensive force.

To my eye, Javin also blatantly missed at least 3 block outs to give up ORs.

[Troublemaker]

Kedsy, where do you go to get the fastbreak points? (It may be a very obvious source that I'm overlooking).

Thanks in advance.

[Kedsy]

Kedsy, where do you go to get the fastbreak points? (It may be a very obvious source that I'm overlooking).

Thanks in advance.

It's at the bottom of the official (GoDuke) box score. It's not there sometimes for non-conference road games (e.g., St. John's), but it's been there for almost all games.

[Kedsy]

To my eye, Javin also blatantly missed at least 3 block outs to give up ORs.

I didn't notice that, but it's possible. Still, he had the most defensive rebounds per minute on the team last night.

[CDu]

To my eye, Javin also blatantly missed at least 3 block outs to give up ORs.

Yep. He definitely missed some box-out opportunities during that second-half run. Not that he was the only one (Bolden missed at least one glaring one too), but there were definitely a few cases of it by DeLaurier to go along with a few over-exuberant fouls.

DeLaurier is currently a high-energy, low-control guy. A few times he ran into other Duke players as both went for a loose ball, causing neither to get it. He's got a lot of physical tools, but he's still very much working to figure out how to use them for good. He will get a fair amount of nice hustle plays, but he is also adding a fair amount of mistakes because he doesn't always "hustle smartly".

It's got to be a tough balancing act, trying to provide energy while also trying to play smart. And when you don't play a lot, it can be especially tough. Hopefully he continues to get more and more comfortable with more time. But right now, he's still really rough around the edges as he tries to figure out what to do out there.

[HereBeforeCoachK]

To my eye, Javin also blatantly missed at least 3 block outs to give up ORs.

My observation too, and he was definitely AWOL on that follow slam by that Coles guy off the bench, which was the highlight moment for Tech all night (and it was a great play by Coles) - but Javin let that happen.

On the other hand, he has moments of incredible quickness and athleticism.......

[HereBeforeCoachK]

Below is a list of Pomeroy's top 25 offenses. Purdue, Michigan State, and West Virginia are the only top 20 defenses in the bunch. Only 7 of the 25 teams are even top 40 on defense.

Is there something about good offense that somehow precludes good defense?


In a way, yes indeed. Rare is the team that is gifted offensively and defensively...and quite often good offensive teams spend more time working on O than on D. Now that second part is not the case with Duke I'd assume...I think their problem is trying to force a square peg into a round hole - ie - they are not a man to man team. It's not who they are, and not who they can become. The most gratifying thing about last night was all the zone...and they're not a great zone team, but they're a very good zone team at times, and if they'd just figure out that's who they are, I think their O and their D would improve, and they'd become the best O team in the country.

[CDu]

My observation too, and he was definitely AWOL on that follow slam by that Coles guy off the bench, which was the highlight moment for Tech all night (and it was a great play by Coles) - but Javin let that happen.

On the other hand, he has moments of incredible quickness and athleticism....

He is definitely a player whose bball awareness hasn't caught up (or even remotely close to caught up) to his energy, hustle, and athleticism. A very raw player still.

If the game awareness catches up to the physical tools, though, look out. Game-changing combination of size, athleticism, and "want-to."

[superdave]

dRating: 0.97


Kedsy - you wouldnt happen to have 1st half rating vs 2nd half rating from last night, would you?

We gave up 26 points in the 1st and 43 in the 2nd. It seemed like they went on cruise control a good bit then realized GT was playing hard.

[Kedsy]

Kedsy - you wouldnt happen to have 1st half rating vs 2nd half rating from last night, would you?

We gave up 26 points in the 1st and 43 in the 2nd. It seemed like they went on cruise control a good bit then realized GT was playing hard.

FIRST HALF
dRating (points per possession): 0.72
turnover percentage: 19.3%
defensive rebounding pct: 81.8%

SECOND HALF
dRating: 1.24
TO%: 8.7%
DR%: 50.0%

So, first half was awesome. Second half was basically St. John's/UNC.

[MrPoon]

FIRST HALF
dRating (points per possession): 0.72
turnover percentage: 19.3%
defensive rebounding pct: 81.8%

SECOND HALF
dRating: 1.24
TO%: 8.7%
DR%: 50.0%

So, first half was awesome. Second half was basically St. John's/UNC.

Even the second half was interesting. First four mins, 1pt against, remaining 16 mins, 42pts. Probably splitting hairs but man was that stretch ugly against a team without its primary PG.

[Kedsy]

Duke defensive stats vs. Virginia Tech:

2-point%: .483
3-point%: .350
%threes: 40.8%
eFG%: 50.0%

TO%: 17.3%
DR%: 89.3 %
FT Rate: 8.2%
dRating: 0.90

A/to: 1.3:1
%assisted: 61.9%
Block%: 8.2% (13.8% of twos)

Opposing fast break points: 9 (17.3% of VaT's points)

Efficiency difference: -4.2 (opposing three efficiency slightly better than opposing two efficiency).

This was an odd one, statistically, for Duke this season. Perhaps because we played zone almost the whole game. Slow pace. High percentage of opponent's made two-point shots (even with three or four missed bunnies early). Forced turnovers a bit better than usual (at least in the second half). Allowed too many fast break points. BUT reversing a recent trend, our defensive rebounding was awesome. And our free throw rate, usually good, was outstanding. And those two things gave us a good overall performance.

[budwom]

Duke plays zone and VT gets all of three offensive rebounds, that's pretty amazing.

[Matches]

Allowed too many fast break points.

Weren't all 9 of those FB points allowed in the first 10 minutes? I think I remember a graphic to that effect when we'd turned it over 7 times before the under-12 timeout. If so the team did an excellent job of course-correcting.

[Kedsy]

Weren't all 9 of those FB points allowed in the first 10 minutes? I think I remember a graphic to that effect when we'd turned it over 7 times before the under-12 timeout. If so the team did an excellent job of course-correcting.

Could be. Seven of the nine VaT fast break points were in the first half, for sure. It's very possible they were all during our early TO-fest.

[El_Diablo]

Could be. Seven of the nine VaT fast break points were in the first half, for sure. It's very possible they were all during our early TO-fest.

Their last score was a one-man runout dunk (with O'Connell trailing)...not sure if that was officially scored as a fast break but it seemed like one.

[pfrduke]

Their last score was a one-man runout dunk (with O'Connell trailing)...not sure if that was officially scored as a fast break but it seemed like one.

It was. Their fast break scores were in the first half at 15:59 (Bibbs 3); 12:47 (Robinson tip-in); and 11:34 (Hill layup) - all of which came off of steals (all by Alexander-Walker, as it happens) - and then in the second half at 0:11 after a missed shot. So there were 31 minutes straight where they didn't score on the break.

[Neals384]

Seem like the appropriate place to quote this off the DBR front page:

“We will play a lot more zone. That will be our primary defense." -Coach K.

[uh_no]

So i was wondering to what degree duke performs worse when our opponents offense improves. In recent KP theory, if an opponent is one PP100 better on offense, they should score ABOUT one PP100 more. So if you took a look at all the opponents duke played, their intrinsic offensive efficiencies, and their measured efficiencies for a given game, you'd expect a slope of about 1. That's not what we see from duke, though:

8078

we see that for every 1 PP100 a team is better, they score on average 1.4 more points in 100 possessions against duke.

Caveats and Conclusions:
1) duke's composition will beat up on crappy offensive teams, but struggles disproportionately against good ones
2) you will almost assuredly have to beat good offensive teams in the tournament
3) I'm not going to calculate this from every team, so while we can't say conclusively whether this is better or worse than anyone else, it aligns with the eye test
4) we have some good performances against good offensive teams....like last night. like UVA. These are countered by several especially atrocious defensive performances against good offensive teams
5) home court advantage you would expect to even out in the long run, so I ignored it here. Even if you make the claim that it disproportionately affects bad teams since we play many cupcakes at home early in the year, I'd counter that the trend is pretty clear across the domain
6) I believe KP switched to a strictly arithmetic method for adjusting per team (o+d-average) from a geometric one (o*d/average). In the latter method, you WOULD expect slightly more than a slope of 1, but only in cases when your defense was really bad or really good. Given that duke's defense is rather middling (100 with a current average of about 105), the adjustment in either case should be almost perfectly linear, and not anywhere close to the 40% we observe