We keep talking about how this year's team has the worst defensive efficiency in the Pomeroy era, so I thought it might be worthwhile to break it down and analyze the pieces.
According to Pomeroy, there are four factors that go into defensive efficiency: defensive rebounding percentage (what percentage of available defensive rebounds do you get); free throw rate (how often do you send your opponent to the free throw line); turnover percentage (what percentage of your opponents possessions end up in a turnover; and effective field goal percentage (shooting percentage counting made three-point shots as 1.5 and made two-point shots as 1.0).
Here's how this year's team stands up to the past dozen Duke teams:
Our free throw rate is actually the 5th best of the 12 years, and our defensive rebound pct is the 3rd best. But the other two factors are the worst figures we've managed in the past dozen years. Having said that, our turnover pct is not all that different from several of the years, including last season and the 2010 national champs. But I think our defensive eFG% is historically bad (48%). Only three years I charted were even close to as bad (2009, 2008, and 2003), and in all of those years we had much better turnover rates.Code:Year DReb% ft rate to% eFG% 2012 65.7 30.1 20.0 48.0 2011 64.9 29.6 21.0 44.5 2010 59.4 34.0 21.4 43.6 2009 62.6 31.0 23.5 47.8 2008 66.0 31.9 24.7 47.5 2007 63.5 29.5 22.1 46.0 2006 69.1 27.6 22.5 46.1 2005 62.7 32.0 21.8 42.2 2004 60.7 31.9 24.4 44.7 2003 63.0 37.6 24.4 47.5 2002 65.5 32.1 25.6 46.0 2001 62.9 28.3 24.9 45.7
So I broke it down further, looking at our D against threes, twos, and total overall. The last column in the table below is what percentage of our opponents' shots were three-pointers.
Our three-point defense seems OK. Six of the past 16 Duke teams were worse and two others were practically identical, so we're middle of the pack for a Duke team in defending the three. So far, so good. Except opponents try fewer threes against us this year than all but three of the past 16 seasons (2006, 2005, and 1997), which suggests the key lies in two-point shots. Where this year's team is tied for worst among the past 16 years in stopping the two-pointer.Code:Year 3 made 3 att 3 pct 2 made 2 att 2 pct totmade tot att tot pct % threes 2012 88 269 0.327 405 849 0.477 493 1118 0.441 0.317 2011 176 543 0.324 720 1667 0.432 896 2210 0.405 0.326 2010 158 559 0.283 725 1643 0.441 883 2202 0.401 0.340 2009 183 542 0.338 722 1544 0.468 905 2086 0.434 0.351 2008 167 507 0.329 713 1518 0.470 880 2025 0.435 0.334 2007 139 441 0.315 626 1371 0.457 765 1812 0.422 0.322 2006 143 471 0.304 809 1744 0.464 952 2215 0.430 0.270 2005 128 420 0.305 655 1586 0.413 783 2006 0.390 0.265 2004 176 543 0.324 702 1614 0.435 878 2157 0.407 0.336 2003 164 474 0.346 675 1415 0.477 839 1889 0.444 0.335 2002 168 555 0.303 724 1559 0.464 892 2114 0.422 0.356 2001 207 602 0.344 821 1869 0.439 1028 2471 0.416 0.322 2000 197 551 0.358 737 1688 0.437 934 2239 0.417 0.326 1999 191 635 0.301 781 1849 0.422 972 2484 0.391 0.343 1998 156 512 0.305 685 1532 0.447 841 2044 0.411 0.334 1997 149 438 0.340 640 1433 0.447 789 1871 0.422 0.306
I assume this means we're giving up a lot more dunks and layups than usual. Which I suppose I should have been able to guess without all the charts. Combine that with the low turnover rate, and it spells bad D, like we saw in the 2nd half against Florida State today.
How can we fix it? I don't know. Perhaps we should pack it in more, clog up the middle and force people to take more threes with people running at them, sort of like we saw from UVa. Maybe we need to take more chances and force more turnovers. Or possibly we just need to get better at rotating and/or stopping the penetration of opposing PGs.
Frankly, I was hoping for more illumination from the numbers, but you can't force the data. Since I did the work, I'm posting the charts, and maybe someone else will have a better idea.
Good stuff Kedsy. Thanks.
Good to see the eyeballs validated that it's our 2 point D as opposed to 3 point D, rebounding, etc. that is the problem. But then the question becomes: what aspect of our 2 point D is the real problem, because most shots taken are 2 pointers. Is it failure to stop dribble penetration, is it failure to recover/close onto shooters by our perimeter guys, poor play against screen/roll, is it failure by our bigs to stop post moves? My eyeballs tell me that it really has been a combination of things, different problems on different nights, different guys having better and worse nights, without a whole lot of rhyme or reason to it. I bet that is a source of frustration for the staff if they perceive what I do.
To me, our inability to stay in front on the perimeter has probably been a more glaring problem than some of the others, but I can't say that it has been the direct result of more baskets surrendered than any other particular problem. The numbers I have compiled in the charting threads shed some light, but they, like the Pomeroy numbers (though much less sophisticated than KenPom obviously) don't tell the whole story.
The best way to tell the real story is to go play-by-play and attribute each hoop to a particular problem and to a particular player(s) who caused the problem. I did that for our game earlier this year vs. Washington, and there is a thread on that, but it would need to be done over a series of games, or even better, the whole year, and it's just too time-consuming for anyone unless it would be their paying job to break it down like that. I guess that's why teams hire video coordinators -- that's what it takes to really drill down into this stuff. If I have time I may do it for this game, since everyone is so upset about the D we played, and I'm curious myself.
It's really cool. We were OK at stopping twos, but not that great, and we more or less sucked at every other component of defensive efficiency -- except we were amazing at just one thing. And that one thing made our defense good enough to win the national championship.
So maybe there's hope for this year's team. Let's get amazing at one thing, and perhaps the rest will follow...
I think it confirms what we see with the eye test: too many lay-ups, and no turnovers. It's not like teams ever really hit threes against us, so to have an eFG% that high means we're really doing a bad job of giving up high-percentage shots inside the arc.
The decent FT rate and rebounding percentage are somewhat of a testament to our big guys not being a major part of the problem, I think. It all boils down to our inability to effectively pressure the ball and keep guards out of the lane, and our lack of length to bother shots once they get there. Bad rotations and ball-screen defense aren't helping things either.
And to add another point to what Kedsy said about the 2010 team being really good at one aspect and weak elsewhere and how that can be effective. It wasn't the only championship team to play like that. As a disclaimer, the era is obviously quite different and teams played quite differently, but the following is a nice illustration.
The team with by far the worst eFG% allowed for a Duke team since the advent of the 3-point line is actually the 1991-92 team. And they weren't great at defensive rebounding or forcing turnovers either but had an extremely low free throw rate. I can only guess by browsing through the stats of the other top teams for context, but this probably wasn't a top 10 defensive squad like the 2010 team, but it obviously did enough.
(Although that comparison only goes so far, since the '92 squad is probably one of the greatest offenses in NCAA history. They had by far the best offensive eFG% for any Duke team and it might the best of any NCAA team in the last 25 years.)
Good summary, not sure if this fits the stats and def. eff., but I view it as a team defense question with the problem often being the second rotation, if not the first. And I think this has a lot to do with playing three guards and two bigs. A defender like Singler can cover and provide a great deal of help side d.
Last edited by tele; 01-22-2012 at 07:33 AM.
Nice post breaking down what many have concluded without the hard data.
I think several people here have thought that this team isn't quick/athletic enough to stay in front of the ball handler with extended man-to-man pressure defense. It seems to me that the only possible solution (or at least the most logical one) is to gravitate back to our 2010 style of defense (i.e. more of a packed in and contain).
Also if we are going to do this it might make the most sense to start both Plumlees and bring Kelly off of the bench b/c they are better rebounders and can take up more space in the paint that Ryan can.
The morning after, I'm struck with a few observations that fit well with the analysis and comments in this thread:
Florida State is a physically mature team with big, long, strong, quick athletes
This year's Duke team does not have that caliber of athletes in abundance
We are especially lacking in not having a versatile, athletic player anywhere between 6'4" and 6'10"
Once an opposing team like FSU can penetrate our perimeter they overmatch our interior defense with a drive or good passing resulting in high quality shots, often lay-ups or dunks
Don't have any ready prescription for this type of defensive weakness; we have to play with the players we have.
Would certainly help to have shot a little better ourselves and win the game by just out-scoring the Seminoles.
I'm afraid our flaws on defense could lead to a real ceiling as to what this team can achieve.
We'll see what Coach K and the staff can do to adjust and improve.
Everyone says we don't have the athletes to be an elite team. I still contend it's the chemistry and lack of leadership. We don't talk ( communicate) with each other on the court like teams in the past. The help defense suffers because of this lack of playing as a "fist" (K's term). I would gladly welcome a new recruit like say some kid from Las Vegas coming in and developIng into our on court leader because no one seems to be accepting the challenge currently.
Another factor: what role, if any, does our offense play into all? If a Duke player has an ill-advised drive, gets knocked to the floor, and the other team has a 5 on 4 run-out that results in a layup, it's not necessarily a problem of us being unable to stop penetration, but rather a problem of offensive shot selection, hustling back, what have you. Maybe these happen so infrequently so as to not affect the numbers. Still, I'm curious as to how offensive productivity (we make the 2d FT; we make the FG) might affect the defensive end.
A this point I'd try sagging just below the arc by the guards, with the bigs packing it in. Shouldn't be as much of an adjustment since most of our guys did it in 2010.
OK, I made a slight mistake. I put opposing Defensive rebound percentage in the table rather than our defensive rebounding percentage.
Here's the correct table:
It actually doesn't change much. This year's team actually has the 2nd best defensive rebounding percentage of the twelve years (rather than 3rd as I reported above). Although the best Duke defensive rebounding team of the past dozen years was the 2007 team, which gave me a little shiver.Code:Year DReb% ft rate to % eFG% 2012 68.5 30.1 20.0 48.0 2011 66.8 29.6 21.0 44.5 2010 67.5 34.0 21.4 43.6 2009 66.8 31.0 23.5 47.8 2008 66.2 40.6 17.1 47.5 2007 69.6 31.9 24.7 46.0 2006 62.2 27.6 22.5 46.1 2005 63.2 32.0 21.8 42.2 2004 62.9 31.9 24.4 44.7 2003 65.1 37.6 24.4 47.5 2002 65.9 32.1 25.6 46.0 2001 63.8 28.3 24.9 45.7
The only substantive difference caused by my mistake is my analysis of the 2010 team. They were tied for 3rd best defensive rebounding pct, rather than the worst. Their free throw rate and turnover pct were both still 2nd worst, though, so for the most part my observation about that team still holds.
Sorry about the mixup.
not enough big players that K thinks are ready to play.
We have Mason and Miles as bigs, 10 fouls. I see Ryan as inside out, as a big 5 more fouls.
So against FSU we had to be careful not to get fouls inside in the second half and it hurt the defense.
This problem will continue unless he plays Josh or Michael more.
In 2010 we had 4 bigs, Brian, Lance and the Plumless. Kyle could and did play inside, as did Ryan some. 6 players, 30 fouls.
The guard heights of the primary contenders in the ACC:
UNC (now w/o Strickland): 6' 4", 6' 6", 6' 8"
Fla St: 6' 5", 6' 5", 6' 5"
NC State: 6' 5", 6' 5", 6' 6"
Virginia: 5' 11", 6'1, 6' 6"
The only team with shorter guards (UVa) we managed to eke out a victory at home against and handled their guards fairly well. It was Scott who wore us down a bit.
We are going to figure out how we handle bigger guards in conference long before we are 'exposed' in the tourney one way or the other, as the games most meaningful to creating separation in the conference race all involve big guards for the rest of the year.
In fact UNC may well have one of the tallest starting five's in college basketball history. At least they aren't also loaded with quick penetrators which is a relief.
This teams greatest strength thus far is the athleticism of the brothers Plum. They have masked a lot of our guard defensive liabilities as they have controlled the interior fairly well even though they are often having to rotate and cover incessantly due to our need to play up on opposing guards and having them dribble by into the lane. At times it's not enough as it was way to easy for both Fla. State and Virginia to get the ball deep in the paint for their best scorers (Scott and James) in the second half, once they realized their advantage there.
Man if Mason leaves next year we are really going to see how we can operate without those two cleaning the glass, creating defensive headaches, and getting up and down the court at amazing speed for their size.
Let the tinkering continue...!
PS - Great stat work Kedsy, good to have objective measures even if they aren't conclusive.