Constructing All-ACC teams (or: holy cow does the media love them some Austin Rivers)

Ed. Note: Great work by quzy, deserving of a FP.

Basketball is fundamentally a team game. As a result, it is a large challenge for basketball statisticians to find proper valuations for individual players within their role of the team framework. It is more difficult than baseball, as baseball has more individual opportunities; it is of a similar difficulty to football, where effective teams have less famous players that allow the prestige players to make plays.

And yet, the awards given at year's end in these sports are for individuals. MVP's, Heisman trophies, All-America teams, All-Star teams - these teams represent statistically dominant players without recognizing their teammates that often allow the individual to shine.

The ACC released the media-voted All-ACC teams today. The team was, as anticipated, dominated by the Triangle teams. For a full rundown of the teams by the BSD crew, please visit here.

What follows is an attempt to construct All-ACC teams without regard to both overall team success or public perception using efficiency statistics.

Methodology: To analyze individuals, you have to remove as much of the team component as possible. There still is no unanimous way of doing this in basketball. What I decided to use was individual (and team) efficiency stats, which I found at Sports Reference. I decided to create two sets of teams using these stats.

The first set of teams is based on pure efficiency ratings. This set is dominated by the teams that collectively have performed well. The second set of teams is based on value - how a player compares to an average player on his own team. This set of teams, in my opinion, illustrates the players that have done the most to carry their teams (hence the value term). I got a player's value rating by comparing his individual efficiency rating to his team's average rating. I weighted both teams by the percentage of minutes the player played, under the theory that you cannot add (or lose) value to your team if you're not playing, and a player who plays at a certain level 80% of his team's minutes is better and more valuable than a player who plays at that same level 50% of his team's minutes.

Example Calculation: I'll use Tony Chennault as an example of the formulae. This season, Wake played 30 games with no overtimes, so 1200 total minutes that a player could have been on the court. Wake had team ratings of 98.8 points scored per 100 possessions on offense and 105.4 points allowed per 100 possessions on defense. Tony Chennault, in 905 minutes, had individual ratings of 93.7 on offense and 108.7 on defense, meaning that he was 5.1 points below average on offense and 3.3 points below average on defense. Therefore, his pure efficiency rating is -15.0 points per 100 possessions. Factoring in the performance of the average Wake player, Tony had a value rating of -8.8 points per 100 possessions (he was 8.8 points worse than the average Wake player). Weighting in Tony's playing time, he gets a weighted efficiency rating of -11.31 points per 100 possessions and a weighted value rating of -6.34 points per 100 possessions. Tony's defensive value is similarly calculated with just the individual and team defensive efficiencies, leading to a defensive value of -3.3 and a weighted defensive value of -2.49.

Note: When compiling this data, I noticed that the defensive efficiency stat used unfairly gives better defensive efficiencies to big men. I believe this to be a systemic flaw in the calculation, but I have no idea a better method of evaluation. As such, the figures are, in my opinion, slightly skewed towards big men who accumulate defensive rebounds and (especially) blocked shots.

Another Note: There are likely better calculations to weight the pure efficiency data than percentage of minutes played. One of the most frequent such methods is usage rate, which can also be found at Sports Reference. For purposes of this study, just know that I have issues with usage rate as it is currently formulated, because it doesn't allow for playing time. A player with a high efficiency and a 40 percent usage rate who plays 15 minutes a game should not, in my opinion, be considered more valuable than a player with an equally high efficiency but a 20 percent usage rate who plays 40 minutes a game.

Another Another Note: When calculating the pure defensive efficiency All-Defense team, I just used individual defensive efficiency ratings with the caveat that only players who played more than 600 minutes (approximately half of a team's minutes) were eligible.

Last Note (I Promise): To be eligible for these teams, a player had to log at least 450 minutes this season (approximately 15 minutes or more per game). Any less than that and small sample size variance creates some bizarre results (hello ACC PoY John Cahill of Boston College!).


All-ACC Pure Efficiency Teams


ACC Player of the Year: Mike Scott - Virginia

First Team All-ACC

Mike Scott - UVA (26.94)

Tyler Zeller - UNC (26.17)

John Henson - UNC (21.09)

Scott Wood - NCSU (17.65)

Andre Young - Clem (17.48)

Second Team All-ACC

Joe Harris - UVA (16.43)

Ryan Kelly - Duke (16.19)

Tanner Smith - Clem (14.76)

Harrison Barnes - UNC (14.68)

Reggie Bullock - UNC (14.62)

Third Team All-ACC

Bernard James - FSU (13.81)

Kendall Marshall - UNC (13.70)

Miles Plumlee - Duke (13.16)

Michael Snaer - FSU (12.53)

CJ Williams - NCSU (12.04)

Honorable Mention (players with a weighted efficiency rating higher than 10.00): Jarell Eddie - VT (11.84), Mason Plumlee - Duke (11.34), Erick Green - VT (11.18), Kenny Kadji - Miami (10.80), Seth Curry - Duke (10.52), Durand Scott - Miami (10.40)

Notables (players who were on the media all-ACC teams or honorable mention): CJ Harris - WF (5.77), CJ Leslie - NCSU (5.70), Lorenzo Brown - NCSU (3.77), Austin Rivers - Duke (3.10), Terrell Stoglin - MD (2.10), Travis McKie - WF (0.26)


Defensive Player of the Year: John Henson - UNC

All-Defense Team

John Henson - UNC (84.9)

Mike Scott - UVA (86.2)

Bernard James - FSU (86.5)

Sammy Zeglinski - UVA (87.6)

Tyler Zeller - UNC (87.8)

Notables: Jontel Evans - UVA (88.7), Michael Snaer - FSU (94.1), Andre Young - Clem (97.6)


Freshman of the Year: Austin Rivers - Duke

All-Freshman Team

Austin Rivers - Duke (3.10)

Malcolm Brogdon - UVA (2.18)

Shane Larkin - Miami (-0.12)

Julian Royal - GT (-0.27)

Lonnie Jackson - BC (-1.01)

Notables: Dorian Finney-Smith - VT (-1.84), Ryan Anderson - BC (-3.08), Nick Faust - MD (-11.53)


All-ACC Value Teams


ACC Player of the Year: Mike Scott - UVA

First Team All-ACC

Mike Scott - UVA (14.59)

Andre Young - Clem (11.94)

Scott Wood - NCSU (11.66)

CJ Harris - WF (11.37)

Tyler Zeller - UNC (11.11)

Second Team All-ACC

Tanner Smith - Clem (9.52)

Daniel Miller - GT (9.43)

Jarell Eddie - VT (8.80)

Lonnie Jackson - BC (8.29)

Erick Green - VT (7.58)

Third Team All-ACC

Ryan Kelly - Duke (7.32)

Sean Mosley - MD (7.13)

Glen Rice Jr. - GT (7.09)

Ryan Anderson - BC (7.05)

Miles Plumlee - Duke (6.39)

Honorable Mention (players with a weighted value rating of 5.00 or higher): Bernard James - FSU (6.26), Travis McKie - WF (5.95),Kenny Kadji - Miami (5.84), CJ Williams - NCSU (5.50)

Notables: Terrell Stoglin - MD (4.93), John Henson - UNC (4.68), Durand Scott - Miami (4.32), Michael Snaer - FSU (4.20), Mason Plumlee - Duke (1.74), CJ Leslie - NCSU (0.83), Seth Curry - Duke (0.22), Harrison Barnes - UNC (-0.84), Lorenzo Brown - NCSU (-2.60), Kendall Marshall - UNC (-4.43), Austin Rivers - Duke (-8.17)


Defensive Player of the Year: John Henson - UNC

All-Defense Team

John Henson - UNC (4.46)

Mason Plumlee - Duke (3.83)

Daniel Miller - GT (3.19)

CJ Leslie - NCSU (2.63)

Carson Desrosiers - WF (2.54)

Notables: Bernard James - FSU (2.52), Jontel Evans - UVA (-1.28), Andre Young - Clem (-2.56), Michael Snaer - FSU (-2.93)


ACC Freshman of the Year: Lonnie Jackson - BC

First Team

Lonnie Jackson - BC (8.29)

Ryan Anderson - BC (7.05)

Ashton Pankey - MD (2.59)

Julian Royal - GT (0.84)

Dennis Clifford - BC (0.00)

Notables: Shane Larkin - Miami (-4.52), Dorian Finney-Smith - VT (-5.03), Austin Rivers - Duke (-8.17), Nick Faust - MD (-9.16)



1) There was a lot of hype concerning Mike Scott mid-year for ACC PoY, most of which has died down as Virginia has skidded over the last few weeks. However, Scott's been the best player in the ACC this season by both metrics, and would rightfully take home the PoY trophy. He still may, but Tyler Zeller - not a bad choice in his own right, but not the correct choice - earned more points in the All-ACC balloting, and thus has to be considered the favorite right now.

2) The North Carolina current starting five are all in the top 12 of the weighted efficiency stat. In fact, they did not have a single player that I looked at have a negative efficiency. They are a juggernaut, and deserved ACC regular season champions. However, if they bow out early in the ACC Tournament (or possibly even lose in the final), they might garner a two seed in the NCAA Tournament. Here's guessing that No. 1 seed will not be thrilled to see Carolina potentially awaiting them in the Elite Eight.

3) It looks like the media hype surrounding John Henson's defense is accurate. He wins both defensive awards here.

4) Glad to see Carson Desrosiers sneak onto the all-defensive value team. He's had a good season, particularly defensively. He'd be higher had he not lost time to Ty Walker in conference play.

5) Most underrated ACC players? Scott Wood and Andre Young both place on both first-teams despite not getting a significant number of votes from the media. Interestingly, Young was recognized for his defense, which isn't necessarily accurate. Wood's weighted defensive value, by the way, was abysmal at -5.04, meaning he might've been the best offensive player in the ACC, with the exception of...

6) I was not kidding when I said BC's John Cahill might've been the ACC's PoY under the value metric had I not weighted the data. BC posted team ratings of 90.7 on offense and 104.5 on defense. Cahill had an offensive rating of 118.5 and a defensive rating of 107.1, so his weighted rating would've been +25.2. Mike Scott's is +18.9. However, Cahill did not meet the minutes standard I'd established (he's played 372 this season).

7) Other underrated ACC players include Tanner Smith and CJ Harris. CJ gets a big bump from being a good player on a statistically bad team. Tanner Smith has been just quietly excellent.

8) Underrated ACC teams include Virginia Tech (whose +4.5 team efficiency is closer to the Miami-NC State-Clemson troika than the Maryland-Georgia Tech-Wake set) and Virginia (who is the second most efficient team in the conference with a +16.0 rating, just ahead of Duke's +13.8).

9) Overrated players? Hello Austin Rivers. I had initially planned this piece to be a head-to-head comparison between Rivers and CJ Harris, but with the release of the media all-conference teams today I decided to do this instead. And does the media love Austin Rivers or what? First-team all-conference, when he is, in fact, the least productive and valuable member of Duke's rotation (and it's not even close). His weighted efficiency is good for a freshman (although Malcolm Brogdon had a higher unweighted efficiency), but just take a look at the weighted value. Woof.

10) Travis McKie still gets no love. He was hurt by a poor defensive metric (106.3 points allowed). As a result, he gets snubbed... again.


Your thoughts? Obviously this had yielded some weirdness, and as such I appreciate all comments on both methodology and the end results. If you want individual results, I'll answer such requests in the comments.

The content of FanPosts is not necessarily the opinions, thoughts or beliefs of Blogger So Dear.

In This FanPost