Browsed by
Tag: UNC team analysis

Carolina-Kentucky: Round 1 Recap

Carolina-Kentucky: Round 1 Recap

Due to life getting in the way, I haven’t had much time to blog recently.

But here are some pieces I wrote following the first matchup:

1.) Defending Monk

2.) Defending Ball Screens (Fox)

3.) Transition D and Late-Game Execution

There obviously no Kenny Williams this time around (who did the lion’s share of the work against Malik Monk in December). And, of course, Theo Pinson has returned to take Williams’ spot as the starting 2 (and likely assignment on Monk). It’s almost comically over-simplistic, but Carolina’s ability to slow down Monk and De’Aaron Fox (and limit their efficiency if that their scoring) will likely determine Sunday’s result. The Heels will need to do a better job against the ball screen than they did in Las Vegas last December.


Wrapping Up the Duke Win

Wrapping Up the Duke Win

While most Carolina fans have probably moved past the Duke game and onto the ACC Tournament (or March Madness/bracketology concerns), I still have a few loose ends to tie up regarding that big win to conclude the regular season. So let’s start tying:

Efficiency by Possession Length

Let’s break the offensive efficiencies for Carolina and Duke down by length of possession:

  • Like usual, Carolina was more prolific in the early offense. The Heels used 46% of their total possessions in the first 10 seconds, compared to just 34% for Duke. Those numbers were strikingly similar to the first matchup, when UNC used 42% of its possessions early vs. 34% for the Devils.
  • Like in the first game (Duke: 126.1, UNC: 121.4), Duke had a slight edge in early-offense efficiency. Carolina’s transition defense tightened up a bit in the second half, however. After allowing 19 early-offense points in 14 first-half possessions, the Heels only allowed 12 in 10 second-half ones.
  • As it’s been doing for much of the season, the Heels won this game with its half-court offensive efficiency. Although Carolina had seven fewer half-court possessions than Duke, it was only outscored 52-48 on trips that lasted longer than 10 seconds. Most of UNC’s big hoops down the stretch (Jackson 3: 22 seconds, Maye layup: 16 seconds, three late Berry hoops: 23, 22, and 22 seconds) occurred in the final half of the shot clock.
  • Limiting Duke’s transition and second-chance 3-pointers was a big key to the win. The Devils got a late transition 3 (Kennard-to-Allen in the corner), but Carolina limited the early-offense kick-out 3s (following offensive rebounds) that can be so deadly against Duke.
  • Like in the previous game against Virginia, many of Carolina’s favorite secondary break options were limited against Duke due to familiarity and good preparation. Well-coached ACC teams will usually do a good job of taking away the back screens, slips, and easy entries that create early offense through UNC’s secondary break. Berry, however, was aggressive in probing the defense early, and able to get some quick baskets off the bounce. In general, the Heels went to more secondary break ball screens and dribble hand-off action to create early offense.
  • Another secondary break wrinkle that I loved seeing was the post-up for Justin Jackson. After cutting backdoor against an overplaying Matt Jones (denying the Meeks secondary reversal pass), Jackson immediately looked to post up the smaller defender. Rather than kicking back to Seventh Woods on the left wing, Meeks instead took a couple of dribbles to the right wing to set up an entry angle to Jackson. This type of secondary option is a great way to get early touches for Jackson on the block against a smaller defender (with the backdoor option available for keeping overplaying defenses honest). Not sure if this was a Duke-specific secondary set that the staff implemented, or if it was just Jackson and Meeks making a play in the moment.


Pinson as a Passer

Despite not scoring for the second game in a row, Theo Pinson’s offensive impact was still profound. Unlike the Virginia game, when the Heels’ offense sputtered (and more Pinson off the dribble may have been an option worth pursuing), it’s hard to argue with 90 points on 72 possessions.  Pinson was clearly Carolina’s top play-maker against Duke, leading the team in assists (7), potential assists (12), and hockey assists (3). His 12 potential assists set up the following shots for the Heels:

  1. A made Hicks layup in secondary after a Pinson post entry
  2. A made Jackson primary break dunk
  3. A missed Meeks face-up jumper after a Pinson post entry
  4. A made catch-and-shoot Maye jumper from just inside the foul line
  5. A missed Maye 3-pointer
  6. A made Jackson lefty layup
  7. A missed Meeks jump hook after a Pinson post entry
  8. A made Maye leaner
  9. A made Hicks layup in secondary after a Pinson post entry
  10. A missed Britt 3-pointer
  11. A missed Jackson 3-pointer
  12. A made Jackson primary break layup (against late Duke pressure)

Pinson also threw two entry passes to Hicks that resulted in unassisted layups (since Hicks backed down smaller defenders off the dribble). Likewise, two of his entry passes turned into hockey assists after inside-out 3-pointers. For the game, Pinson threw a team-high nine post entry passes. Those passes resulted in six made baskets, two missed baskets, and a turnover. The clip below shows one of the Pinson post entries to Hicks that didn’t result in an assist. Still, Pinson’s ability to recognize the mismatch (Kennard on Hicks) and get the ball to the right spot is one reason why his presence is so important to UNC. On this play, Berry failed to take advantage of either of his mismatches (taking Jefferson off the dribble, or feeding Hicks in the post). Pinson, however, immediately capitalized on Carolina’s advantage—something he’s been excellent at doing all season. None of Pinson’s assists against Duke were super-flashy. And his two turnovers could have easily been more (he had some questionable passes that were deflected, but not outright stolen). Still, his ability to seamlessly mix in the simple pass with the high-risk/high-reward one has paid great dividends for the UNC offense, while allowing Berry to focus on his strengths (scoring, perimeter ball movement) rather than his weaknesses (feeding the post).


Carolina’s Defense / Guarding the 3

In the first matchup against Duke, Carolina’s defense allowed 13 made 3-pointers on 27 attempts. Of those attempts, I charted 24 to be open or lightly contested (including all 13 of the makes). So how much did UNC improve its perimeter defense in the second Duke game? Of the 19 Duke 3-pointers (already a big improvement!), only 12 were categorized as open/lightly contested. Duke went 5-of-12 on its clean 3s, a little less lucky than the 13-of-24 it shot on them in Cameron. So, while better shot luck also played a role, the bigger factor was the Heels’ ability to cut the rate of clean 3-point looks in half relative to the first Duke game.

That reduction in clean 3-point looks came at a cost, however. Since Carolina was more committed to sticking to shooters, it put a bigger onus on its on-ball and interior help defenders to stop dribble penetration (rather than using its wings to help early). Duke relentlessly attacked off the bounce, drawing a ton of free throws (usually against UNC’s primary defender, but sometimes against late-helping bigs). It more than doubled its free throw rate from 28.1 in the first matchup to 64.8 this time around. Using the wings to help early against the drive sets up the drive-and-kicks that Duke’s 3-point attack thrives on. But not helping with the wings will put a ton of pressure on ball defenders to curtail wing penetration without fouling (especially difficult against Duke since they’re aggressive, talented, and coached to draw/exaggerate contact). There’s not necessarily a right and wrong way to defend Duke—UNC allowed 1.27 PPP in the first game and 1.19 in the second (both pretty bad). It’s really just a matter of trade-offs: what are you hoping to take away, and what are you willing to live with? Duke’s good enough offensively t0 take advantage of what you give it (penetration/drawing fouls/finishing in the paint last Saturday). That said, I thought the defensive adjustment to take away 3s was the appropriate one. While it’s annoying to watch your rival parade to the line 35 times, that strategy did take away most of the back-breaking, momentum-generating 3s that the Duke offense has historically feasted on. With a different crew of referees (and/or some better UNC defensive discipline), that strategy could have been even more effective.

One play (late in the game) in which the Heels did help early from the wing is highlighted below. Berry starts the defensive possession with good ball pressure to blow up a Duke ball screen. He then makes an excellent help-and-recovery close-out to content a Frank  Jackson 3. This was more of a fake-and-retreat move, as Berry was already recovering back to Jackson before Tatum even released the kick-out pass. If Berry was longer (like Jackie Manuel, Danny Green, or Theo Pinson), this type of play would be even more effective. It represents the ultimate form of defense against Duke: help early from the wings to prevent deep penetration/fouls, but still recover to shooters in time to contest the 3. It takes a perfectly timed help-and-recover (plus some combination of length/lateral quickness) and, of course, introduces the possibility of overhelping. These help decisions are really hard to make in real time, but Berry did a great job on this late possession.



OK, on to Brooklyn!

Closing Out a Championship

Closing Out a Championship

Throughout the season, we’ve spent countless words detailing Carolina’s late-game execution in crunch-time situations. Oftentimes this year (including the first Duke game), the Heels have been out-executed down the stretch. Even in some wins (like at Clemson), the Heels’ late play (and decision-making) left plenty to be desired. But Saturday night, on the season’s biggest stage, the UNC close-and-late performance was top-notch. And, sometimes, it’s less about execution and more about just stepping up and making winning plays.

We’ll start this close-and-late breakdown as close as it can get: tied at 71 with 6:15 on the clock following a Luke Kennard tip-in. On the court for the Heels was the (regular) starting five–a unit that played 7:03 of the final 9:07, leading 19-10 over that period (the Heels trailed 8-4 in the 2:04 without all five starters down the stretch).

UNC1 (74-71): After running some active (but somewhat frantic and unfocused) freelance motion, UNC dialed up dribble hand-off action between Isaiah Hicks and Justin Jackson with about 10 seconds left on the shot clock. Duke’s Frank Jackson, an explosive freshman scorer, made the cardinal sin of going under the exchange, leaving his namesake free for a clean 3-point look from the top of the key. The Heels’ Jackson, who had missed his first six 3-pointers (including two from the top of the key, a spot where he’s shooting just 36.7% (18-49) this season), calmly drilled this clutch opportunity to put the Heels back in the lead.

DU1 (74-71): Using a ball screen from Amile Jefferson, Jayson Tatum, Duke’s stretch 4, attacked off the dribble (as he was looking to do all game). Unusually, the Heels opted not to switch this big-big exchange, with Kennedy Meeks flat-hedging while Hicks worked hard to recover to a driving Tatum. Hicks recovered just enough to force Tatum to fade left while attempting to finish to his right, with Jackson arriving late to get a help-side hand up. Make no mistake, though, this was a pretty good (and close) look for Tatum, who certainly didn’t have his best night as a finisher (some due primarily to good Carolina defense, some not so much). Meeks corralled the defensive rebound.

UNC2 (77-71): In its secondary break, Carolina went immediately back to the identical dribble hand-off action from the previous possession between Jackson and Hicks. This time, Frank Jackson was conscientious to fight over the top of the exchange. A solid Hicks screen, however, allowed Justin a driving lane as Frank frantically tried to recover. Jefferson, stuck a bit in no man’s land, was forced to step up to stop Jackson’s drive, allowing Carolina’s Player of the Year candidate to thread a perfectly delivered pocket pass to a rolling Hicks. Hicks, guarded by an overmatched Frank Jackson now, easily finished at the rim while drawing the foul. It was his 13th “and-1” of the season (and second of the game), tying Meeks for the most on the team. He knocked down the free throw, the 10th time he’s completed a 3-point-play opportunity.

DU2 (77-73): After Nate Britt and Luke Maye checked in for Theo Pinson and Hicks, Duke immediately exploited the Britt-Kennard matchup (with Britt giving up several inches and dozens of pounds). It used a floppy set to curl Kennard off a right-block screen. He then rolled in a floater over Britt in the paint. Britt defended it well positionally; he simply wasn’t big enough to adequately stop it.

UNC3 (79-73): Carolina went back to its freelance passing game, again using Jackson in a two-man game. This time, it was Maye who set a ball screen for Jackson on the right wing. Duke, who had been switching all exchanges all game, actually hedged and recovered here (with Frank Jackson on the ball and Jefferson as the hedger). Jefferson was late to get back to a rolling Maye, and Jackson was able to hit him for an open layup.

DU3 (79-75): Also sticking to what was working, Duke again ran Kennard off a curl out of floppy (this time using a left-block screen). After receiving the ball in the paint, he took Britt to the hoop to draw a foul (then knocked down both free throws).

UNC4 (79-75): For the fourth straight trip, the Heels utilized Jackson in two-man action. On this possession, it was Meeks setting a ball screen for Jackson on the left wing. He drove baseline then, upon being cut off by the Duke help defense, threaded a wrap-around bounce pass to Joel Berry in the opposite short corner. It was a clean look for Berry, though he was a bit off-balance (and not really ready to catch and shoot, a bit surprised by the odd angle of the Jackson pass, perhaps) which led to a missed 16-footer. Although

DU4 (79-78): With Jackson along one baseline (after his pass) and Berry along the other (after his shot), Duke was able to grab the defensive board and push tempo (following a strong Tatum outlet pass to Kennard). Britt was left by himself in transition to stop Kennard and Allen and, given Duke’s trademark floor spacing, was unable to prevent Allen’s 3 after he sprinted to the left corner. This was just high-level transition offense and shot-making by Duke to take advantage of bad Carolina floor balance.

UNC5 (81-78): With Duke on a 5-0 run, Berry made one of his signature momentum-shifting shots. After Jefferson switched onto him following some perimeter exchanges, Berry drove middle from the left wing in an attempt to get to his preferred right side. Tatum made a strong help rotation, forcing Berry back to the left-side of the rim (and his left hand). This was a fantastic finish from Berry, who’s not always at his best when forced to his left.

DU5 (81-79): UNC switched Britt onto Allen, opting for Jackson’s size on Kennard. Duke used a pick-and-pop to create an iso opportunity for Tatum against Maye (an action it used about a dozen times on Saturday night to isolate Tatum against Hicks/Maye). He attacked immediately to draw a foul on Maye, then split a pair of foul shots to cut the lead to two.

UNC6 (81-79): As Duke was consistently isolating his bench players, Williams went back to his starters with Pinson and Hicks returning to the court. Using a Hicks ball screen, Jackson settled for a contested 18-footer over Jefferson (who switched this time). While Jackson’s getting plenty of praise (deservedly so) for his mid-range game, it’s probably worth making a distinction between 3-level and 4-level scoring (treating the four scoring levels as 1.) close, 2.) 5-10′ (post moves and floaters, generally), 3.) 10-20′ (catch-and-shoot or pull-up mid-range jumpers), and 4.) 3-pointers). Jackson’s been great between 5 and 10 feet this season, almost always utilizing his lethal floater from this distance. But, after this miss, he’s shooting just 24.2% (8-33) from 10 to 20 feet on the season. And, on mid-range pull-up jumpers like this attempt, he’s made just 1 of 17 all year (5.9%). Given that data (and how well-contested the shot was—not to mention that it was a really long 2), this was definitely a win for the Duke defense and a settle by the Carolina offense (although, if you’re going to settle, settle with your best scorer, I guess).

DU6 (81-79): Pinson was back on Kennard, and Duke went back to its right-block floppy screen for him. This time, given Pinson’s size/physicality, Kennard cut to the right wing rather than curling to the right elbow like against Britt. He got a pretty clean 3-point look off, but rushed it a bit due to Pinson’s length/impending close-out. It missed short, with Tatum out-battling (and possibly shoving) Hicks for the offensive board. The Devils reset their offense, going back to the Tatum pick-and-pop iso set. On this occasion, Hicks and Jackson (defending Allen) switched the ball screen, leaving Tatum isolated against Jackson. Undeterred, the brash rookie immediately looked to attack. But Carolina’s wily veteran was one step ahead, moving his feet well to force a Tatum push-off/offensive foul. It was only the fourth offensive foul that Jackson’s drawn this season (all in ACC games), but this one was certainly at a critical time.

UNC7 (83-79): Following a Meeks pindown screen out of a Carolina box set, Berry received the ball on the left wing. Meeks immediately sealed Jefferson after setting the screen, giving Berry an option. He could have taken one hard dribble to the left and spun in a lefty entry pass to Meeks to lead him perfectly to the hoop for a layup. Or, he could have taken one hard dribble to the right to launch a contested 16-foot, left-elbow jumper over the longer Allen. The percentage play, especially in Carolina’s post-centric system, was probably the former—an entry to Meeks. Berry, of course, chose the latter, knocking down the tough mid-range jumper. If nothing else, this play provides a perfect view into the crunch-time mind of Berry. When the chips are down, and the choice is to trust his passing or shot-making abilities, he’ll fall back on his scoring ability nearly every time. The good news for UNC fans is that Berry is a cold-blooded assassin in situations like this. He’s not afraid of taking big shots, and can create/hit them even when well-defended. On the year, Berry’s now shooting 50.0% (18-36) on mid-range jumpers off the dribble.

DU7 (83-79): Allen used a little brush screen by Kennard beyond the arc to drive on Jackson, before wildly throwing his body into a helping Meeks to again get to the line (he had 11 FTAs on just 4 FGAs on Saturday night). As possible cosmic intervention from his earlier foul-drawing theatrics (or, more likely, late-game fatigue or good, old-fashioned choking/bad luck), Allen missed both free throws with Hicks grabbing the defensive board.

UNC8 (85-79): After turning down a Hicks ball screen late in the shot clock, Berry drove right on Kennard, then banked in a short floater over a helping Allen. Learning from his earlier offensive foul drawn by the helping Jefferson, Berry didn’t over-penetrate this time, opting for the floater rather than getting all the way to the rim. Though not as prolific with the floater as Jackson, Berry’s actually been more accurate with his this season, converting 57% of his 21 such attempts this season (speaking of 4-level scorers!).

DU8 (85-79): A Tatum drive-and-kick, which hoped to set up a clean 3-point look for Frank Jackson, was thwarted by a great fake-and-retreat maneuver by Berry. After faking a help rotation, he quickly recovered to contest Jackson’s 3 after the kick-out pass (this fake-and-retreat style of defending the drive-and-kick was memorably used by Raymond Felton on his late deflection/steal to essentially seal the 2005 national championship against Illinois). Berry’s good close-out forced a Jackson brick off the backboard which, when rebounded by Pinson, gave the Heels the ball and a six-point lead with only 1:2o left on the clock. This would essentially wrap it up for the Heels (despite a silly Pinson foul that allowed Kennard to convert a 3-point play and cut the Duke deficit to three; Theo immediately made up for it by hitting Jackson for a transition layup, his seventh assist of the game).

When the pressure was on, Carolina did a great job of getting the ball to its best scorers. Jackson and Berry drove all the action, either as shooters or playmakers. The Heels leaned on two-man action (ball screens and dribble hand-offs), mixing in a box set or two along the way. On the defensive end, Carolina made a strong Duke offense work hard for its looks. The help rotations were crisp, as were the hedges/recoveries. It certainly wasn’t perfect defensive execution (and even that won’t stop an elite scorer like Kennard from getting buckets), but Saturday’s sense of urgency on that end will be what’s required for another march through March.

Big Game Berry + defensive urgency + late-game execution bodes well for a Carolina postseason run. Buckle up: it’s about to get fun!


UNC-Duke: Game-Day Notes

UNC-Duke: Game-Day Notes

Hey! I’ve been knocked out most of the week with the flu (did watch a copious amount of old UNC-Duke games on ESPN Classic in between naps), so didn’t get a chance to do much posting. I’m feeling mostly better today, though, so thought I’d write up a few pre-game notes for tonight’s huge matchup.

  • The earlier UNC-Duke this season had 67.5 possessions, moving Duke’s record against the Heels to 8-3 in the Roy Williams era in games with fewer than 70 possessions. When UNC’s able to ramp the tempo above 70, it improves to 8-9 in the rivalry under Williams.
  • The reason for the above stat, generally, is that Duke has out-executed Carolina in the half-court. Oftentimes, that’s as simple as spreading the court and running high ball screens to set up drive and kicks for its shooters. In the first matchup, Duke also ran some of its NBA-inspired sets (floppy sets for a curling Kennard, or horns sets for Allen dribble hand-offs) to get its scorers the ball in advantageous spots. With Frank Jackson in the starting lineup this time, there might be a return to more of an emphasis on high screen action. Since Jackson, Kennard, and Allen are all players who can finish in the paint as well as kicking to open shooters, they can be tough covers. It’s not as simple as sticking to shooters (as that will risk foul trouble to UNC’s bigs and/or finishes at the rim), and all Carolina fans are aware of the dangers of overhelping against the Duke drive (open 3s!).
  • In that first matchup, the offensive efficiencies by possession length were:
    • 1-10 seconds: UNC—121.4 in 42% of possessions, Duke—126.1 in 34% of possessions
    • 11-30 seconds: UNC—112.8 in 58% of possessions, Duke—126.7 in 66% of possessions
  • Duke used more of its possessions in the halfcourt, and also used them more efficiently than Carolina. As the possession grew longer, Duke became more efficient, scoring 1.37 PPP in trips that took 18 seconds or longer.
  • I charted 19 Duke high screens in the first UNC matchup. Those screens resulted in six offensive resets (i.e., no immediate action directly resulting from the screen/subsequent attack). Of the 13 that did result in immediate action, Duke shot 6-of-9, including 4-of-7 on 3s. It also drew two fouls while committing only one turnover. Again, with Jackson seeing more minutes, there might be even more pick-and-roll offense. The Heels will need to defend it better to walk away with a win.
  • In ACC games, UNC is using 52% of its possessions in the early offense (seconds 1-10). On the season, it’s 56%. The typical Roy Williams Carolina team has been closer to 60% in this metric. In any case, the Heels will need to find a way to improve upon its early offense rate of 42% from the first Duke matchup (while maintaining its early offense efficiency in the 120 range). Part of that is through more transition (primary/secondary break) opportunities, but another big part of “early offense” is creating second-chances/quick put-backs (although this, too, is often easier in the open court/against recovering defenses). An important metric to keep an eye on tonight is how many early offense possessions the Heels are getting (and, obviously, how effectively they’re converting them).
  • Speaking of early vs. half-court offense, Carolina is coming off a season-worst half-court PPP of 0.52 versus Virginia. The Heels scored just 15 points on 29 possessions of between 11-30 seconds against the Cavs. Their early offensive efficiency (1.o4 PPP in 48% of possessions) wasn’t great, either, of course. In two games against UVa. this season, UNC was shut out (o points in 11 possessions) in the final six seconds of the shot clock.
  • After the UVa. game, Carolina is now less efficient in the halfcourt than in the early offense for the first time this ACC season. In conference games, the Heels have scored 1.16 PPP in seconds 1-10 of the clock (52% of possessions) and 1.15 PPP in seconds 11-30 (48% of possessions). In all games, UNC is still slightly more efficient in the halfcourt (a first for the RoyW era): 1.17 PPP in seconds 1-10 (56%) and 1.18 PPP in seconds 11-30 (44%). It’s been an excellent half-court team this season that will need to bounce back from a very poor offensive performance in Charlottesville.
  • In the first Duke matchup, UNC threw only 19 post entry passes, down from its season average of about 24 per game. Part of that was no Isaiah Hicks. A bigger part, perhaps, was how well/aggressively Amile Jefferson fronted (or three-quartered) the post against Kennedy Meeks. Jefferson was really physical, and worked really hard to deny Meeks easy entries (and second-chances) in Durham. That, in my opinion, was an underrated aspect of Carolina’s loss in the first edition of the rivalry. The Heels were fairly successful when they did feed the post (8-of-12 shooting, one foul, three turnovers, three offensive resets)—they just didn’t do it as often as usual. That’s not surprising against Duke, who always tries to take away post entries by pressuring the ball and fronting the post. Rather than trying to force-feed the paint, the Heels will need to attack off the dribble early and often. Let the bigs get their touches on the offensive glass. There, of course, will be opportunities (especially in the secondary break) to make clean post entries. But the default mentality should probably be to attack off the bounce rather than to probe for a post entry (a deviation from how the Heels attack most opponents).
  • UNC’s defense has been performing much better overall than when it played Duke the first time around (in game 12 of the ACC season). Let’s break down the Heels’ defense by ACC segment:
    • Games 1-4: 97.1 defensive efficiency, 90.6 adjusted defensive efficiency (8th in nation if maintained over the course of the season)
    • Games 5-8: 105.2 defensive efficiency, 97.0 adjusted defensive efficiency (48th)
    • Games 9-12: 120.5 defensive efficiency, 108.8 adjusted defensive efficiency (251st)
    • Games 13-17: 91.1 defensive efficiency, 83.0 adjusted defensive efficiency (1st)
  • As seen, the first Duke game completed a terrible four-game stretch of defense for the Heels in which they were allowing 1.21 PPP Even adjusting for the strength of opposing offenses, Carolina allowed 1.09 PPP—a mark that would place it 251st in the country if maintained season-long.
  • Since the first Duke game, however, UNC has allowed just 0.91 PPP (down to 0.83 when adjusted for opposing offensive strength). While some better shot luck has played into that stronger defensive number, there’s no denying that the Heels have been significantly better on the defensive end since the first loss to Duke.
  • In the first Duke matchup, UNC allowed 24 open or lightly contested 3s (Duke made 13 of them). That number will need to come down this time around (or the Heels will need to hope for lots of shot luck). In the last game against Virginia, UNC allowed only 13 of the Cavaliers’ 24 3s to be open or lightly contested (they made 8 of those 13, compared to just 2-of-11 that were well-contested). Against Duke in round 1, 89% of the 3s that UNC allowed were open/lightly contested. Last game against Virginia, only 54% were. If tonight’s number is closer to 54% than 89%, the Heels will probably emerge victorious.

I’ll be back after the game with some quick statistical tidbits, then later this weekend with some more detailed charting analysis.

Enjoy the game!


Pack-Line Problems

Pack-Line Problems

Tony Bennett’s pack-line defense has traditionally produced some of the best defenses in the country. Including his three-year stint at Washington State, Bennett’s teams have been in the top 25 in adjusted defensive efficiency nine times in his 11 seasons as a head coach. Six times Bennett has fielded a top-10 defense, including four top-5 defenses in the last six seasons. This year, the Cavaliers lead the nation in adjusted defensive efficiency. So, suffice it to say, the way that Bennett teaches his pack-line principles is very effective (including hard hedges, immediate big-to-big doubles on post entries, and clogging driving lanes with help defenders rather than overplaying the wings).

Let’s take a closer look at some of the Carolina offensive struggles on Monday night.

Shot Distribution

On Monday night, the UNC shot distribution looked as follows:

  • Close: 9-21 (42.9%) –> 44% of FGAs
  • 5-10′: 2-6 (33.3%) –> 13% of FGAs
  • 10-20′: 2-5 (40.0%) –> 10% of FGAs
  • 3-pointers: 4-16 (25.0%) –> 33% of FGAs

On the season (entering last night’s game), the Heels’ distribution of FGAs was:

  • Close: 60.2% –> 42% of FGAs
  • 5-10′: 41.4% –> 14% of FGAs
  • 10-20′: 35.3% –> 13% of FGAs
  • 3-pointers: 37.3% –> 30% of FGAs

Carolina created the same shots it always does—in fact, a slightly higher proportion at the rim, and a slightly lower fraction from the low-efficiency mid-range. The rate of 3-pointers was up a tick, but only due to a few late-game, desperation attempts. The obvious discrepancies between the Monday night and year-to-date numbers, of course, are the shooting percentages—particularly at the rim and from behind the arc. Of UNC’s 16 3-pointers, I classified two as open, 11 as lightly contested, and three as contested. The Heels’ two primary shooters, Justin Jackson and Joel Berry, combined for 11 of the 16 attempts behind the arc (including a Berry attempt that was erroneously credited to Kennedy Meeks). It was a different story at the rim, however, where Virginia blocked eight of Carolina’s 21 attempts. Many of the non-blocked close attempts were also well-contested (often using Bennett’s signature style of going straight up with the hands while using the lower body to bump/displace the shooter). The Cavs’ rim protection was excellent on Monday night, but close attempts are exactly what Roy Williams’ offense is trying to create. Those weren’t shot selections issues—just a combination of stellar UVa. paint defense and some problems finishing through contact for the Heels. Overall, the shot selection for Carolina was satisfactory. Jackson took a couple contested 3s, plus a very bad, off-balance long 2. Britt had a contested mid-range attempt that UNC can probably live without. There was only one late-clock situation that required a tough shot (a Seventh Woods’ pull-up jumper). But, in general, the Heels got the shots they needed to in order to win the game. Based on season averages, Carolina will score about 43 points on 21 close attempts and 16 3-pointers. Against Virginia, the Heels managed only 30 points on those 37 attempts.

Even given the UNC turnover issues (the Heels turned it over on 25% of their possessions, including 40% in the first half—their year-to-date average entering the game was 16.3%), it did enough to win the game had it simply knocked down a couple more 3s and finished a couple more close attempts. In just 56 offensive possessions, Carolina threw a staggering 30 post entries (it averages about 23 per game on the season). Those passes resulted in eight made field goals, 10 missed field goals, eight turnovers, one foul (non-shooting), and three offensive resets. The glaring number there, of course, is the eight turnovers. Most of UNC’s miscues were a function of trying to feed the post (and the subsequent action following the hard post-to-post Virginia double teams). Let’s take a closer look at how Carolina handled the big-to-big doubles:

UNC vs. the Big-to-Big Double

I charted 13 times in which the Cavaliers immediately send a big-to-big double following a UNC post touch. Let’s see what happened on those plays, in chronological order:

  1. Meeks, left block: Meeks attempted to hit a diving Hicks at the front of the rim, but a helping Kyle Guy was able to disrupt the play from behind to force a turnover. This is exactly how the Heels wanted to attack the double. Meeks’ pass was a split-second late, and Hicks needs to be stronger with the catch. Had this been successfully completed, however, it’s an easy layup/dunk.
  2. Hicks, right block: This time, Hicks was able to successfully complete the pass to the diving Meeks. With UVa. point guard Ty Jerome helping down (and giving up five inches and 70 pounds), Meeks simply needs to finish this opportunity at the rim. Good execution, bad finish.
  3. Hicks, right block: Hicks, after catching the entry pass too far off the block, used an escape dribble to reset the offense.
  4. Meeks, left block: Following the Hicks escape dribble/reset, UNC immediately entered the ball to Meeks on the opposite block. He was stripped by a doubling Devon Hall while trying to make a pass. This is a case of Meeks needing to be stronger with the ball.
  5. Meeks, right block: Meeks immediately turned baseline (away from the approaching double) to bank in a short jump hook. This was a quick decisive move by Meeks, who, given his proclivity for turning left shoulder, will generally do better against post doubles when receiving it on the right block.
  6. Meeks, left block: Meeks was forced to pass it back to Britt in the ball-side corner here, a win for the UVa. defense since the ball stayed on the same side of the court. With the possession sputtering following the post double/kick-out, Britt settled for (and missed) a contested mid-range jumper.
  7. Maye, left block: Maye kicked it to the opposite wing here to Seventh Woods. Had this been Berry at point guard, it would have resulted in a clean 3-point look. Woods, a reluctant perimeter shooter, shot-faked, then traveled on his drive to the hoop. This was well-executed by Maye/UNC on the post double, but just a personnel issue in this particular lineup.
  8. Meeks, left block: Again, Meeks passed to the opposite (right) wing—this time for a clean inside-out Britt 3-pointer. This is Carolina’s bread-and-butter—a post touch leading to an inside-out look.
  9. Meeks, left block: For the third consecutive post double, a UNC big (Meeks again) on the left block looked diagonally to the right wing. This time, it was Berry receiving the pass and missing a lightly contested 3-pointer. Can’t argue with the execution or shot selection here.
  10. Meeks, right block: Like his earlier make, Meeks again spun quickly to the baseline to attempt a jump hook. This one was better defended by Virginia, but still a strong, decisive move by Meeks by attacking before the double can arrive.
  11. Hicks, left block: Hicks used an escape dribble to relocate to the left wing. Meeks then filled in Hicks’ vacated spot on the left block to receive a post entry from him. Meeks turned it over by trying to spin around Jack Salt (setting a solid wall) in the paint. This was vintage Roy Williams basketball; Meeks just needs to be more efficient in the paint.
  12. Maye, left block: Maye used a single escape dribble to create some space, then kicked it opposite to Berry on the right wing. This time, Berry knocked down the clean look. Great work by Maye here against the post double.
  13. Meeks, left block: Meeks, this time spinning middle, was able to get off a clean jump hook in the paint. He missed, but no issues with the shot selection here.

So on 13 post doubles (and 12 possessions), Carolina scored eight points. Meeks made 1-of-3 shots while fighting through doubles/shooting before they arrived. The Heels also made 1-of-3 3s created from inside-out passes following a big-to-big double team. After attempting to hit the diving big on the first two tries, UNC got away from that option later in the game. While it certainly wasn’t a clinic on defeating Virginia’s post double (Brice Johnson was much more effective in last year’s match-ups, creating more close opportunities for his diving fellow post), Carolina’s execution here was adequate. It certainly wasn’t the reason the Heels lost the game. More problematic, perhaps, was UNC’s execution on its ball screens (against Virginia’s hard hedging strategy).

Attacking the Hard Hedge

Bennett’s defensive philosophy includes hard-hedging of ball screens, meaning the help defender aggressively moves into the ball-handler’s path to force him laterally (or even backwards) while the on-ball defender recovers. This technique used to be (as recently as the middle of last season) Roy Williams’ preferred one against the ball screen, too. But due to some physical (Meeks) and mental (Brice Johnson and Hicks’ proclivity for picking up cheap fouls by bumping the dribbler) limitations, Williams moved to a flat hedge technique designed to curtail dribble penetration and force mid-range jumpers. One could, of course, argue that if Johnson/Hicks were allowed to be as physical with their hedges as Virginia’s big were last night, Carolina would still be employing the hard hedge. But that’s a bit of a digression.

“Attacking” is probably the wrong word for how UNC responded to the Cavs’ hard hedge last night. To successfully beat this technique, ball-handlers generally need to turn the corner or split the defenders to get into the paint. The Heels did neither consistently last night, instead allowing the Virginia helping big to force them laterally (or, too often, backwards) and force an offensive reset/turnover. This Carolina team, while having a variety of guards/wings that can get to the basket off the bounce, lacks that Ty Lawson-style attacker who can turn the corner on anyone, As such, it’s sometimes susceptible to an aggressive ball screen defense like Bennett used on Monday night. Another way to beat the hard hedge is by slipping screens. This is a core option of Carolina’s secondary break, but the Heels only slipped a single screen on Monday (resulting in a Hicks travel after a great Virginia help rotation).

Carolina used 31 high screen against Virginia (the vast majority of which were hard hedged). Those actions resulted in the following outcomes: 3-of-10 shooting, three drawn fouls (one shooting foul drawn by Brandon Robinson), five turnovers, and 13 offensive resets (where UNC just had to restart its offense, generally as a result of being pushed out deep by the hedger). On 31 ball screens, the Heels created only seven points. Breaking it down by Carolina ball-handler:

  • Berry: 11 screens—4 missed shots (Jackson pick-and-pop, Hicks missed lay-up as after pocket pass to roller, Britt missed 3 after drive-and-kick, Pinson missed 3 after perimeter pass), 4 resets, 2 fouls (when Berry aggressively drove into the hedger to force the whistle), and 1 TO (a Berry ball-handling turnover near the UVa. bench)
  • Pinson: 9 screens—4 resets, 3 TOs (Bradley charge after a pocket pass, Hicks charge after a pocket pass, Hicks travel after slipping a screen), 2 made shots (pick-and-pop with Hicks who hit a 12-footer, pass to Maye who entered the ball for a Bradley layup)
  • Jackson: 8 screens—4 resets, 2 missed shots (a Jackson long, contested 2 off the bounce, a missed Bradley layup), 1 made shot (a Bradley dunk after Jackson hit Pinson as a pressure release, who whipped it in to a rolling Bradley), and 1 TO (a Jackson ball-handling turnover when trying to split the defenders)
  • Robinson: 1 screen—1 foul (successfully split the hedge to draw a foul at the rim)
  • Woods: 1 screen—1 missed shot (a Maye pick-and-pop 3)
  • Britt: 1 screen— 1 reset

The Heels only tried to split the hard hedge three times: Robinson’s foul, Jackson’s turnover, and another time by Jackson when he found Meeks in the paint, but the ball was deflected out of bounds. And, as mentioned, there was only one attempted slip (the Hicks travel). What did happen was plenty of side-to-side dribbling. If Carolina meets Virginia again in the ACC Tournament, it will be interesting to see what (if any) adjustments it makes in attacking the hard hedge.

Virginia’s obviously a very disciplined and well-drilled defense. It executes its pack-line principles excellently, while also trying to take away its opponents’ go-to sets. In Carolina’s case, that meant shutting down the secondary break by bumping/holding cutters and hard hedging ball screens. The Heels got an early lob to Meeks off of a secondary back screen, but otherwise the Cavs shut down most of the initial looks via physical defense/keeping UNC from getting to its spots in a timely manner. Rather than continuing to run secondary without creating good scoring chances, Carolina could have tried more quick hitters out of its 1-4 set. Very early in the game (to make the score 4-0), the Heels ran Jackson off an elbow curl to create a short floater for him. As he was being guarded by the smaller London Perrantes, going back to that curl repeatedly might have made sense. UNC didn’t run it again after that early Jackson hoop. Virginia was also well-scouted on Carolina’s use of the box sets. The Heels didn’t have a ton of success with its box formations, as physical defense and scouting conspired to take away most of the options. Carolina, anticipating Virginia’s help defense/hedge, was able to slip Bradley (after he screened in an attempt to free Jackson coming through the elevator doors) on the final play of the first half. This was a nice call by the bench, but resulted in a missed Bradley attempt at the rim following another good Cavalier help rotation.

While it’s easy to be critical of the coaching staff after the team lays an egg offensively, I actually thought Carolina got a lot of the shots it wanted. Plenty of post touches/close attempts, as well as clean looks for its best 3-point shooters. Certainly the bigs need to be stronger with the ball, and with finishing through contact. A few wrinkles against the hard hedge (more slips, or pressure release passes) might be a nice adjustment, as would be moe quick hitters/Jackson curls in the early offense (rather than such a steady diet of secondary break). Really, though it’s easy for us as Carolina fans to view things through a Heels-centric lens, the Virginia defense deserves a ton of credit for its tremendous effort and execution of Bennett’s defense. He’s a terrific defensive coach and, sometimes, you just need to tip your cap to the opponent (even if it’s a physical, hand-checking, body-bumping one that maybe took advantage of some favorable officiating).

I’ll be back later with a bit on Carolina’s defense (spoiler alert: I actually thought it was even better than against the Cavs in Chapel Hill), then we’ll be on to Duke!



The Saturday Clipboard

The Saturday Clipboard

A few charting-related nuggets to pass along before the Carolina-Pitt game tips off at noon:

First, let’s break down Carolina’s top scorers (Jackson and Berry) by their early offense vs. half-court offense splits.

  • As seen, each player uses roughly half his weighted shots (FGAs + 0.475*FTAs) in each segment. Combined, Berry and Jackson score 16.6 points per game in the first 10 seconds of the shot clock, and 16.7 in seconds 11-30.
  • Both players shoot 2-pointers better in the early offense (due to transition opportunities), but 3-pointers better in the halfcourt. Jackson’s 3-point split is more dramatic. Also, not surprisingly, both players draw significantly more fouls in the early offense (against oftentimes unset/scrambling/transitioning defenses).
  • Both players also shoot more 3s (as a proportion of total FGAs) in the halfcourt. The combination of 3-point volume and efficiency from Berry and Jackson in seconds 11-30 is why Carolina’s offense has been so good and balanced (between early and halfcourt) this season.
  • Jackson 3-point percentage actually gets higher and higher and the shot clock gets shorter and shorter:
    • 1-10: 31.3% (26-83)
    • 11-17: 41.8% (28-67)
    • 18-24: 48.6% (18-37)
    • 25-30: 60.0% (6-10)
  • Only three Tar Heels have taken double-digit FGAs in the final six seconds of the shot clock (Meeks and Hicks each have nine FGAs).
    • Berry: 50.0 FG% (11-22), 60.0 3Pt% (6-10), 67.7 TS%
    • Jackson: 62.5 FG% (11-16), 60.0 3Pt% (6-10), 83.2 TS%
    • Britt: 11.8 FG% (2-17), 20.0 3Pt% (1-5), 14.7 TS%
    • Berry’s actually been trending in the wrong direction here (after a really efficient start to the season in late-clock situations). Jackson’s been consistently great all year with an expiring shot clock; Britt’s been consistently bad.
  • UNC’s leaders in off-hand FGAs:
    • Meeks: 13-17
    • Jackson: 9-15
    • Britt: 9-11 (doesn’t really have an “off” hand, I guess—these are lefty attempts (all at the rim), though)
    • Berry: 5-7
    • Pinson: 3-7
    • Hicks: 3-7
    • Woods: 4-5
    • Williams: 4-4
    • Maye: 2-4
    • Bradley: 2-2
    • Meeks has been using his left hand more and more from the left-side of the rim, and has been steadily raising his close FG% from that side. It’s still at just 50.0% (31-62), though, compared to 65.4% (17-26) from the close middle and 69.2% (63-91) from the close right. Bradley, likewise (who still doesn’t use his left hand much), is shooting 58.1% (25-43) on close left attempts. That’s below his close middle (63.0% on 17-27) and close right (71.4% on 20-28) marks.
  • Jackson’s left wing/right wing 3-point splits continue to be extremely pronounced. He’s made 35-of-67 3s from the left wing (52.2%), but only 15-of-55 (27.3%) from the right wing. From the top of the key, he’s somewhere in between at 35.7% (15-42).
  • The Maye-Bradley frontcourt has been heavily used by Roy Williams in the ACC, and has had fantastic +/- results. It’s actually the second-most-used frontcourt in conference games, and has the highest efficiency margin of any combination.
    • Hicks-Meeks: 224 ACC minutes, +18.1 efficiency margin
    • Maye-Bradley: 93 ACC minutes, +33.5 efficiency margin
    • Maye-Meeks: 85 ACC minutes, +6.5 efficiency margin
    • Pinson as 4: 57 ACC minutes, +3.9 efficiency margin
    • Jackson as 4: 55 ACC minutes, -7.9 efficiency margin
    • Hicks-Bradley: 49 ACC minutes, +17.0 efficiency margin
    • Bradley-Meeks: 14 ACC minutes, +24.9 efficiency margin
  • As the above data shows, Carolina’s small-ball lineups have not been effective (from a +/- perspective) in the ACC. Pinson’s efficiency margin splits by position have been:
    • As a 2: 72 minutes, +26.4 efficiency margin (110.4-84.0)
    • As a 3: 65 minutes, +32.9 efficiency margin (115.9-83.1)
    • As a 4: 57 minutes, +3.9 efficiency margin (124.8-120.9)
    • The offensive efficiency has been great with Pinson at the 4. However, the team’s inability to get consistent stops has more than offset any gains in scoring production. The defense has been terrific in Pinson’s wing minutes (whether at the 2 or the 3). His minutes have been pretty evenly split across all three spots so far; since Kenny Williams’ injury, of course, they’ve been shifting more heavily to the 2.
  • Berry-Pinson has also clearly been UNC’s best backcourt against top competition. In minutes against Pomeroy Tier A&B opponents (top-100, venue-adjusted competition), Berry-Pinson has an efficiency margin of +27.1 in 60 minutes. Berry-Williams and Berry-Britt have both played 221 minutes against Tier A&B foes, with respective efficiency margins of +14.5 and +6.9.


Better Defense or Luckier Defense?

Better Defense or Luckier Defense?

It’s been well-documented that Carolina’s defense has been playing much better recently. After dropping to the mid-40s in Pomeroy’s adjusted defensive efficiency rankings, the Heels have climbed all the way back to No. 20 with dominant back-to-back home performances against Virginia and Louisville. Narratives being what they are, you’ve surely heard plenty of buzz about Carolina’s defensive “effort” and “focus” being better over the past couple of games. But how much is due to actually playing better (or harder) defense versus simply playing luckier defense? Let’s dive a little deeper inside the numbers to investigate.

We’ll start by breaking UNC’s ACC campaign into three segments: 1.) the first six games (Georgia Tech-Syracuse); 2) the next six games (Boston College-Duke); and 3.) the last three games (@NC State-Louisville). The first table below summarizes the Heels’ defensive Four Factors by season segment.

As seen, Carolina’s defensive efficiency (both adjusted and unadjusted—strength of opposing offense has stayed very constant across segment: 114.1, 114.9, and 115.3, respectively) got dramatically worse over games 7-12 before making a huge improvement in the last three games. So what’s happened? Defensive rebounding and keeping opponents off the foul line have been the two factors in which the Heels have been consistently solid this ACC season. The DR% has spiked up a bit lately, but that certainly wasn’t the reason for Carolina’s mid-schedule defensive swoon. The other two factors—eFG% and forced TO%—have been the more volatile ones. eFG% has been especially noisy. In an effort to explain the variability in opposing shooting percentages, let’s look at another table:

This data breaks down Carolina’s eFG% defense (and opponents’ shot distribution) by how well the Heels contested the shot. Generally these categories can be described as:

  • Open: wide-open shot without even a late closeout/contest
  • Lightly contested (LC): the close-out is either a step late (common on Carolina’s help-and-recover 3-point defense), or otherwise not strongly contested
  • Contested: a well-positioned defender gets a hand up to force a tough shot (think of a solid wall in the post, or a perimeter defender who’s right in the face of an opposing shooter)
  • Heavily contested: generally a blocked shot (or one that’s not blocked, but just thrown wildly in the direction of the basket)

The proportion of shots that the Heels have contested well (contested + heavily contested) has stayed fairly consistent segment-over-segment. It was 36.2% in the first six games, 33.6% in the next six, and 34.7% over the last three. What’s been more volatile is opponents’ eFG% across levels of contestedness. While open and heavily contested shots have been consistent (always very good or very bad), both contested and (especially) lightly contested efficiency has varied wildly. It’s not surprising that in the second segment—when Carolina’s defense was slumping—opponents’ were shooting the best on these types of shots (i.e., opponents’ luck was inversely correlated with UNC’s defensive effectiveness). Likewise, over the past three games, opponents have been shooting their worst on lightly contested/contested shots.

Ken Pomeroy’s famously (at least in basketball analytics circles!) illustrated that 3-point defense shouldn’t be defined by percentage (which is mainly randomness/shot luck), but rather volume (i.e., reducing opponents’ attempts from behind the arc). While that’s generally true, charting stats can shine a little more light on the subject. That is, it’s better to give up 20 3-point attempts if 10 are lightly contested and 10 are contested than it is to give up 15 3-points attempts if all 15 are lightly contested. Open/lightly contested 3s go in about 40% of the time (an eFG% of 60%), while contested ones are usually between 10-20% shots (eFG% of 15-30%). Since teams (at least well-coached ones) don’t generally settle for too many bad 3-pointers, Pomeroy’s conclusion is true in the global sense. So what’s driving UNC’s recent run of 3-point% defense effectiveness: better shot contesting or better luck?

Here are the game-by-game 3-point shooting splits for UNC’s ACC foes. This list shows their 3-point% on open/lightly contested 3s, as well as the proportion of total 3s that were open/lightly contested (with the remainder, of course, being well-contested).

  • Georgia Tech: 30.0% (3-10) on open/lightly contested; 90.9% of total 3s open/lightly contested
  • Clemson: 44.0% (11-25); 89.3%
  • NC State: 42.9% (6-14); 60.9%
  • Wake Forest: 47.8% (11-23); 85.2%
  • Florida State: 39.1% (9-23); 82.1%
  • Syracuse: 47.4% (9-19); 79.2%
  • Boston College: 47.8% (11-23); 85.2%
  • Virginia Tech: 42.9% (9-21); 84.0%
  • Miami: 46.7% (7-15); 75.0%
  • Pitt: 54.5% (12-22); 75.9%
  • Notre Dame: 47.4% (9-19); 73.1%
  • Duke: 54.2% (13-24); 88.9%
  • NC State: 50.0% (8-16); 66.7%
  • Virginia: 13.3% (2-15); 75.0%
  • Louisville: 29.4% (5-17); 85.0%

To summarize by season segment:

  • ACC games 1-6: 44.1% (49-111); 78.7%
  • ACC games 7-12: 49.2% (61-124); 80.5%
  • ACC games 13-15: 31.3% (15-48); 76.2%

A couple key takeaways: 1.) UNC’s been slightly better at contesting 3s lately (as only 76.2% of opponents’ 3s have been open/lightly contested in the last three games); 2.) UNC’s been significantly luckier at defending 3s lately (Virginia and Louisville have combined to make just 21.9% (7-32) of its open/lightly contested 3s).

So it’s maybe one part better perimeter defense to ten parts luckier perimeter defense. It should also be noted that Carolina had been especially unlucky in the first dozen games of the conference season. Historically (since 2008, at least, when I began charting it), UNC’s opponents have made between 37 and 42% of their open/lightly contested 3s. That jumped the whole way to 46.8% in the first 12 games of the 2017 ACC season (an unsustainable/unlucky number).

Another quick note: the Heels were also playing better-shooting teams during its slump. During games 7-12, the Heels played the top-4 3-point shooting teams (by percentage) in the ACC: Virginia Tech, Duke, Notre Dame, and Pitt (who combined to make exactly half of their 86 open/lightly 3s against UNC). In games 1-6, the average 3-point rank of UNC’s opponents was 10.5. It fell to 5.8 in games 7-12, before rising again to 10.0 over the last three games. Because of how Carolina defends (helping off the wings on shooters, emphasis on defending the paint), it will always be susceptible to teams that spread the floor with good shooters to set up drive and kick opportunities. Its shouldn’t be too surprising that the Heels’ defensive slump coincided with facing several such teams.

All of that isn’t to say that Carolina’s defense hasn’t improved lately. As shown, the Heels are contesting shots (both 2s and 3s) slightly better. They’ve also been even more dominant on the defensive glass. Most importantly, UNC’s forced TO% is also climbing back up recently. That’s been driven by increased disruption by defensive catalyst Joel Berry. His forced TO / 40 by season segment have been:

  • ACC games 1-6: 3.11
  • ACC games 7-12: 2.13
  • ACC games 13-15: 4.89

Berry, whose individual defensive slump (not coincidentally) mirrored the team’s, has been much better on the defensive end over the past few games. That clearly makes a difference in UNC’s effectiveness irrespective of shot luck.

Luck is important in basketball. Illinois made just 12-of-40 3s (most of them lightly contested) in the ’05 title game against the Heels. Likewise, Carolina’s lost to hot-shooting teams in the NCAA Tournament (2011 Kentucky in the Elite 8, countless other examples I’m probably repressing). And luck’s been very important to UNC’s recent defensive surge. But if the Heels continue to combine their top-5 offense with three strong defensive factors (DR%, FTA Rate, and TOF%), they’ll be able to survive all but the worst cases of bad shot luck (and/or bad 3-point defense/allowing too many lightly contested 3s). And, if it continues its recent shot luck into the postseason (or, ideally, starts running a few more shooters off the line with well-timed help-and-recover close-outs), Carolina could give Roy Williams his third national title.



Facing Top Defenses in the Williams Era

Facing Top Defenses in the Williams Era

Despite allowing a startling 1.45 PPP (90 points on 62 possessions) in its last game against Virginia Tech (at home, no less!), Louisville remains fifth in the nation in adjusted defensive efficiency. Always stout defensively under Rick Pitino, the Cardinals have ranked in the top-5 in this metric for a staggering seven straight seasons (assuming they can hold on to it this year).

So how has Carolina performed against top-10 defenses (based on Pomeroy’s adjusted defensive efficiency) in the 14-year Roy Williams era? Let’s take a look. All these numbers are from the 2003-04 season through the Virginia game this year. It should be noted that these are using the end-of-year numbers rather than the time-of-game ones.

  • UNC vs. teams with top-10 offenses and defenses: 4-9
  • UNC vs. teams with top-25 offenses and defenses: 20-28
  • UNC vs. teams with a top-10 defense only: 31-22
  • UNC vs. teams with a top-10 offense only: 27-26

Louisville this season, ranked 17th in offensive efficiency and fifth in defensive efficiency, falls into the top-25/top-25 bucket. Not surprisingly, Carolina has struggled some to beat this elite, balanced teams during the Williams era. In the halcyon days under RoyW (2005-09), the Heels went 9-4 against teams with this statistical profile (including 1-1 against top-10/top-10’s—an ’05 championship-game win over Illinois, and a Final Four loss to ’08 Kansas). But in the other nine season under Williams, UNC has gone just 10-23 (UNC was 1-2 last year against top-25/top-25—splitting with Virginia, and losing to Villanova in the championship game; the Heels split with Florida State and Kentucky in their only two such games this season).

As seen from the records above, Carolina has played better against teams with elite (top-10) defenses/non-elite offenses (58.5 winning percentage) than it has against elite offenses/non-elite defenses (50.9%). So far this season, UNC is 2-1 against both of these types of teams: wins over Oklahoma State and Wake Forest, and a loss to Duke in the elite offense/non-elite defense tier, and wins against Wisconsin and Virginia, and a loss to Georgia Tech in the elite defense/non-elite defense bucket.

Not surprisingly, the Heels were also much better against these types of opponents during the 2005-09 high-water period. They went 9-5 against elite offense/non-elite defense teams (vs. 18-21 in all other Williams seasons), and 15-5 against elite defense/non-elite offense opponents (vs. 16-17).

Not a ton to read into this, probably. It’s not breaking news that good, balanced teams are tough to beat. Ranking fourth in offense and 26th in defense, Carolina itself is right on the cusp of being an elite/elite team. Perhaps after tonight’s game, the Heels will find themselves back in that rarefied air.

Shifting gears, let’s briefly discuss how good UNC’s post quartet of Kennedy Meeks, Isaiah Hicks, Tony Bradley, and Luke Maye has been in terms of assist-to-turnover ratio this season. In some cases (Bradley), it’s been more about great ball security. In other cases (Maye), it’s been a function of an above-average assist rate. Meeks has been pretty solid in both A:TO areas. Even Hicks, who’s the worst in the rotation at 0.69, has improved considerably from his sophomore (0.32) and junior 0.57) marks.

As the table below shows, this post rotation currently has the best A:TO of the Williams era:

A couple notes from the table:

  • Stretch 4s are (pretty obviously) always solid in this metric: Noel, Hairston, Jawad Williams, and Watts posted some of the best ratios on the list. Luke Maye fits that mold currently (with the A:TO to match).
  • Henson (from 0.39 to 0.98) and McAdoo (from 0.43 to 1.23) made dramatic and impressive A:TO improvements from their sophomore-to-junior seasons. Both improvements were driven primarily by drastic reductions in turnover rates.
  • The Tylers (Hansbrough and Zeller) were never really into the whole “passing” thing. Still legendary Carolina posts, of course—just not stellar A:TO numbers. It’s not always that crucial of a stat for a post player, but it’s a nice feature for this UNC team since it lacks a go-to post scorer like Hansbrough or Zeller (and instead depends on ball movement and passing democracy).
  • Joel James was obviously a terrific Tar Heel ambassador and locker room/bench presence. A pretty solid fourth big, too. But his career 0.25 A:TO (including 0:15 as a senior in ’16) hasn’t really been missed very much. It was better than Alex Stepheson’s career mark (as a Tar Heel) of 0.20. Believe it or not, Stepheson’s A:TO actually declined to 0.14 (14:101) in his two seasons at USC.
UNC’s Early Offense

UNC’s Early Offense

Yesterday, we looked at Carolina’s efficiency in the early offense (first 10 seconds of the shot clock) versus the half-court (seconds 11-30 of the clock). Against Virginia, true transition (i.e, primary break) opportunities are always at a premium, but that doesn’t mean that a team can’t create plenty of “early offense” chances against Tony Bennett’s team (through things like the secondary break, put-backs, and BLOBs/special situations). And, as seen in the piece from yesterday, the Heels have been more dominant in the half-court this season than in their (generally) preferred early offense.

A big storyline going into yesterday’s game was: who would win the battle of tempo? Since it’s much easier to slow down a game than speed it up, a better way to phrase the question might be: which team would win the early-offense battle, and which would win the half-court battle? Of course, if the same team won both of these facets, that team would obviously win the game (and possibly even dominate it). Yesterday, that team was North Carolina.

Let’s start by breaking down each team’s offensive efficiency by shot-clock segment:

Not surprisingly, Carolina had the clear advantage in early-offense opportunities. The Heels used 44% of their possessions within the first 10 seconds, nearly double the rate of Virginia (23%). In conjunction with UNC’s efficiency advantage in the half-court (a +29.7 margin in seconds 1-10), that gave the Heels a huge +18 (30-12) advantage in early-offense points. Despite having significantly fewer half-court opportunities than the Cavaliers, Carolina compensated by being dramatically more efficient with those chances (a half-court efficiency margin of +43.1). That resulted in a +6 in half-court scoring for North Carolina on Saturday night (35-29). When combining those two UNC advantages, it’s no surprise that the game was a blowout.

The one shot-clock segment that Virginia did control on both ends was late-clock situations. Carolina has held scoreless in its seven possessions in the final six seconds of the clock (0-4, with misses by Berry, Jackson, Britt, and Meeks with 3 TOs (by Berry, Jackson, and Woods (although it was erroneously charged to Britt in the box-score)). Seconds 25-30 of the clock was actually UVa’s most efficient segment, as it scored seven points in eight such possessions. Carolina, as it’s been all season, was especially lethal in seconds 18-24 of the shot clock. That’s generally a sweet spot that occurs after the offense has made the defense shift/probed for openings, but before it’s constrained by an expiring shot clock. In ACC games, the Heels have posted an offensive efficiency of 131.8 in that segment (in 132 possessions). Against the ‘Hoos, it was an even more impressive 177.8. This has also been UNC’s ACC opponents’ most-efficient half-court segment (as it generally is, perhaps for the “sweet spot” hypothesis I postulated above), but the Heels held UVa. to just 0.69 points per possession in seconds 18-24 (on a healthy 16 possessions).

Let’s quickly recap how the Heels created their 26 early-offense opportunities against Virginia, leading to 30 points. As mentioned earlier, the Cavs rarely give up true fast-break points since they generally concede crashing the offensive glass in favor of floor balance/getting back in transition defense. But that’s part of the beauty of Roy Williams’ secondary break system. These are listed chronologically:

  1. Out of one of Carolina’s signature secondary-break actions, Pinson threw a lob for a Hicks dunk. Hicks received a Berry back screen after setting a ball screen for Pinson. Pinson continues to set up UNC’s big for easy hoops: his three assists against Virginia were all to post players (two to Meeks, and this one to Hicks) for a dunk, a layup, and a short hook shot.
  2. Using a secondary-break ball screen from Meeks, Pinson drove the lane but was called for a push-off/offensive foul. His five turnovers this season are three bad passes and two offensive fouls (to go along with 30 assists, plus seven FT assists).
  3. After Berry picked up a backcourt steal, he missed a floater in the lane following the live-ball turnover.
  4. This one was created by another live-ball turnover—this time it was Britt stripping a driving Perrantes with Jackson picking up the loose ball. Jackson pushed it coast-to-coast to draw a foul in a rare primary break opportunity against UVa. He split the free throws.
  5. Jackson hit a secondary break 3 after coming off a Maye screen to receive a dribble hand-off from Britt near the top of the key (shading towards Jackson’s preferred left wing). This transition opportunity was preceded by a long 3-point miss from UVa with a second left on the clock, leading to a long rebound by Bradley.
  6. Bradley tipped around an offensive rebound several times before it was eventually secured by Maye. Maye immediately shoveled it back to Bradley for a FT assist at the rim. Bradley split a pair of free throws, and this was more evidence of the chemistry that’s developed between Carolina’s back-up frontcourt duo.
  7. Berry carelessly lost his dribble out of bounds when attempting to start Carolina’s secondary break from the right wing. It was Berry’s team-high 10th ball-handling turnover of the season (although Woods’ per-40 rate of ball-handling TOs is nearly three times as high as Berry’s).
  8. In a seldom-used baseline out of bounds (BLOB) set, Roy Williams called a play to create a look for the red-hot Jackson. He curled off a staggered double screen from Hicks and Robinson to receive a Woods pass from his left-wing hot spot (Jackson’s made 34-of-65 3s (52.3%) from the left wing, including 2-of-4 vs. Virginia). Although he missed this one, I thought it was a great call by Williams to get his leading scorer a shot.
  9. Woods waved the trailing Hicks out of his usual secondary spot at the top of the key in order to set up a quick hitter out of UNC’s 1-4 alignment. Jackson curled off of a Hicks screen to receive a pass from Woods and hit a floater in the paint while drawing an “and-1.” He’s convert the old-fashioned 3-point play to give UNC a 25-12 lead.
  10. In another secondary break staple, Hicks slipped a screen to receive a pass from Jackson for a dunk. I’m sure the staff worked on this one in practice, as Virginia’s ball screen defense makes it susceptible for the secondary slip. Jackson had a downright Pinsonian game passing the basketball. His six assists resulted in two dunks (to Hicks), three layups (two to Meeks, including an “and-1” and one to Berry), and Pinson corner 3.
  11. After a Pinson steal, he pushed the ball in the primary break to Jackson on the left wing. As Jackson looked to pull the ball back rather than attack the hoop, he was called for a travel.
  12. Bradley blocked a driving layup by Darius Thompson to launch a primary break opportunity. Jackson corralled the defensive board and immediately pushed it himself, hitting Berry for an easy layup as he filled the right wing in transition.
  13. Jackson fed Bradley with a secondary break post entry pass to the left block. Bradley, who had established deep position, wasn’t doubled by Virginia, and missed a good look at a short jump hook over his left shoulder (his go-to post move/location).
  14. After a Meeks block, Pinson grabbed the defensive board and went coast to coast for a primary-break “and-1.” Another example of great Carolina defense fueling its transition game (like the Bradley block above).
  15. Once again a Meeks blocked shot got the Heels out in transition. This time, Berry missed a layup from the right side after making a nifty behind-the-back, hesitation-dribble drive (the quintessential “everything but the finish” play).
  16. Meeks controlled another second-half defensive rebound, throwing an outlet to Jackson who missed a transition 3-pointer. This was a tough, contested 3 off the dribble, and immediately led to a Virginia run-out/open Shayok layup. Jackson didn’t do much wrong on Saturday night, but this shot selection qualifies as one of his poor decisions.
  17. In another secondary break action, Pinson curled off a Meeks screen, then hit the rolling big for a lefty layup. A great pass by Pinson, and another example of the secondary break creating a quick score (although not one that’s considered “fast break” points in the box score).
  18. Woods threw a secondary-break entry to Meeks on the left block, and the ‘Hoos immediately sent their big-to-big post double. Hicks, the trailing big in secondary, cut hard from his top-of-the-key position to receive a Meeks pass for an open dunk. Hockey assist to Woods, and a great job of attacking Virginia’s post-trapping scheme with a well-timed dive to to rim.
  19. After Britt missed a secondary-break corner 3 that was created by a Berry-Bradley pick-and-roll, Jackson crashed the glass to tip in the miss for his only second-half hoop.
  20. Maye picked up a 3-second violation while trying to establish deep post position against an undersized Devon Hall (playing the 4 in UVa’s small-ball formation). This is, of course, rarely called, and is the cost of doing business in the secondary break/Roy Williams system.
  21. Following a Perrantes drive and miss, Virginia’s floor balance was uncharacteristically out of sync (this too-frequently happens to UNC, too, following Berry’s drives). This enabled Jackson to push it himself following a defensive rebound and hit Hicks for a primary-break dunk. Hicks flew down the floor on this play, simply out-running the Virginia bigs. The ability of Carolina’s starting wings (Pinson/Jackson) to defensive board and push the pace themselves is turning into a huge weapon for the Heels.
  22. Running the same 1-4 quick-hitter set that resulted in his earlier “and’1,” Jackson curled off another Hicks screen, but this time missed the floater in the paint. Using this set more has been a nice adjustment that takes advantage of Jackson’s skill-set/ability as a curler.
  23. In another secondary set, Hicks, rather than receiving the reversal pass from Berry, set a screen for Pinson to curl off of. This allowed Pinson to get into the paint off the dribble and finish a contested scoop shot at the rim. Pinson’s ability to penetrate and finish at the rim has obviously given the Heels’ offense a whole new dimension lately.
  24. After throwing a secondary-break pass to Jackson from the top of the key, Meeks followed his pass to set a ball screen on the left wing. Jackson tried to split the Virginia hard hedge, resulting in a ball-handling turnover. Again, no huge issues here—just the cost of doing business in the secondary break.
  25. Pinson hit Maye on the right block with a secondary-break entry pass, then the Cavs came immediately with their big-to-big double. Maye quickly found an alertly-cutting Jackson, who missed a layup that he’ll generally finish. Meeks, however, was in perfect position for a tip-dunk—more evidence of how good offensive ball/player movement sets up Carolina’s elite offensive rebounding game.
  26. On another right-block entry from the right wing, Pinson got the ball to Meeks in deep post position (too deep to double). Meeks turned immediately and banked in a short jump hook to cap off his 13-point second-half performance.

As seen in the recap above, Carolina used a variety of secondary break sets to create early offense against Virginia. It also mixed in a couple of opportunistic primary breaks off of live-ball turnovers or defensive boards by its wings/blocked shots by its bigs. While this game was undisputedly played at Virginia’s pace (59.5 possessions—only the seventh game of the 14-year Williams era played below 60 possessions; UNC’s won all seven), the Heels were still able to create their share of early offense. In an average game, UNC uses about 55% of its possessions in the first 10 seconds (down from a Williams-era average of about 60%). That dropped to 44% on Saturday night. But, as discussed, Carolina’s impressive half-court efficiency this season (particularly in the possession-length sweet spot of 18-24 seconds) has enabled it to win both fast and slow. That combination of early-offense and half-court efficiency figures to make this Tar Heel team an especially tough out in March.





UNC vs. Virginia: Tempo-Free Season Box Scores

UNC vs. Virginia: Tempo-Free Season Box Scores

No big game can ever have too many previews, so…

Here’s a look at tons of pace-adjusted stats for the teams and players that will collide tomorrow at 8:15 in the Dome of Dean. Specifically, the team stats (besides pace and offensive efficiency) are per 70 possessions. The team’s percentile rank among the 351 teams in D1 are also given.

Team And Opponent Stats

Stat    UNC          UVA Opp              UVA         UNC Opp      
Pace     74.4   92%                        61.1   0%               
OffEff  117.3   99%     89.4  99%         110.8  88%     96.5   80%
2P%      52.4   79%     44.1  92%          53.4  87%     47.1   70%
3P        6.9   40%      6.8  66%           7.6  60%      7.9   24%
3PA      18.7   26%     21.0  46%          19.6  36%     22.7   19%
3P%      37.1   74%     32.3  85%          38.9  90%     34.9   49%
eFG%     53.4   78%     55.0  94%          55.0  90%     53.4   60%
FT       15.5   74%     12.9  70%          11.2   5%     11.8   88%
FTA      22.1   75%     18.8  67%          15.7   2%     16.5   92%
FT%      70.1   53%     68.5  73%          71.5  62%     71.3   32%
P        82.1   99%     62.6  99%          77.6  88%     67.5   80%
OR       14.7  100%      7.8  99%          10.1  48%      8.6   90%
R        41.1  100%     32.7  88%          37.3  87%     28.9  100%
A        17.1   98%     10.7  96%          16.7  96%     11.2   88%
B         3.1   44%      2.8  80%           4.8  89%      4.0   10%
S         6.9   75%      5.9  54%           6.5  62%      5.9   57%
PF       16.6   88%     17.0  11%          18.5  56%     18.8   47%
TO       11.5   88%     14.8  87%          11.4  89%     13.5   60%
No surprises here. Ours is the better offense, theirs the tougher D. The numbers say they 
should out-block us, but we have a shot at getting some steals and deflections and they
shouldn't exactly parade to the free throw line.

Our offensive boards should be a clash of titans (and could be quite busy, given the 2P% 
and block stats of their D vs. our O); in theory we should own the glass at their offensive 
end, but when teams don't strategically cede DR's to us to control our transition game 
they often have second-chance success.

We can't count on fast-break points or Berry/Jackson long bombs. We have to solve the
Bennett mystery and execute our formula of paint and 2nd-chance points.

Player stats, presented in a pet normalization of mine that I introduced in my first post about
Theo's extraordinary stat line. These are the averages and shooting totals the players would have
if they played 30 mpg for 35 games at a pace of 70 with 35 rebounds per game at each end. It's
intended to paint a very intuitive picture of what a starter who produced like the player in
question would look like. Stats are divided into two rows so as to fit on this page.

UNC Players
Name              Ht    Wt   Class  G   MPG       
   2P%   3P  3PA  3P%   eFG%  FT   FTA  FT%          P     OR   R     A    B    S    PF   TO 
Justin Jackson    6-8   200  Jr     27  31.4      
   51.9  82  212  38.9  54.8   82  108  76.1         16.8  1.2   4.4  2.3  0.2  0.5  1.3  1.4
Joel Berry        6-0   195  Jr     25  29.6      
   51.2  83  199  41.9  57.6   87  102  85.5         14.4  0.4   3.1  3.9  0.1  1.5  2.2  2.0
Kenny Williams    6-4   175  So     26  23.7      
   52.5  44  129  33.8  51.4   31   48  63.3          7.5  1.6   4.0  2.6  0.4  1.1  1.8  1.4
Kennedy Meeks     6-10  260  Sr     27  23.6      
   53.0   0    0   0.0  53.0   91  147  61.7         15.2  4.5  11.3  1.4  1.1  1.2  2.8  1.6
Isaiah Hicks      6-9   235  Sr     26  23.4      
   61.0   0    0   0.0  61.0  128  159  80.4         15.3  2.4   6.7  1.4  0.9  0.4  3.6  2.1
Nate Britt        6-1   175  Sr     27  19.7      
   39.0  36  105  33.9  44.0   30   49  61.5          7.1  0.5   2.8  3.7  0.1  1.7  2.7  1.6
Theo Pinson       6-6   205  Jr      8  17.5      
   59.3  21   78  27.3  53.9  100  142  70.0         11.2  2.3   8.3  5.5  0.2  1.6  2.4  0.8
Tony Bradley      6-10  235  Fr     25  15.0      
   56.5   0    0   0.0  56.5  144  229  62.8         14.8  5.8  10.7  1.3  1.3  0.5  3.4  1.4
Luke Maye         6-8   230  So     22  13.7      
   50.6  26   69  38.1  52.0   40   79  50.0         11.1  3.4   7.9  2.4  0.5  1.0  3.8  1.9
Brandon Robinson  6-5   160  Fr     27   8.9      
   41.9  29  112  25.9  40.5   71  104  68.0          7.6  1.2   4.6  2.6  0.4  0.9  2.8  1.4
Seventh Woods     6-2   175  Fr     27   8.7      
   37.5   8   42  20.0  36.0   89  156  56.8          6.9  0.7   5.1  5.1  0.2  1.9  2.4  4.2

UVA Players
Name              Ht    Wt   Class  G   MPG       
   2P%   3P  3PA  3P%   eFG%  FT  FTA  FT%          P     OR   R    A    B    S    PF   TO 
London Perrantes  6-2   192  Sr     25  31.7      
   47.5  77  196  39.1  53.2  77   96  79.4         13.8  0.4  3.4  4.2  0.1  0.7  1.1  1.9
Isaiah Wilkins    6-7   230  Jr     25  28.5      
   57.3   7   11  57.1  58.8  56   78  71.7          9.1  3.3  7.9  1.8  1.7  1.3  2.2  1.5
Devon Hall        6-5   209  Jr     25  26.4      
   45.7  44  114  38.7  50.0  64   81  79.5         11.2  0.7  5.6  2.5  0.1  0.7  2.6  1.1
Marial Shayok     6-5   213  Jr     25  21.6      
   49.7  31  101  31.1  49.0  79  103  76.1         15.3  0.6  4.3  1.8  0.5  1.6  2.2  1.9
Darius Thompson   6-4   196  Jr     25  20.2      
   57.7  46  135  33.9  54.7  39   60  64.0         10.7  0.3  3.1  4.3  0.5  1.6  1.7  1.9
Jack Salt         6-11  110  So     25  18.5      
   57.1   0    0   0.0  57.1  39   79  50.0          7.1  2.9  7.1  0.8  1.1  0.6  5.0  1.3
Kyle Guy          6-3   165  Fr     25  17.4      
   44.4  95  196  48.6  58.5  64   84  76.7         15.1  0.3  2.8  2.4  0.1  0.8  2.1  1.1
Mamadi Diakite    6-9   195  Fr     23  11.7      
   65.5  13   42  30.0  62.3  46   92  50.0         11.0  2.4  7.0  0.6  3.8  0.7  6.1  0.8
Jarred Reuter     6-7   243  So     24  11.3      
   60.8   0    0   0.0  60.8  70   86  81.0         11.6  2.9  7.7  1.9  0.2  0.6  4.2  2.7
Ty Jerome         6-5   190  Fr     25  11.0      
   81.8  73  165  43.9  71.4  40   56  71.4         11.5  0.1  3.2  4.1  0.2  1.0  4.1  2.6

Berry and Jackson’s 132 treys put them in a three-way tie for 29th-most by a duo in the country. Marcus Keene and Braylon Rayson of Central Michigan are the maddest bombers with 172. Clearly we need somebody handcuffed to Kyle Guy as soon as he enters the game.

Hopefully we’ll get our first extended look at the Pinson-Hicks effect. My own play-by-play charting has Isaiah at 23 P, 8 R, and 7.3 FTA per 30 minutes on 63 & 91 shooting with Theo vs. 16, 7, and 4.5 on 61 & 79 without him.

Three of their best rebounders are foul-prone. Hopefully that comes into play late in the game. Early would be fine, too.

I have Kennedy as tied for 14th-best rebounder in the country and Tony tied for 24th. There are reasons why we probably won’t see significant minutes for those two together with Joel, Theo, and Justin, but it would be oh-so-interesting to see how that would work.