Browsed by
Tag: UNC history

Lawson ’09 vs. Berry ’17

Lawson ’09 vs. Berry ’17

Although Joel Berry’s been great this year, his season still pales in comparison to Ty Lawson’s sublime 2009 campaign—the G.O.A.T. point-guard statline in Carolina history (with apologies to a couple of Phil Ford seasons, Kenny Smith in ’87, Raymond Felton in ’05, etc.). He does stack up quite favorably to Lawson in some key categories, while falling well short in some others.

Let’s break it down with a series of side-by-side comparisons for: I.) Shooting/Scoring; II.) Passing/Turnovers; III.) Defense; and IV.) On-Court Impact.

I. Shooting/Scoring Comparison

  • From a pure scoring volume and efficiency standpoint, Berry ’17 and Lawson ’09 are nearly indistinguishable. It’s how they get their points where the differences lie.
  • Lawson was a better and (significantly more) frequent close finisher than Berry. He attempted nearly 50% more shots / 40 minutes at the rim than Berry (5.66 vs. 3.95), and also made a higher percentage (62.4% vs. 58.8%). For each point guard, most of that close offense was created off the bounce. Factoring in Lawson’s FTA Rate in ’09—over twice as high as Berry’s this season (a little closer if you look at FTMade Rate since Berry’s at 91.2% vs. “only” 79.8% for Lawson)— and his ability to finish through contact (nearly quadruple the number of “and-1s” / 40), and it’s clear that he was the vastly superior scorer at the rim.
  • While Lawson’s better at the rim, the edge at the other two scoring levels (mid-range and behind the arc) would probably go to Berry ’17. Though Lawson made a higher percentage of his 3s in ’09 (47.2% vs. 42.6%), Berry’s attempting nearly twice as many from behind the arc per-40. Each point guard was super-efficient from the top of the key, and most prolific from the right wing. Lawson, in very limited attempts, was also money from the corners. ’09 Lawson was a more dangerous transition threat from behind the arc (though, again, Berry’s shoots transition 3s much more frequently), and both were deadly off the dribble and in the half-court.
  • While neither point guard made his living with the floater (each was more comfortable pulling up for a jumper or (especially in Lawson’s case) getting the whole way to rim), Berry was more efficient with that shot. Both point guards were lethal on mid-range (10-20′) pull-up jumpers.

II. Passing/Turnover Comparison

  • The biggest differentiator between Lawson ’09 and Berry ’17 was in the passing metrics. Lawson’s assist and potential close assist rates were significantly higher than Berry’s this year. He also created a higher percentage (relative to all potential assists) of open shots for his teammates. Despite creating more and better opportunities for others, Lawson was able to maintain a lower rate of turnovers / 40 than Berry. Combining those two factors, Lawson ’09 had more than double the A:TO (factoring in FT assists) of Berry ’17. For a point guard, that’s obviously a huge, glaring advantage.
  • Each point guard had a very similar turnover distribution. Berry commits passing turnovers at a higher rate (by over a half-turnover / 40), but all other turnover categories look nearly identical.
  • Lawson also created drive-and-kick 3-pointers at nearly triple the rate of Berry. Having wing snipers like Wayne Ellington and Danny Green waiting to catch and fire helped here. But Lawson was also better at getting into the paint to create for others (in addition to himself).
  • Although I didn’t include this data in the table, each point guard had a similar post-entry passing profile. Lawson threw 10.6 post entries / 40 with a Success:Failure (made FGs + fouls / missed FGs + TOs) of 0.90 in ’09. Berry’s currently at 8.6 and 0.79 in those categories. Slight advantage Lawson, but having Tyler Hansbrough in the post is certainly a nice luxury for an entry passer.

III. Defensive Comparison

  • In the early part of the season (through Maui), this is the one area in which I would have given the clear advantage to Berry. His Stop% was up in the low 70s through the first half-dozen games, and he was applying consistent ball pressure to fuel Carolina’s 22 defense (and set up its preferred wing overplays/denials). But, post-ankle injury and illness, Berry has been a significantly less disruptive defensive force. Fatigue’s been an issue, too, as the Heels demand so much of Berry on both ends in big games.
  • The two point guards have been equally disruptive (as measured by forced turnovers and deflections), but Lawson was better at denying opponents scoring opportunities (in large part due to keeping them out of the paint a little better than Berry does). ’09 Lawson allowed a couple fewer FGAs and points per-40 compared to ’17 Berry.
  • Though Lawson’s defensive consistency was vastly improved by his junior season, it was still somewhat sporadic. But, when engaged and motivated (see the ’09 national championship game), it’s hard to deny that he could be a disruptive defensive force and lockdown on-ball defender. This category’s close (with plenty of time for Berry to rewrite the script), but I’d give the slight edge to Lawson.

IV. Plus/Minus/On-Court Impact Comparison

  • Each point guard had a huge and profound offensive on-court impact in his respective season. The ’09 Heels were also slightly better on defense with Lawson on the court, while the ’17 Heels (especially in ACC play) have been significantly worse on that end in Berry’s minutes. This is partially a tribute to how well and hard the Carolina bench units (generally some combo of Woods/Britt/Robinson/Maye/Bradley, plus a starter or two) have defended. It’s also probably an artifact of the noisiness and general unreliability of +/- data—especially in a smaller (half-season) sample in Berry’s case.
  • Suffice it to say, each point guard made his team better. Though, again, I’d give ’09 Lawson the advantage for on-court impact (assuming that quality of back-up PGs—SR Frasor/FR Drew II in ’09 vs. SR Britt/FR Woods in ’17—was roughly equal between the seasons).

In terms of pure scoring ability/efficiency, Berry has been downright Lawsonian this season. He does it a bit differently (more from behind the arc, less at the rim), but just as effectively. It’s the other areas of point guard play (play-making and ball protection, primarily), however, that made Lawson’s 2009 campaign such a historically great one, and have separated it from what Berry’s accomplished so far in 2017.

Jackson & Berry: Best-Shooting UNC Duo Ever?

Jackson & Berry: Best-Shooting UNC Duo Ever?

Justin Jackson and Joel Berry combined to make 6-of-11 3-pointers in the big home win over Florida State. That’s nothing new, though—that combo is shooting 41.5% (88-212) from behind the arc this season on a healthy 11.8 attempts per game (accounting for the two games that Berry missed). So how does that compare to the greatest Carolina 3-point shooting tandems of all-time?

To answer that question, let’s use Points Above Replacement Shooter (PARS), a metric that combines shooting efficiency and shooting volume. It assumes a replacement-level 3-point shooter makes 30.0% of his shots and, unlike here where we used PARS/1,000 minutes, we’ll use PARS / game for this analysis. All 3-point attempts per game are pace-adjusted.

As seen in the table, Berry and Jackson are currently third on UNC’s all-time list for combined PARS / game for a pair of teammates. While it will be difficult to maintain their lofty percentages as the schedule continues to intensify, it’s a safe bet that this duo will remain in the top 5 on this list all season. We’ll keep an eye on this leaderboard as the season progresses, but it’s safe to say that the Berry/Jackson combo has exceeded even the most optimistic Tar Heel fan’s expectations in terms of 3-point shooting.

I’m charting the exciting win over the Seminoles this evening, and will be posting a game story at some point this weekend. So stay tuned for that.

The Best (and Worst) UNC Performances vs. Wake Forest

The Best (and Worst) UNC Performances vs. Wake Forest

Similar to this piece on NC State, let’s take a look back at the Carolina-Wake rivalry since Roy Williams has been back in town. While the Heels have certainly gotten the better of the Deacs during the Williams era, it hasn’t been nearly as one-sided as the series with the Wolfpack. UNC is 12-5 against Wake Forest under Williams, and they’ve exceeded their expectations (based on the difference in seasonal KenPom efficiency margins between the teams) in eight of those contests. On average, the games have gone just as predicted, with the Heels outperforming their baseline by just a half-point per game (unlike the nearly 4-point per game difference versus NC State).

3 Best Williams-Era Performances vs. Wake Forest

1. @UNC 105, Wake Forest 72: February 22, 2014

  • Adjusted seasonal efficiency margin difference between teams: +13.5
  • Adjusted game efficiency margin: +40.07
  • Overperformance Score: +26.57
  • Six Heels scored in double-digits, as UNC posted a sweet 57.4/73.3/82.2 shooting split (FG%/3Pt%/FT%). Leslie McDonald hit 5-of-6 3s, while Marcus Paige connected on 3-of-5. In the paint, Brice Johnson and Kennedy Meeks combined to make 13 of 16 field goals.

2. @UNC 104, Wake Forest 67: February 10, 2007

  • Adjusted seasonal efficiency margin difference between teams: +24.5
  • Adjusted game efficiency margin: +41.05
  • Overperformance Score: +16.57
  • Games 2 and 3 on this list were from within a 2.5-week stretch during the 2007 season, when UNC beat an overmatched Wake team by a combined 65 points (192-127). In this one, the second game of the season series, Reyshawn Terry scored 23 points with a virtually flawless shooting line (7-8 from the field, 3-3 on 3s, 6-6 from the line). The Heels’ starting frontcourt of Terry/Brandan Wright/Tyler Hansbrough made 18 of its 21 shots from the field. No Carolina player logged more than 22 minutes and 14 Heels scored, as the bench was used early and often in this easy win.

3. UNC 88, @Wake Forest 60: January 24, 2007

  • Adjusted seasonal efficiency margin difference between teams: +24.5
  • Adjusted game efficiency margin: +40.71
  • Overperformance Score: +16.21
  • Once making the adjustments for pace and game location, the 2007 games against Wake Forest were almost identical in terms of performance above expectation. While UNC’s frontline dominated the Deacs in Chapel Hill, it was the freshman backcourt that starred in Winston-Salem. Ty Lawson and Wayne Ellington combined for 33 points on efficient 13-of-20 shooting (including 4-9 on 3s). The Heels’ stingy defense held WFU to 32.2% shooting from the field, while forcing 19 turnovers. For Wake, only Kyle Visser (16 points, 5 rebounds) reached double-digits.

3 Worst Williams-Era Performances vs. Wake Forest

1. Wake Forest 82, @UNC 69: January 20, 2010

  • Adjusted seasonal efficiency margin difference between teams: -0.8
  • Adjusted game efficiency margin: -24.28
  • Underperformance Score: -23.48
  • While Wake Forest had a slightly better KenPom ranking then the Heels in 2010 (No. 58 vs. No. 61), the metrics certainly didn’t predict a 13-point road victory for the Demon Deacons. UNC’s backcourt was badly outplayed in this one: Drew/Ginyard/Strickland/McDonald combined for just 20 points on 7-30 shooting (4-16 on 3s), while Ish Smith and C.J. Harris had 40 points on 15-of-28 shooting. As a team, UNC made just 6 of its 26 3-point attempts.

2. @Wake Forest 95, UNC 82: January 15, 2005

  • Adjusted seasonal efficiency margin difference between teams: +7.9
  • Adjusted game efficiency margin: -11.71
  • Underperformance Score: -19.61
  • A couple of 14-1 (3-0 ACC) teams squared off in this high-profile showdown in Winston-Salem. In a battle of star point guards, Chris Paul (26 points on 18 FGAs, 8:1 A:TO, 5 steals) outplayed Raymond Felton (16 points on 18 FGAs, 5:2 A:TO, 3 steals). With Paul orchestrating Skip Prosser’s pick-and-roll heavy offense like a maestro, the Deacs posted an offensive efficiency of 123.4 They also knocked down all 32 of their free throws, which never hurts.

3. @Wake Forest 73, UNC 67: January 5, 2014

  • Adjusted seasonal efficiency margin difference between teams: +13.5
  • Adjusted game efficiency margin: -3.04
  • Underperformance Score: -16.54
  • Even on the road, the Heels were a substantial favorite over Jeff Bzdelik’s Demon Deacons, who would finish 17-16 (6-12 ACC) with a KenPom of 118. Once again in a loss to WFU, UNC’s guards let it down, as Paige and McDonald combined to make just 6 of 25 shots (including 2 of 15 3s). They also committed seven turnovers with only seven assists. The Heels lost despite an overwhelming 53-34 rebounding advantage in which UNC controlled both the offensive (OR%: 48.0) and defensive (DR%: 78.4) backboards.

Perhaps the most memorable Carolina-Wake game of the Roy era didn’t crack either of these lists. It was the triple-OT classic in 2004—Williams’s first season back—that the Demon Deacons won 119-114. Jawad Williams and Felton logged 50 and 48 minutes, respectively, but their double-doubles (17 points/12 rebounds for Williams; 22 points/11 assists for Felton) weren’t enough to secure the win. Wake had seven players in double-digits, including four with at least 18 points. The Deacs were led by 24 from Eric Williams and 20 from Justin Gray.

UNC vs. Expectations against Wake Forest

On average, Carolina’s been expected to win by 14.3 points / 100 in its 17 Williams-era games against Wake Forest. The Heels have actually won by an average of 14.8 points / 100 in those contests, an overperformance score of +0.50. UNC has exceeded its baseline expectations in 8 of those 17 games. Let’s break down those numbers a little further, splitting them out by venue and WFU coach:

  • In Chapel Hill:
    • Record: 6-2
    • Underperfomance Score: -0.90
    • Exceeded expectations: 3 times in 8 games (37.5%)
  • In Winston-Salem:
    • Record: 6-3
    • Overperformance Score: +1.75
    • Exceeded expectations: 5 times in 9 games (55.6%)
  • Against Skip Prosser:
    • Record: 4-2
    • Average adjusted seasonal efficiency margin difference: +12.0
    • Average adjusted game efficiency margin: +16.29
    • Overperformance Score: +4.31
    • Exceeded expectations: 4 times in 6 games (66.7%)
  • Against Dino Gaudio:
    • Record: 2-2
    • Average adjusted seasonal efficiency margin difference: +7.4
    • Average adjusted game efficiency margin: +2.26
    • Underperformance Score: -5.09
    • Exceeded expectations: 1 time in 4 games (25.0%)
  • Against Jeff Bzdelik:
    • Record: 4-1
    • Average adjusted seasonal efficiency margin difference: +20.2
    • Average adjusted game efficiency margin: +21.38
    • Overperformance Score: +1.22
    • Exceeded expectations: 2 times in 5 games (40.0%)
  • Against Danny Manning:
    • Record: 2-0
    • Average adjusted seasonal efficiency margin difference: +22.1
    • Average adjusted game efficiency margin: +20.55
    • Underperformance Score: -1.55
    • Exceeded expectations: 1 time in 2 games (50.0%)
Justin Jackson’s Development as a Shooter

Justin Jackson’s Development as a Shooter

After making just 1.17 3-pointers / 40 minutes at a rate of 29.7% through his first two collegiate campaigns, Justin Jackson has improved those numbers to 3.17 and 38.7% after 17 games of his junior season. Both his 3-point volume (per-40) and efficiency have significantly increased season-over-season—a testament to the feedback he received from NBA scouts, and the hard work he put in all summer. But just how rare is it for a Tar Heel to make the type of perimeter improvement that Jackson has this year? Let’s dig deeper into the data to answer that question.

For the sake of this analysis, we’ll look at two primary numbers: 1.) 3-point percentage (efficiency) and 2.) pace-adjusted 3-point attempts per 40 minutes (usage/volume). In Jackson’s case, his season-by-season marks in these metrics are:

Using these two concepts of 3-point proficiency, we can create a metric called Points Above Replacement Shooter per 1,000 Minutes—or PARS/1000. Since 2002, the national average has hovered between 33.9% and 35.1%. The average for those seasons in 34.5%. For the sake of this analysis, we’ll consider a “replacement-level” shooter to be one who connects on 30% of his 3s. While that number is somewhat arbitrary, it doesn’t make a difference for the sake of the rankings/ordering. Alternatively, we could use the concept of an average shooter here rather than a replacement-level one, too. 1,000 minutes is used because a.) it’s a round number, and b.) it’s about the number of minutes an average collegiate starter will play in a season (35 games at 26.6 MPG).

Jackson’s PARS/1000 this season can be computed as follows:

  • 7.62 3-pointers per 40 minutes –> 190.5 3-pointers per 1,000 minutes
  • a replacement-level shooter would score 171.5 points on those attempts (190.5 * 0.300 * 3)
  • Jackson would score 221.2 points on those attempts (190.5 * 0.387 * 3)
  • thus, Jackson has a PARS/1000 of 49.7 (221.2 – 171.5)

As a sophomore, Jackson’s PARS/1000 was actually negative (since he fell slightly below the 30% mark) at -2.5. Thus, his season-over-season change in this metric was (49.7 – (-2.5)) = +52.2. So how does that mark compare historically to other UNC shooters? Let’s take a look at the table below:

The major assumption used here is that a player must log at least 10 minutes/game in each of the seasons for which the PARS/1000 change is being measured. This will ensure that we’re only including players who were in the rotation in both years of the comparison. It excludes the freshman-to-sophomore leaps of some memorable UNC shooters like Hubert Davis, Donald Williams, Shammond Williams (and Kenny Williams this season), as well as Wes Miller’s sophomore-to-junior jump and Pearce Landry’s junior-to-senior one. If we were to raise the minutes/game threshold to 15, Jackson would actually climb to No. 2 on the list behind Okulaja (Hairston and Curry played 13.0 MPG as freshmen, McDonald played 10.3, and Graves (prior to his suspension) played 11.2 as a sophomore).

The players on this list can be split into three primary buckets:

  1. Those who improved both their 3-point volume and efficiency significantly. This includes (at least to date) Jackson and Berry this season. It also includes Britt’s jump between his freshman and sophomore seasons (a time period in which—stop me if you’ve heard this one—he actually changed shooting hands!). Others on this list include McDonald, Okulaja, Graves, Stackhouse, Cota, Paige, and Jawad Williams.
  2. Those who improve their 3-point efficiency significantly. This group includes Hairston, Curry, Felton, Calabria, Bullock (twice!), Davis, Boone, Green, Lawson, and Donald Williams. For some on this list (notably, Curry, Calabria, and FR-to-SO Bullock), 3-point volume actually went down (as part of a volume-efficiency trade-off). For some (Felton, Boone, Green, D. Williams) it went up slightly. For some (Hairston, SO-to-JR Bullock, Davis, Lawson), it stayed nearly the same. In all cases, it was the increase in 3Pt% rather than an increase in the volume that was driving the improvement.
  3. Those who increase their 3-point volume significantly while maintaining a high percentage. This is the rarest type on the list, including just Fox, Bucknall, Scott, and Noel. Scott moved from point guard as a freshman to off the ball (alongside Felton) as a sophomore. Noel’s volume increased to help compensate for the losses of Felton, McCants, Scott, and Jawad/Marvin Williams from the ’05 champs. Fox and Bucknall likewise stepped up as upperclassmen to help fill perimeter voids (the loss of Ranzino Smith in Bucknall’s case, and the losses of Lebo/Bucknall in Fox’s).

In the offseason, I’ll play around a little more with this data (career PARS/1000 leaders, categorizing UNC’s historical 3-point shooters into buckets by career shooting progression, etc.). For the remainder of this season, the ability of Jackson and Berry to maintain their places on this list will help to determine just how special Carolina’s season ends up being. The Heels will need both to continue being high-volume, high-efficiency options from behind the arc.

An interesting aside about Jackson’s 3-point shooting in 2017: he continues to be significantly more effective from the left side of the floor than the right side from the perimeter (and as a penetrator too, actually; his floater percentage from the left paint is much better than from the right paint). Here are his 3-point splits by shot location:

  • right corner: 25.0% (2-8)
  • right wing: 20.0% (6-30)
  • top of the key: 44.4% (8-18)
  • left wing: 51.3% (20-39)
  • left corner: 45.5% (5-11)

Or, summing those up: 21.1% from the right side (on 38 attempts) and 50.0% from the left side (on 50 attempts). The top of the key was right about in the middle percentage-wise (and, you know, shooting-wise) until Jackson hit 2-of-2 there against NC State to bump that percentage until the mid-40s. Just something to keep an eye on as the season progresses.


Slow ACC Start? Nobody Panic!

Slow ACC Start? Nobody Panic!

Confession time: I started researching/writing this piece over the weekend before Carolina’s historic 51-point drubbing of NC State. Its thesis seemed much more relevant at that time. The data’s still interesting (although it just confirms conventional wisdom, rather than debunking it), though, so let’s proceed as if the Heels are still off to a “slow” start.

Obviously Carolina’s ACC opener against Georgia Tech was less than ideal. As I wrote here, it was (at the time) the sixth-worst loss of the Roy Williams era relative to expectations. And, depending on how the Jackets’ campaign progresses, it could end up looking worse and worse by the end of the season (in fact, it’s currently moved up (down?) to No. 4 on the list of worst losses based on Georgia Tech’s tepid efforts against Duke and Louisville). But certainly a slow start to the conference season hasn’t doomed the Heels in past years.

To prove this point, let’s refer to the table below. For each of Roy Williams’s 13 (full) seasons in Chapel Hill, it lists the performance above expectation for each of six season segments: 1.) the non-conference schedule; 2.) the first three games of the ACC season; 3.) the first half of the ACC season; 4.) the second half of the ACC season; 5.) the ACC Tournament; and 6) the NCAA Tournament (or, in the case of 2010, the NIT). Performance above (below) expectation is the number of points per 100 possessions better (worse) that UNC plays relative to its baseline. The baseline (for a single game) is computed by using the difference in KenPom seasonal adjusted efficiency margins between the two teams.

Table 1: UNC’s Performance Above Expectation by Season Segment in the Roy Williams Era

So what are some key takeaways from the table? Let’s go through it season segment-by-season segment:

  • The Heels have generally had above-average non-conference showings in the Williams era. Even the 2010 team boasted some impressive out-of-conference performances (notably, wins over Ohio State and Michigan State, a blowout of Hassan Whiteside-led Marshall, and a close loss at loaded Kentucky).
  • Since strong showings in his first two seasons (the 2005 ACC start was especially stout: UNC beat Virginia Tech 85-51 in Blacksburg, followed by impressive home wins over Maryland (109-75) and Georgia Tech (91-69)), Carolina has really struggled in the first three games of the ACC schedule under Roy Williams. The stretch between 2009 and 2014 was especially underachieving. It included the 0-2 start in 2009 (including a home loss to a mediocre Boston College team), a bad 20-point road loss to Georgia Tech in 2011, the 33-point drubbing in Tallahassee in 2012, an 0-2 start in 2013, and an 0-3 start in 2014 (including bad losses to below-average Wake Forest and Miami teams). The conference starts over the past three seasons haven’t been as bad (relative to expectations), however. This year’s includes both a positive (NC State) and negative (Georgia Tech) outlier, as well as an as-expected performance (close win) at Clemson.
  • While not as poor on average as the “first three games” subset, the performance in first half of the ACC season has not met expectations under Williams. In four separate years (2oo4, 2008, 2010, and 2016), it’s been the worst (relative to expectations) of any season segment. Only in 2007 has it been the best.
  • The second half of the ACC season, on the other hand, has generally exceeded expectations. The ’06 Heels are the canonical example of a team that peaked during the ACC stretch run. The 2012 and 2014 teams were also playing their best basketball late in the conference season.
  • Then comes the ACC Tournament—Ol’ Roy’s favorite cocktail party. Consistent with its coach’s famous disdain for the event, Carolina has tended to underperform in this event. It should be noted, though, that’s there’s a clear split between years 1-9 (-5.82) and years 10-13 (+4.86) with respect to ACCT performance. And, of course, some key injuries (Ty Lawson and John Henson, most memorably) must be accounted for in a sample this small.
  • Finally, the NCAA Tournament: as seen in the table’s bottom line, it’s been the part of the season in which Carolina has been at its best relative to expectations. The 2005 and 2009 championship runs speak for themselves, but the 2008 (even including the Kansas debacle), 2011, and 2016 Tournament runs were also very impressive.

That UNC is playing its best hoops at the right times (end of the ACC regular season into the NCAA Tournament) is certainly a testament to Williams’s greatness and how he manages his rosters and develops his teams. He’s more interested in using early games to cultivate talent (see some of the crazy early-season rotations/lineups that fans sometimes lament), a strategy with double-pronged benefits of developing underclassmen/bench parts and keeping key starters fresh. A quintessential system coach, Williams is also more committed to establishing and perfecting Carolina’s core identity than he is to making dramatic early-season tweaks (e.g., switching ball screens, going small, playing extensive zone, etc.) to win any single (non-critical) game. It’s almost certainly true that Carolina’s early-ACC foes are doing more UNC-specific game-planning for the Heels than the Heels are doing for them. By late-ACC season (and definitely into the NCAAs), however, Williams is far more likely to tweak the system on a game-by-game basis to account for the strengths and weaknesses of opponents. And, in the meantime, UNC is using most of its practice and game time to focus on mastering its core concepts (e.g., the secondary break, the free-lance passing game, its overplaying man-to-man defense and its help-rotation principles, etc.).

So, while all of this analysis might not ease your mind after the Heels lay an(other) early-season egg, rest assured that Roy Williams’s teams will generally be peaking at the right time. Given what a complex calculus it can be to pull off that feat consistently, it’s something Carolina fans should never take for granted with its current coach.

Carolina’s Top (and Bottom) Performances vs. NC State

Carolina’s Top (and Bottom) Performances vs. NC State

Pending any winter-weather catastrophes in the Triangle, Carolina and NC State are slated to renew their rivalry on Saturday in a prime-time 8pm showdown.

And, frankly, ever since Roy Williams’s return to Chapel Hill before the 2003-04 season, calling it a “rivalry” is probably polite to the ‘Pack. Carolina has won 25 of its 28 match-ups against NC State during that timeframe. But, of course, not all of those victories are equal in terms of level of performance (don’t get me wrong: any win over the Wolfpack is OK in my book). Let’s take a look at UNC’s top-three and bottom-three showings versus NC State in the 28 games coached by Williams.

To measure a game’s performance (relative to expectations), I compare its actual efficiency margin (adjusted for venue) to the difference in seasonal (adjusted) efficiency margins between the two teams. All adjusted efficiency numbers are courtesy of For example, so far this season, UNC has an efficiency margin of +25.8 while NC State’s is +12.8. This suggests that the Heels would be expected to outscore State by 13 points / 100 possessions on a neutral floor. So that’s the “expectation” for Saturday night’s game. In a 75-possession game (and using a straight 3.75-point home-court advantage adjustment per Pomeroy), that would equate to a 13.4-point Carolina victory. But, say, in a 65-possession game in Raleigh, it would result in just a 4.6-point UNC win. The overperformance (or underperformance) score can be interpreted as the points / 100 possessions better (or worse) a Carolina outcome was relative to its baseline expectation.

Read more about this metric’s methodology here.

3 Best Williams-Era Performances vs. NC State

1. UNC 95, @NC State 71: February 22, 2006

  • Adjusted seasonal efficiency margin difference between teams: +6.1
  • Adjusted game efficiency margin: +43.36
  • Overperformance Score: +37.26
  • This 24-point road win over NC State (KenPom No. 32) began a remarkable four-game stretch to end the 2006 ACC regular season: the Heels would follow with a 24-point home win over Maryland (58), a 45-point home win over Virginia (82), and a 7-point road win over Duke (2). David Noel had a 25-point, 15-rebound double-double for the Heels, as four starters scored at least 15 points. Point-guard platoon Bobby Frasor and Quentin Thomas, while not scoring, combined for 11 assists and only two turnovers. This would, probably not coincidentally, be the final UNC game of the Herb Sendek era at NC State.

2. UNC 77, @NC State 63: January 26, 2010

  • Adjusted seasonal efficiency margin difference between teams: +1.8
  • Adjusted game efficiency margin: +26.49
  • Overperformance Score: +24.69
  • Even in Roy Williams worst year (by far) in Chapel Hill, his Heels easily handled NC State on two occasions. Two of UNC’s five ACC wins, and two of its three double-digits ones, came at the expense of Sidney Lowe’s Wolfpack. Deon Thompson scored 20 points on 9-of-14 shooting, while Larry Drew II added 18 points with a 7:1 A:TO. Dexter Strickland scored 14 points on only seven FGAs in 17 high-quality minutes off the bench.

3. UNC 71, @NC State 64: February 29, 2004

  • Adjusted seasonal efficiency margin difference between teams: -0.6
  • Adjusted game efficiency margin: +17.06
  • Overperformance Score: +17.66
  • According to the Pomeroy rankings, NC State (9) actually had a slightly stronger team than UNC (11) in Williams’s first year in Chapel Hill. Despite that, the Heels escaped with a pair of close wins: a two-point victory at the Dean Dome, and this more-impressive seven-point win in Raleigh. Playing on Leap Day, each team made eight 3-pointers—but it took NC State more than twice as many long-range attempts (34 to 16) to do so. Rashad McCants led the way with 22 points (including four 3s), five rebounds, three assists, and a pair of steals. Wolfpack star Julius Hodge was held to 13 points on 4-of-11 shooting in 38 minutes.

3 Worst Williams-Era Performances vs. NC State

1. NC State 58, @UNC 46: February 24, 2015

  • Adjusted seasonal efficiency margin difference between teams: +7.9
  • Adjusted game efficiency margin: -26.3
  • Underperformance Score: -34.2
  • This one ranks as the worst home loss of the Williams era. UNC just couldn’t make shots, converting just 37.2% (16-43) of its 2-pointers, 25.0% (3-12) of its 3s, and 55.6% (5-9) of its free throws. BeeJay Anya controlled the defensive paint, blocking six Carolina shots. NC State didn’t exactly put on an offensive clinic in this one, but its super-physical, slow-down strategy turned it into a half-court grinder. With Terry Henderson’s claim that the ‘Pack will be more like the Bad Boy Pistons, will Saturday’s game take on the same slugfest tone? Optimistically, perhaps, NC State (led by lightning-fast PG Dennis Smith, Jr.) is playing at its fastest tempo of the Mark Gottfried era (50th in the nation with an adjusted pace of 72.1 possessions / 40) so may be more interested in running with the Heels than turning it into a half-court donnybrook.

2. @NC State 83, UNC 79: February 3, 2007

  • Adjusted seasonal efficiency margin difference between teams: +20.6
  • Adjusted game efficiency margin: -0.34
  • Underperformance Score: -20.94
  • In the first UNC-NC State rivalry game of the Sidney Lowe era, the Wolfpack came away with a resounding home upset over the Heels. Rather than ushering in a paradigm-shifting new regime, however, this outcome proved to be a total fluke. Carolina would go on to beat Lowe’s teams 10 straight times until he was unceremoniously dismissed following the 2011 season.

3. @UNC 89, NC State 80: February 18, 2009 

  • Adjusted seasonal efficiency margin difference between teams: +20.1
  • Adjusted game efficiency margin: +7.5
  • Underperformance Score: -13.40
  • When one of your three “worst” performances is a mere nine-point victory, you know you’ve had a pretty solid run of dominance. The Heels, who ended the season atop the KenPom rankings, were a heavy home favorite over the 81st-ranked Wolfpack in a game that ended up closer than expected. NC State connected on 11-of-20 3-pointers to keep this one respectable—Javier Gonzalez went 4-4 behind the arc, scoring 18 points on just eight FGAs. The Heels were led by a vintage Ty Lawson ’09 performance: 17 points on nine FGAs with a 9:1 A:TO. Tyler Hansbrough had 27 points on 10-15 shooting from the field. He also made all seven of his FTAs. Random “Psycho T” trivia: he made 72-of-79 career free throws against NC State in nine games—a sizzling 91.1%. Against all other opponents, he shot “only” 78.3% from the charity stripe.

UNC vs. Expectations against NC State

On average, Carolina’s been expected to win by 11.6 points / 100 in its 28 Williams-era games against NC State. The Heels have actually won by an average of 15.4 points / 100 in those contests, an overperformance score of +3.80. UNC has exceeded its baseline expectations in 19 of those 28 games. Let’s break down those numbers a little further, splitting them out by venue and NC State coach:

  • In Chapel Hill:
    • Record: 12-1
    • Overperfomance Score: +1.96
    • Exceeded expectations: 9 times in 13 games (69.2%)
  • In Raleigh: 11-2
    • Record: 11-2
    • Overperformance Score: +7.44
    • Exceeded expectations: 10 times in 13 games (76.9%)
  • Neutral Sites: 2-0
    • Record: 2-0
    • Underperformance Score: -7.93
    • Exceeded expectations: 0 times in 2 games (0.0%)
  • Against Herb Sendek:
    • Record: 6-0
    • Average adjusted seasonal efficiency margin difference: +6.4
    • Average adjusted game efficiency margin: +20.21
    • Overperformance Score: +13.81
    • Exceeded expectations: 5 times in 6 games (83.3%)
  • Against Sidney Lowe:
    • Record: 10-1
    • Average adjusted seasonal efficiency margin difference: +16.8
    • Average adjusted game efficiency margin: +19.90
    • Overperformance Score: +3.10
    • Exceeded expectations: 8 times in 11 games (72.7%)
  • Against Mark Gottfried:
    • Record: 9-2
    • Average adjusted seasonal efficiency margin difference: +9.3
    • Average adjusted game efficiency margin: +8.33
    • Underperformance Score: -0.96
    • Exceeded expectations: 6 times in 11 games (54.5%)

A few insights from the data above:

  1. Under Williams, UNC has been especially dominant on the road in this series.
  2. Relative to expectations, the Heels have struggled in their only two ACC Tournament match-ups against NC State. Of course, no one’s going to complain too much about any postseason victory.
  3. Although NC State was at its (relative) strongest in the Sendek era, his teams performed poorer against the Heels than either Lowe’s or Gottfried’s.
  4. Even though NC State was (relatively) weak in the Lowe days, UNC still managed to exceed its (lofty) expectations in those games (on average).
  5. Although the 2-9 record might not reflect it, NC State has been significantly more competitive with Carolina during Gottfried’s tenure. Even if you throw out the 2015 @UNC outlier (underperformance score of -34.15), the Heels’ overperformance score against Gottfried (+2.36) would still be worse than against either of his two immediate predecessors. When including it, Carolina’s actually (slightly) underperformed relative to expectations against the ‘Pack.
The Worst Performances of the Roy Williams Era

The Worst Performances of the Roy Williams Era

Saturday’s ACC opener in Atlanta against Georgia Tech was obviously a clunker, as the heavily-favored Heels lost by double digits to the Yellow Jackets. While it was clearly Carolina’s worst performance of this season, how did it compare to other forgettable games from Roy Williams’s 13.5 seasons in Chapel Hill?

A few methodological notes: This list is purely statistical in nature. It’s looking at the difference between two components: 1.) the Pomeroy adjusted efficiency delta between the two teams, and 2.) the actual efficiency margin (adjusted for location) of the game. The higher the positive difference between these components, the more UNC exceeded its expectations in a given game. The higher the (absolute) negative difference, the more the Heels fell short of expectations. This analysis is using only the final Pomeroy numbers from a given season (or year-to-date numbers for the current campaign), so does not take into account how much a team’s performance may have improved or declined during the year. It’s also not accounting for things like injuries, suspensions, etc.

So let’s take a look at UNC’s 12 worst performances of the Roy Williams era according to this metric.

1.) Florida State 90, UNC 57: January 14, 2012

  • Adjusted seasonal efficiency margin difference between teams: +8.3 (with positive meaning UNC’s was better; negative meaning it was worse)
  • Adjusted game efficiency margin: -38.5
  • Underperformance Score: -46.8
  • UNC rolled into Tallahassee with a 15-2 record, including wins in its first two ACC contests. Much like the recent Georgia Tech game, the Heels combined terrible 3-point shooting (4-21) with poor ball protection (22 turnovers) in a miserable offensive performance. Kendall Marshall and Harrison Barnes combined for 12 turnovers against only 4 assists. As a team, Carolina made just 45.0% of its free throws (9-20). FSU made 12 of its 27 3-pointers, including 8-of-10 from the otherwise-forgettable Deividas Dulkys.

2.) Santa Clara 77, UNC 66: November 19, 2004

  • Adjusted seasonal efficiency margin difference between teams: +28.5
  • Adjusted game efficiency margin: -12.7
  • Underperformance Score: -41.2
  • This one obviously needs a huge asterisk, as the Heels were without suspended star point guard Raymond Felton for their 2004-05 season opener. Quentin Thomas made the start in his Carolina debut, and combined with fellow PGs Melvin Scott and Wes Miller to score five points (on 2-10 shooting) with an 8:5 assist-to-turnover ratio in 47 minutes. UNC was dominated in the paint in this one, as Santa Clara shot 57.6% inside the arc. The Broncos also shot 29 FTs to UNC’s 24, and out-rebounded the Heels 36-31.

3.) Georgia Tech 78, UNC 58: January 16, 2011

  • Adjusted seasonal efficiency margin difference between teams: +15.5
  • Adjusted game efficiency margin: -21.7
  • Underperformance Score: -37.2
  • Saturday’s game was only the second-worst UNC performance in Atlanta. It was topped (bottomed?) by this dud from the 2010-11 season, notable because it marked the final game that Larry Drew II would start as a Tar Heel. The reins were officially handed to Kendall Marshall after this one, and Drew would transfer out of the program after coming off the bench for four games. He was shut out with just two assists in 21 minutes in this one, but wasn’t alone in his ineptitude. The Heels shot 27.6% from he field and had 18 turnovers compared to just nine assists. Carolina-killer Iman Shumpert (30) and Glen Rice, Jr. (24) combined for 54 points on 59.5% (22-37) shooting, nearly outscoring UNC by themselves.

4.) NC State 58, UNC 46: February 24, 2015

  • Adjusted seasonal efficiency margin difference between teams: +7.9
  • Adjusted game efficiency margin: -26.3
  • Underperformance Score: -34.2
  • This one ranks as the worst home loss of the Williams era. UNC just couldn’t make shots, converting just 37.2% (16-43) of its 2-pointers, 25.0% (3-12) of its 3s, and 55.6% (5-9) of its free throws. BeeJay Anya controlled the defensive paint, blocking six Carolina shots. NC State didn’t exactly put on an offensive clinic in this one, but its super-physical, slow-down strategy turned it into a half-court grinder.

5.) Boston College 85, UNC 78: January 4, 2009

  • Adjusted seasonal efficiency margin difference between teams: +19.6
  • Adjusted game efficiency margin: -14.3
  • Underperformance Score: -33.9
  • After starting the 2008-09 season 13-0, the Heels dropped their ACC opener to a mediocre (final Pomeroy ranking of 74) Boston College team in Chapel Hill. UNC would lose its next ACC game, too (at Wake Forest), before righting the ship to win 13 of its final 14 regular-season conference games. The Heels, of course, also won the national championship in 2009, beating each of its NCAA Tournament foes by double digits. Tyrese Rice, who scored 46 in a 2008 loss in Chapel Hill, had 25 points (on 13 FGAs) and eight assists in this one, getting the better of Ty Lawson. Lawson shot just 3-for-13, with 10 points, four assists, and four turnovers. He would go on to better days, too, on his way to the ACC Player of the Year award.

6.) Georgia Tech 75, UNC 63: December 31, 2016

  • Adjusted seasonal efficiency margin difference between teams: +22.9
  • Adjusted game efficiency margin: -10.6
  • Underperformance Score: -33.5
  • As of today, UNC’s Saturday stinker versus Georgia Tech is the sixth-worst performance of the RoyW era. Depending on how the Heels and Yellow Jackets end the season, this one has the potential to move up or down on this list. Like so many of the other games in this group, bad shooting (41.3% on 2s, 19.2% on 3s) and ball protection (20 turnovers) doomed the Heels here. I’ll be posting a little more about the Heels’ specific failures in this game soon.

7.) UNC 83, Gardner Webb 80: November 19, 2005

  • Adjusted seasonal efficiency margin difference between teams: +30.6
  • Adjusted game efficiency margin: -1.0
  • Underperformance Score: -31.6
  • UNC’s only win among its 12-worst performances, the Heels needed a buzzer-beating David Noel 3 to beat Gardner Webb in the Dean Dome. The Bulldogs would finish the 2005-06 campaign ranked 263rd in the KenPom ratings. UNC’s point guard platoon of Bobby Frasor and Quentin Thomas combined for just two points (on 1-9 shooting) with eight turnovers (although they did have 11 assists). Within its first 13 games, the post-championship ’06 Heels actually accounted for three of the clunkers on this list (see also #8 and #11). In one of Roy’s best coaching jobs, however, he turned this young team into a late-season juggernaut that was routinely dominating ACC foes in the second half of the conference schedule.

8.) Miami 81, UNC 70: January 14, 2006

  • Adjusted seasonal efficiency margin difference between teams: +9.1
  • Adjusted game efficiency margin: -21.4
  • Underperformance Score: -30.5
  • After winning its first two ACC games, the ’06 Heels dropped a home game to Frank Haith’s mediocre (18-16 (7-9 ACC), KenPom #55) Miami squad. The Heels couldn’t keep the Hurricanes’ guards out of the paint, as Guillermo Diaz, Robert Hite, and Anthony Harris combined for 65 points and 29 FTAs (of which they made 24). After dropping the next game to a less-than-stellar (KenPom #82) Virginia team in Charlottesville, Williams would shake up the starting lineup by inserting Wes Miller for Marcus Ginyard. In his first start, Miller’s six 3s led the Heels to a road victory over Florida State, helping to turn the Heels’ season around with his perimeter punch and floor-spacing presence.

9.) Virginia 75, UNC 60: January 31, 2010

  • Adjusted seasonal efficiency margin difference between teams: +2.7
  • Adjusted game efficiency margin: -27.6
  • Underperformance Score: -30.3
  • In Tony Bennett’s first season at Virginia, he brought a bad Cavs team (15-16 (5-11 ACC), KenPom #76) into Chapel Hill and left with a 15-point win. Sylven Landesberg and Sammy Zeglinski combined for 48 efficient points, as the Heels made just 32.4% of its 2-pointers and half of their 22 free throws. Starting posts Ed Davis and Deon Thompson teamed up for just 11 points (3-10 FGs, 5-12 FTs) and seven turnovers, flummoxed by Bennett’s hard big-to-big doubles in the paint.

10.) Pittsburgh 89, UNC 76: February 14, 2015

  • Adjusted seasonal efficiency margin difference between teams: +15.2
  • Adjusted game efficiency margin: -14.9
  • Underperformance Score: -30.1
  • In the second (and final) game of the Isaiah Hicks-Brice Johnson starting frontcourt experiment, Pitt picked apart the Carolina defense, scoring 89 points on 62 possessions (offensive efficiency of 143.5). That included 35 points on 17 possessions (offensive efficiency of 205.9!) when Hicks and Johnson were on the floor together; those lineups, although explosive offensively, were really terrible on the defensive end in 2014-15. They’d improve on that end in 2015-16 (with Johnson becoming more of an assertive defensive 5), eventually becoming the crunch-time frontcourt for the ’16 national runners-up. Pitt had an astounding 30 assists against only five turnovers in this one (UNC’s A:TO was  likewise stellar 25:4), making 69% of its 2-pointers.

11.) USC 74, UNC 59: December 21, 2005

  • Adjusted seasonal efficiency margin difference between teams: +15.0
  • Adjusted game efficiency margin: -14.8
  • Underperformance Score: -29.8
  • In a pre-Christmas trip to the West Coast to battle Tim Floyd’s Trojans, UNC left its game back in Chapel Hill. The Heels had 25 turnovers in this one (7 by Bobby Frasor, 6 by David Noel), as poor point guard play once again doomed them (Frasor and Thomas combined for five points (2-10 shooting) and nine turnovers). USC hit 12 of 21 from behind the arc, including a combined 9-of-13 by Gabe Pruitt and Lodrick Stewart (TWINS!). For Carolina, Tyler Hansbrough made 6 of 7 from the field (his attempts were limited by the USC ball pressure/UNC’s inability to feed the post), while all other Heels made just 14 of 47 (29.8%).

12.) Duke 82, UNC 50: March 6, 2010

  • Adjusted seasonal efficiency margin difference between teams: -15.4
  • Adjusted game efficiency margin: -44.1
  • Underperformance Score: -28.7
  • Duke led 53-26 at the half in Durham, and this one could have probably been even uglier. Kyle Singler, Nolan Smith, and Jon Scheyer combined for 65 points, making 20-21 free throws. The Heels attempted just five 3s (making only one), while converting only 34.9% of their 43 2-pointers. Carolina’s starting guards/wings (Larry Drew II, Marcus Ginyard, and Will Graves) made just 2 of 12 shots, scoring only nine points in their 89 minutes on the court. Ugh. Bad memories. Moving on.

So the point of this exercise wasn’t to totally depress you as Tar Heel fans. It was to attempt to frame the poor performance on Saturday in some historical context. Under Roy Williams, Carolina has had a game that bad about every other year or so. Some otherwise excellent teams (see 2009 and 2012) have laid some huge eggs. It happens.

I’m hoping to break down the recent Georgia Tech loss in a little more detail soon (just returning from holiday travel that has me a bit behind on charting) to determine what was just fluky outlier stuff and what may portend future issues/systemic weaknesses. I’m also planning to use this metric of UNC single-game performance relative to expectations to write some happier pieces—UNC’s best showings of the RoyW era, how UNC improves as the season progresses, and how UNC over-performs expectations in the NCAA Tournament, to name three.


The 10 Best Lineups of the Roy Williams Era

The 10 Best Lineups of the Roy Williams Era

With both the Joel Berry-Kenny Williams-Justin Jackson-Isaiah Hicks-Kennedy Meeks lineup (Efficiency Margin of +61.1 in 85 minutes) and the Berry-Williams-Jackson-Hicks-Tony Bradley one (Efficiency Margin of +64.8 in 28 minutes) off to ridiculously good +/- starts this season, let’s take a look at some historically strong Williams-era combinations.

To qualify for this list, a lineup needed to play at least 50 minutes together in a given season. These aren’t strictly ranked by raw net efficiency; it’s a combination of +/-, total usage, and some “eye test” considerations (and probably some personal biases of mine, too, if we’re being honest). We’ll count them down from No. 10 to No. 1. No current combos are included on the list, although, if they can maintain anywhere close to their current pace, the two mentioned above are serious candidates to join it.

Honorable Mention:

  • Joel Berry-Marcus Paige-Justin Jackson-Isaiah Hicks-Brice Johnson (2015-16)
    • 142 minutes, Offensive Efficiency: 127.6, Defensive Efficiency: 103.8, Efficiency Margin: +23.8
    • After being terrible as a defensive combination in 2014-15, the Hicks-Johnson frontcourt improved to average the next season. Those lineups were always explosive offensively, though, and this became RoyW’s go-to crunch-time lineup (along with the Pinson-at-the-4 small-ball unit) during last year’s postseason run (it was the quintet on the court for the final 5:21 against Villanova, outscoring the ‘Cats 17-10 in that time).
  • Ty Lawson-Wayne Ellington-Danny Green-Brandan Wright-Tyler Hansbrough (2006-07)
    • 71 minutes, Offensive Efficiency: 112.2, Defensive Efficiency: 81.8, Efficiency Margin: +30.4
    • An all-underclassmen lineup (3 FR/2 SO), but five guys who stuck around the NBA for a long time. Green was still raw and mistake-prone as a sophomore, but this was a successful and disruptive defensive unit.
  • Kendall Marshall-Leslie McDonald-Harrison Barnes-John Henson-Tyler Zeller (2010-11)
    • 67 minutes, Offensive Efficiency: 122.3, Defensive Efficiency: 97.3, Efficiency Margin: +25.1
    • This lineup played way fewer minutes than the 2010-11 starting 5 (Dexter Strickland in for McDonald), which logged 321. Though it was a bit worse defensively than the starters (97.3 vs. 92.8), the spacing benefits of having a shooter at the 2 more than made up for it on the offensive end (122.3 vs. 111.3).

10.) Ty Lawson-Wayne Ellington-Reyshawn Terry-Brandan Wright-Tyler Hansbrough (2006-07)

  • 279 minutes, Offensive Efficiency: 121.0, Defensive Efficiency: 94.3, Efficiency Margin: +26.7
  • The 2006-07 starting 5, this lineup was solid on both ends. This team was super-deep, of course, and didn’t really miss a beat when subbing in Marcus Ginyard or Danny Green for one of the wings.

9.) Marcus Paige-J.P. Tokoto-Justin Jackson-Isaiah Hicks-Kennedy Meeks (2014-15)

  • 69 minutes, Offensive Efficiency: 130.0, Defensive Efficiency: 80.9, Efficiency Margin: +49.1
  • This lineup actually has the highest net efficiency of any on the list, albeit in only 69 minutes. Meeks, at his healthiest as a sophomore, was a +/- monster in 2014-15. From that perspective, the Hicks-Meeks frontcourt (Efficiency Margin of +33.7 in 226 minutes) was the Heels’ most successful (Johnson-Meeks was +15.3 in 591 minutes; Hicks-Johnson was -23.9 in 87 minutes, including an abysmal defensive efficiency of 139.1).

8.) Kendall Marshall-Dexter Strickland-Harrison Barnes-John Henson-Tyler Zeller (2011-12)

  • 247 minutes, Offensive Efficiency: 115.6, Defensive Efficiency: 79.6, Efficiency Margin: +36.0
  • Although this lineup actually has a higher net efficiency than the post-Strickland-injury lineup (with Reggie Bullock at the 2), it is aided by playing against a much softer strength of schedule. This lineup had better (unadjusted) defensive numbers than the post-Strickland version, but I preferred the length with Bullock at the 2. Both 2s were excellent defenders, though.

7.) Raymond Felton-Rashad McCants-Jackie Manuel-Marvin Williams-Sean May (2004-05)

  • 70 minutes (note: I’ve only charted 21 games in 2004-05), Offensive Efficiency: 125.0, Defensive Efficiency: 94.3, Efficiency Margin: +30.7
  • As noted, my charting for 2004-05 is incomplete—it includes the Maui games, the NCAA Tournament games, and about 10 of the conference games (among the 21 I’ve charted). It’s a pretty representative sample across the season, and I’m fairly confident in concluding that this was UNC’s best offensive lineup in ’05, but the Jawad-at-the-4 combo was UNC’s best defensively. This was obviously the crunch-time lineup that RoyW was leaning on by the end of the season; I reserve the right to move this one up (maybe way up) once I finish my 2005 charting project.

6.) Kendall Marshall-Reggie Bullock-Harrison Barnes-John Henson-Tyler Zeller (2011-12)

  • 283 minutes, Offensive Efficiency: 126.1, Defensive Efficiency: 93.0, Efficiency Margin: +33.1
  • As discussed in 8.), this lineup played against a much tougher average opponent than the version with Strickland at the 2. While it’s debatable which 2011-12 lineup was stronger defensively, the evidence is pretty unambiguous that having a shooter/floor spacer at SG helped offensively. Wrote about this over five years ago (!) at Shane Ryan’s old site. Ha.

5.) Ty Lawson-Marcus Ginyard-Reyshawn Terry-Brandan Wright-Tyler Hansbrough (2006-07)

  • 72 minutes, Offensive Efficiency: 125.5, Defensive Efficiency: 83.0, Efficiency Margin: +42.5
  • Smallish sample size here, but this was actually my favorite of the 2006-07 lineups (and there were hundreds and hundreds of them that year!). Although Ellington (even as a freshman) was undisputedly a better talent, I thought Ginyard fit in perfectly with this unit. Alongside four high-usage scoring options, Marcus was free to do what he did best—play lockdown wing defense, crash the offensive glass/make smart cuts to the hoop, and get out in transition for easy buckets. And, personally, I thought the sophomore-year version/role was peak Ginyard—I liked him a lot more as an instant-energy, high-motor bench guy (for 15-20 MPG) than a 30+ MPG starter.

4.) Ty Lawson-Wayne Ellington-Danny Green-Deon Thompson-Tyler Hansbrough (2008-09)

  • 351 minutes, Offensive Efficiency: 127.6, Defensive Efficiency: 94.4, Efficiency Margin: +33.2
  • The ’09 starting 5, this was a really effective offensive unit that could lock down on defense when motivated (see the NCAA Tournament run). It had a slightly worse net efficiency than the Davis-for-Thompson version of this lineup, but both were consistently dominant. If you’d like to move this one up solely for how well they started the ’09 title game against Michigan State, I won’t object!

3.) Ty Lawson-Wayne Ellington-Marcus Ginyard-Danny Green-Tyler Hansbrough (2007-08)

  • 119 minutes, Offensive Efficiency: 131.4, Defensive Efficiency: 99.1, Efficiency Margin: +32.3
  • The only small-ball combo on the list, this was Carolina’s crunch-time “Death Lineup” during its 36-3 season in 2008. As usual, playing small meant sacrificing a bit of defensive efficiency. But, when protecting a lead, this unit was so offensively efficient (and so good at the foul line) that teams were unable to match them score for score.

2.) Raymond Felton-Rashad McCants-Jackie Manuel-Jawad Williams-Sean May (2004-05)

  • 123 minutes (note: I’ve only charted 21 games in 2004-05), Offensive Efficiency: 113.9, Defensive Efficiency: 75.7, Efficiency Margin: +38.1
  • From a pure +/- perspective, this was one of the best defensive units of the Roy Williams era— five upperclassmen playing excellent team defense fueled by a great on-ball guy in Felton and a lockdown wing stopper in Manuel. I can’t put it first simply because RoyW had moved to Marvin over Jawad in March close-and-late situations (although still mixing in Jawad situationally in crunch-time).

1.) Ty Lawson-Wayne Ellington-Danny Green-Tyler Hansbrough-Ed Davis (2008-09)

  • 120 minutes, Offensive Efficiency: 135.9, Defensive Efficiency: 100.1, Efficiency Margin: +35.8
  • As good the the ’09 starting 5 was offensively, it actually got stronger on that end with Davis in the lineup (for Thompson). Even though Davis was a better rim protector with a higher 2009 Stop% than Thompson (63.1% vs. 60.3%), this quintet had a (slightly) worse defensive efficiency than the starters. It’s probably safe to say that lineups 1.) and 4.) were very close on both ends—I’ll take the Davis version just for pure rim protection/offensive rebounding reasons.


The Blueprint of a Great Carolina Offense

The Blueprint of a Great Carolina Offense

Yesterday, I had a post comparing the 2009 Tar Heels to the current team. Today, I’ll briefly extend that comparison to include a side-by-side look at the two offenses in general.

As seen in the table below, the 2009 version of Carolina looked exceedingly similar to the modern-day version—at least the first 13 games of it. And that’s the huge caveat in this analysis: I’m comparing whole-season numbers in 2009 (including an entire ACC campaign, plus a dominant six-game run through the NCAA Tournament) t0 non-conference numbers in 2016-17. It will be interesting to see how the comparison holds up as the schedule strength ramps up this season; one would expect, however, that UNC’s (unadjusted) offensive efficiency will fall a bit as the quality of opposing defenses rises. In 2009, the average defense that Carolina faced (per had an adjusted efficiency of 95.1 (six points better than the national average of 101.1). So far this season, the Heels have faced an average adjusted defensive efficiency of 99.5 (a little over three points better than the national average of 102.6).

Some conclusions to draw from the table include:

  • Each team’s distribution of possessions among turnovers, free throws, 3-pointers, and 2-pointers is nearly identical—like eerily similar
  • 2009 did shoot slightly better from behind the arc (as detailed in yesterday’s analysis) and significantly better from the foul line. 2017, however, has been a bit better on 2-pointers so far, and also is grabbing a higher percentage of possible offensive rebounds (41.5% vs. 38.9%).
  • The biggest difference between the two offenses involved their distribution of 2-pointers: while each shot a nearly identical proportion of 2s from 5-10′ (post moves and floaters), the ’09 team preferred—or settled for—(inefficient) mid-range 2s over (efficient) close 2s. While the ’09 group was significantly better from the mid-range than the current Heels, those were still very low-efficiency shots relative to the 5-10′ or (especially) close 2s.
  • Interestingly, the current post trio of Kennedy Meeks/Isaiah Hicks/Tony Bradley has been much more prolific and efficient from close range than their ’09 counterparts (Tyler Hansbrough/Deon Thompson/Ed Davis). This year’s group of bigs has attempted 14.1 close FGAs per game, making them at a rate of 68.3% (led by Hicks’s absurd 83.6% on close attempts). In ’09, the three primary posts combined for only 10.1 close FGAs per game, converting them at 63.7%. Again, this is partly due to the whole-season vs. partial-season nature of this comparison. But, it should also be remembered that the ’09 version of Hansbrough wasn’t as dominant (or explosive) as the NPOY one of 2008. After missing the first four games with a stress reaction in his shin, “Psycho T” returned as a slightly more perimeter-oriented version of himself. It will be worth following this metric as the season progresses to see if the 2017 Heels can remain more prolific and efficient than the ’09 team near the rim.
  • While the 2009 frontcourt wasn’t generating as many close opportunities as the current post rotation is, that was partially mitigated by how effectively Ty Lawson was able to get to the rim. He averaged 5.7 close FGAs / 40, making 62.4% of them. By comparison, Joel Berry is currently shooting 59.3% on his 3.6 shots / 40 at the rim. He did attempt seven (making five) against Kentucky, however, so it’s possible that he’ll be able to ramp up his aggressiveness against better defenses/opponents (at least in a fast-paced, open-court game).
  • Although we can expect the current (unadjusted) offensive efficiency of 119.5 to drop once the competition stiffens, the 2017 Heels should continue to be successful be following the Roy Williams blueprint: avoid turnovers, crash the offensive glass, get plenty of opportunities in the paint/at the rim. And, if it’s able to keep its fraction of possessions with low-percentage mid-range 2s so low, it’s possible that this year’s UNC offense can be just as effective as the 2009 championship edition.