FanPost

What is Slow or Fast Play in the NBA? Revisited: Offense

Kyle Terada-USA TODAY Sports

SUMMARY INTRODUCTION

Almost a month ago, I published "What is Slow or Fast Play in the NBA?" on the two SBNation sites. It provided a gentle metaphorical introduction to a variety of measures for how fast an NBA team plays. This piece will be shorter and reality based. Furthermore, I have sharpened the focus to only how fast a team’s offense is. With one exception, all the data are as of January 2, 2016.

First, a few words on the differences between the unit of activity called a possession used in advanced metrics and the unit called a play used by the NBA Player Tracking System. Play has several advantages over pace including counting offensive rebounds as starting new plays and providing separate values for each team for each game. (Pace is an average across a game’s two teams.)

Two tables contain orderings of NBA teams by different measures or metrics as well as giving the values ordered upon, some have supplementary information. The first table has seven orderings of teams. The bullets note the rankings of Golden State, San Antonio, and Oklahoma City within parentheses.

These three teams while all outstanding teams have three different styles. Informally, the might be described as GSW, work as a group to kill’em quick, SAS, work as a group to move’em and confuse’em until they’re easy to kill, and OKC, use one or two assassins to kill’em. Unsurprisingly, GSW is fast by most measures, so for them mainly this will be a journey across the tops of columns. The others are more up and down.

The fastest team is ranked number one. Some ordering in the tables also have supplemental information identified here after a hyphen. The first table’s orderings of teams are by:

  1. Pace, average "possessions" per 48 minutes, as calculated by Basketball Reference (GSW 2, OKC 15, SAS 26) – pace as calculated by NBA.com

  2. Plays per 48 minutes (GSW tied 3-5, OKC 7, SAS 29)

  3. Minutes per game of physical possession of ball (GSW 1, OKC 12, SAS 28) [fewer minutes treated as faster]

  4. Number of touches per game in the frontcourt (SAS 3, GSW 6, OKC 30)

  5. Average seconds per touch for all touches (GSW 1, SAS 8, OKC 30) – average dribbles per touch

  6. Passes received, that is completed (SAS 3, GSW 7, OKC 30)

  7. Average player speed on offense (GSW tied 1-2, SAS 8, OKC 20) – total distance traveled by all team members on offense and percentage of team’s plays that are transition plays

The second table covers play-related data and contains three orderings:

  1. Seconds per play (GSW 1, OKC 9, SAS 29)

  2. Plays per 48 minutes (GSW tied 3-5, OKC 7, SAS 29) – game clock minutes per 100 plays and plays per game

  3. Touches per play (SAS 1, GSW 13, OKC 30) – passes per play and average seconds per touch

Possibly, no two numbers shows the contrasts in styles among GSW, SAS, and OKC as well as the contrast between 8, seconds per play, and 10, touches per play. Golden State’s plays finish quickly with a slightly above median number of touches – kill’em quickly. San Antonio plays take a long time but with the most touches – move’em and confuse’em until easy to kill. Oklahoma City’s plays take a moderate amount of time with few touches – a few assassins do the job.

THE TABLES

The first table below shows the order of NBA teams from fast to slow by each of seven measures each of which is immediately to the right of each column of teams ids. I have included three other columns of related data for information. The three teams mentioned in the list above Golden State’s (GSW), San Antonio’s (SAS), and Oklahoma City’s (OKC) have typeface changes to aid finding them. GSW is in bold, SAS in bold italics, and OKC underlined. In addition, the Clippers are in italics (LAC) and Cleveland has a dotted underline (CLE).

If the reader read the Summary introduction and feels comfortable, he or she can jump to directly reviewing the tables picking up the text again at Final Remarks and Conclusions. However, the text does work through the tables providing some context.

The first ordering of teams is by the traditional measure, pace or "" per 48 minutes, as defined by Basketball Reference. I have included the values as defined by NBA.com. That these are different highlights merely one of the problems with pace. GSW ranks second by this measure, OKC 15th, SAS 26th. Thus, by the traditional measure the three include a fast team (GSW), a middle of the pack team (OKC), and a slow team (SAS).

The second ordering is by the number of the unit used by the NBA Tracking System, called a play. As mentioned above, it has several advantages over pace including counting offensive rebounds as new possessions and providing separate values for each team for each game. (Pace is average across a game’s two teams.) Unfortunately, the statistic published from the tracking system is not plays per game but the amount in game clock minutes 100 plays would proportionally take. Nevertheless, it allows us to calculate plays per 48 minutes and order the teams albeit with a number of ties. Here GSW ranks tied for third through fifth, OKC seventh, and SAS 29th.

Luckily, the next stat published from the Tracking System is convenient. The time of possession is the time the team physically has the ball and does not include such time as when a shot is in the air. Thus, for a game the two teams’ times of possession add up to significantly less than 48 minutes. However, it does directly measure the time a team physically has the ball on offense. GSW is ranked first by this automated measurement of time, OKC 12th, and SAS 28th.

Table on Offensive Speed Measures

Rk

Team

Basket

Ball

Refer

Pace

NBA

STATS

PACE

Tm

Plays

per

48 Min

Tm

Time
Of
Poss

Tm

Front
CT
Touches

Tm

Avg
Sec
Per
Touch

Avg
Drib
Per
Touch

Tm

Passes
Received

Tm

Player
Speed
Off

Team

Dist.
Miles
Off

%

Plays

Trans

1

SAC

99.8

102.13

SAC

113.2

GSW

17.7

UTA

362.4

GSW

2.36

1.75

NYK

349.1

GSW

4.74

9.07

17.0

2

GSW

99.4

101.87

BOS

112.4

WAS

17.8

NYK

356.6

BOS

2.44

1.89

UTA

345.9

ORL

4.74

9.57

12.7

3

PHO

98.6

100.88

DET

111.9

BOS

18.3

SAS

354.7

NYK

2.48

1.73

SAS

345.1

CHA

4.72

9.56

10.4

4

BOS

98.3

100.95

GSW

111.9

HOU

18.5

ATL

339.2

WAS

2.51

1.97

DAL

334.7

WAS

4.68

8.91

19.1

5

WAS

98.0

100.30

PHO

111.9

ATL

18.6

BOS

335.7

ATL

2.53

2.00

PHI

327.1

NOP

4.67

9.09

12.5

6

HOU

97.6

100.03

HOU

111.1

PHO

18.6

GSW

334.9

DAL

2.57

2.09

BOS

322.9

PHI

4.66

9.55

14.7

7

PHI

97.6

99.94

OKC

110.9

NOP

18.7

PHI

333.8

PHO

2.57

2.09

GSW

322.9

POR

4.64

9.46

12.5

8

IND

97.1

99.46

CHI

110.3

SAC

18.7

DAL

332.5

SAS

2.58

2.09

ATL

321.8

SAS

4.63

9.72

10.4

9

ATL

96.5

98.72

IND

110.1

IND

19.1

ORL

327.7

NOP

2.59

2.05

ORL

313.7

CHI

4.60

9.35

11.4

10

NOP

96.5

98.68

LAL

109.8

LAC

19.1

CHI

324.3

UTA

2.62

2.01

TOR

310.4

SAC

4.57

8.95

18.4

11

CHI

96.3

98.89

PHI

109.8

MIL

19.1

TOR

323.9

PHI

2.64

2.26

NOP

310.1

ATL

4.55

8.89

13.6

12

LAC

96.2

98.54

DEN

109.3

OKC

19.2

MEM

318.1

CHI

2.68

2.05

CHI

309.1

DAL

4.54

9.34

10.7

13

LAL

96.2

98.76

POR

109.1

BRK

19.4

NOP

315.0

IND

2.68

2.03

PHO

308.8

BOS

4.52

8.79

15.4

14

CHA

96.0

98.18

WAS

109.1

CHA

19.4

CHA

314.8

ORL

2.69

2.13

MEM

308.7

MIL

4.52

8.97

12.1

15

OKC

96.0

98.70

BRK

108.8

CHI

19.4

PHO

314.0

HOU

2.71

2.05

CLE

308.4

NYK

4.52

9.27

11.4

16

DAL

95.7

97.79

ORL

107.9

NYK

19.4

IND

313.3

CHA

2.72

2.26

CHA

307.9

UTA

4.52

9.49

10.7

17

MIN

95.7

98.09

MIN

107.6

ORL

19.5

DEN

310.4

MIL

2.72

2.01

WAS

304.9

DEN

4.50

9.15

11.4

18

BRK

95.6

97.76

LAC

107.4

DAL

19.6

MIL

310.2

LAC

2.74

2.23

IND

303.9

OKC

4.50

8.92

15.3

19

DET

95.5

97.93

NOP

107.4

MIN

19.6

CLE

309.0

SAC

2.75

2.24

MIL

301.3

PHO

4.48

8.90

14.6

20

DEN

95.3

97.62

NYK

107.4

DEN

19.7

MIA

305.8

MEM

2.78

2.18

MIA

300.1

BRK

4.47

9.01

11.8

21

POR

94.9

97.24

ATL

106.9

MIA

19.7

BRK

305.5

CLE

2.79

2.19

LAC

296.3

LAL

4.46

9.16

12.7

22

ORL

94.7

97.00

CHA

106.9

POR

19.8

WAS

305.0

DEN

2.80

2.23

DEN

295.8

HOU

4.43

8.67

17.7

23

MIL

94.1

96.39

DAL

106.4

LAL

19.9

LAC

301.4

MIA

2.83

2.30

BRK

288.8

IND

4.43

8.88

16.1

24

MEM

93.8

96.06

TOR

106.4

MEM

19.9

POR

298.5

BRK

2.84

2.35

HOU

286.0

DET

4.42

9.15

13.3

25

NYK

93.7

96.41

CLE

106

PHI

19.9

HOU

290.3

TOR

2.85

2.39

POR

285.1

MIA

4.42

8.93

12.5

26

SAS

93.7

95.62

MEM

106

CLE

20.0

SAC

288.6

POR

2.92

2.37

SAC

282.6

MIN

4.40

8.86

12.2

27

TOR

93.4

95.86

MIL

106

DET

20.0

LAL

284.8

MIN

2.96

2.38

LAL

277.0

TOR

4.36

9.24

11.8

28

CLE

93.1

95.39

UTA

105.7

SAS

20.0

MIN

282.4

DET

2.97

2.50

DET

276.2

CLE

4.32

8.98

12.6

29

MIA

92.3

94.46

SAS

104.8

UTA

20.3

DET

277.9

LAL

2.98

2.21

MIN

274.2

MEM

4.29

8.87

10.9

30

UTA

91.5

93.96

MIA

102.6

TOR

20.5

OKC

277.5

OKC

2.98

2.43

OKC

262.1

LAC

4.26

8.51

12.0

The next measures are activity based. The first is the number of frontcourt touches per game. I chose frontcourt touches over total touches because I think that it is more directly related to offense. Here, SAS reverses its location and comes in third, GSW sixth, and OKC plunges to last. The average seconds per touch could easily have an effect on the perception of how fast a game is if players quickly move the ball on. GSW are first in this combination of time of possession and number of touches, SAS is eighth, and OKC remains last. I have included dribbles per touch simply because some may think it is additional activity although others think it slows the game down.

More passes can increase the sense of the speed of the game. Number of passes mirrors the number of touches as additional touches mainly occur via passes. So the order is the same with SAS third, GSW seventh, and OKC is once again last.

Does OKC play a slow game? Most would not say so. The next measure, average player speed, has OKC is higher at 20th. GSW and SAS are near the top with GSW in a tie for first and second, SAS in eighth. I have included columns with information regarding number of miles traveled on offense and the percentage of the team’s plays that are transition plays. Unlike all the other data that is through January 1, this last column on transition plays is through January 9.

Some differences between the results involving similar measures have straightforward explanations. OKC’s significant difference between pace (number of possessions) and number of plays reflects its frequent capture of offensive rebounds (30.5%). Is OKC the slowest team because it ranked lowest in three measures involving touches and passing? Few who have had to run with them would say so.

Other aspects can be of interest. A team with high speed and low percentage of transition plays probably executes their half court offensive with more and/or faster player movement.

Because LAC had the slowest player speed, I italicized LAC, and you can pick out its rank on the other measures. Is LAC the slowest offense in the NBA? Some might say so, but the fact that others would not points to peoples perceptions fitting different measures. The traditional measure, pace, had the Clippers tenth.

LAC has the slowest average speed of player movement, but JJ Redick can still gain separation and execute coming off screens at high speeds. Presumably, this might be offset in part by the players setting picks for him needing to stand still. However, much more than this probably lies behind LAC having had the slowest average player speed.

Below is a second table that emphasizes measures based on the concept of plays. To me, one of the most intuitive measures is seconds per offensive play. Some use fewer seconds than others to execute each possession of their offense. GSW is almost an outlier with its low seconds per play. Unsurprisingly, this behavior is never characterized as hurrying to give the other team a chance to score as some other quick-shoot offenses have been. Seconds per play is also more under the control of the offensive team than the total number of plays per game as this depends in part on how much time the opponent uses on offense.

Data per Plays

Rk

Team

Seconds

per

Play

Team

Plays

per

48 Min

Game Min

per 100

plays

Plays

per

Game

Team

Touches

per

play

Passes

per

play

Team

Avg
Sec
Per
Touch

1

GSW

9.4

SAC

113.2

42.4

113.6

SAS

4.43

3.29

GSW

2.36

2

BOS

9.7

BOS

112.4

42.7

113.2

UTA

4.37

3.25

BOS

2.44

3

WAS

9.8

DET

111.9

42.9

114.4

NYK

4.36

3.25

NYK

2.48

4

HOU

9.9

GSW

111.9

42.9

113

DAL

4.28

3.13

WAS

2.51

5

PHO

9.9

PHO

111.9

42.9

112.2

ATL

4.13

3.01

ATL

2.53

6

SAC

9.9

HOU

111.1

43.2

112.1

PHI

4.11

2.98

DAL

2.57

7

CHI

10.3

OKC

110.9

43.3

111.9

MIA

4.05

2.91

PHO

2.57

8

IND

10.3

CHI

110.3

43.5

112.9

TOR

4.05

2.91

SAS

2.58

9

OKC

10.3

IND

110.1

43.6

110.8

MEM

4.04

2.91

NOP

2.59

10

ATL

10.4

LAL

109.8

43.7

110.2

CLE

4.01

2.89

UTA

2.62

11

NOP

10.4

PHI

109.8

43.7

109.8

NOP

3.99

2.87

PHI

2.64

12

DET

10.5

DEN

109.3

43.9

109.7

CHA

3.97

2.87

CHI

2.68

13

BRK

10.6

POR

109.1

44

109.7

GSW

3.97

2.86

IND

2.68

14

LAC

10.6

WAS

109.1

44

109.1

ORL

3.97

2.87

ORL

2.69

15

ORL

10.7

BRK

108.8

44.1

109.6

BOS

3.96

2.86

HOU

2.71

16

CHA

10.8

ORL

107.9

44.5

109.2

MIL

3.96

2.83

CHA

2.72

17

DEN

10.8

MIN

107.6

44.6

109

WAS

3.91

2.8

MIL

2.72

18

LAL

10.8

LAC

107.4

44.7

107.7

LAC

3.89

2.76

LAC

2.74

19

MIL

10.8

NOP

107.4

44.7

108.1

IND

3.87

2.75

SAC

2.75

20

MIN

10.8

NYK

107.4

44.7

107.7

PHO

3.87

2.76

MEM

2.78

21

NYK

10.8

ATL

106.9

44.9

106.9

CHI

3.85

2.74

CLE

2.79

22

POR

10.8

CHA

106.9

44.9

107.6

DEN

3.84

2.7

DEN

2.80

23

PHI

10.9

DAL

106.4

45.1

107.1

BRK

3.75

2.64

MIA

2.83

24

DAL

11.0

TOR

106.4

45.1

106.7

POR

3.71

2.6

BRK

2.84

25

CLE

11.2

CLE

106

45.3

107.1

HOU

3.66

2.55

TOR

2.85

26

MEM

11.2

MEM

106

45.3

106.3

LAL

3.64

2.52

POR

2.92

27

UTA

11.4

MIL

106

45.3

106.6

MIN

3.64

2.52

MIN

2.96

28

MIA

11.5

UTA

105.7

45.4

106.4

SAC

3.6

2.49

DET

2.97

29

SAS

11.5

SAS

104.8

45.8

104.8

DET

3.53

2.42

LAL

2.98

30

TOR

11.5

MIA

102.6

46.8

103.2

OKC

3.46

2.35

OKC

2.98

The members of the 10 seconds or less per play club and the 11+ group fit my impression of most people’s sense of what teams play fast or slow. However, SAS that many feels plays relatively fast is next to the bottom. It simply takes time to execute a "move’em and confuse’em so they are easy to kill" strategy, but a large amount of often fast movement occurs while SAS is doing move’em, not to mention confuse’em. Indeed, opponents have often mentioned the need to be prepared to defend for a long 24 seconds. This also points to lamentably missing statistics on distance and speed of ball movement.

The second ordering in this second table is a repeat from the first table but provides the raw statistic from NBA.com, minutes for 100 plays. Plays per game vary from per 48 minutes when overtimes are involved.

Touches per play tends to be about one larger than passes per play as usually – and hopefully – the last member of the team to touch the ball on a play does not make a pass. Average seconds per touch is repeated from the first table for ease of comparison.

FINAL REMARKS and CONCLUSIONS

For live-ball-related measures, I want to lament again that the speed and distance traveled of the ball are not published from the NBA Player Tracking System. The data exists to produce such information. It would not only be useful, but many people watch basketball by watching what happens with the ball and such a measure should fit their intuition of how fast a game is.

Also of interest for some purposes would be tabulations of the wall clock times between the beginning and ends of games. The measures discussed above and in tables do not address the time absorbed by such activities as free throws and timeouts. Large amounts of such time can create an impression of a slow game.

The conclusions from this survey are straightforward:

  • No single measure fits what everyone means when they say a game is slow or fast because they do not mean the same thing

  • Multiple measures exist related to how slow or fast games are and more of them should be in common use

  • Pace, the traditional measure, is flawed compared to play and neither serves to express the full set of important aspects related to how slow or fast a game or team is.

Note:

I was happy to see Gomez’s use of multiple measures in his recent article on Pounding the Rock, Calling the Spurs "slow" is just wrong by Jesus Gomez Jan 6, 2016. I hope he continues and others begin to use a wider range of the variety of measures available.

This FanPost is a submission from a member of the mighty Golden State of Mind community. While we're all here to throw up that W, these words do not necessarily reflect the views of the GSoM Crew. Still, chances are the preceding post is Unstoppable Baby!