Game Charting Insights: Touches & Dribbles

For touches we track the touches from a realistic attacking point. So if a rebounder dumps off the ball to the PG to bring across halfcourt, we don't care about the 'rebound touch.' Only touches from the point where a player is in range to at least take a three are counted. Each time a player touches the ball is also a separate count, so if it goes Williams-Wade-Shaq-Wade that counts as 4 for touches, (eg count both Wade's touches). For dribbles, we just noted 0, 1, 2, 3, etc ...simply the number of times the ultimate possesion event player bounces the ball.

Touches per Possession and resulting performance

The obvious first use of this information is to understand "typical" NBA numbers for a possession and how the touches correspond to resulting efficiency.

Touches	Freq.	FG%	FTA/FGA	TO/FGA	PtsPoss
1	23%	.491	.49	.21	99.6
2	30%	.495	.31	.13	105.7
3	23%	.478	.26	.12	98.5
4	13%	.448	.21	.10	97.4
5+	11%	.457	.16	.08	101.0

Freq. = Frequency of the touch number among possessions
PtsPoss = Points per 100 possessions (offensive rebounds not considered)

So the median number of touches on an NBA possession works out to be 2, again only starting to count from the point where a player is in reasonable attacking position. In terms of average number of touches per possession, that works out to be 2.7 touches. So in other words, despite all the talk of ball movement and the like, it often comes down to the first ballhandler trying to break down the defense and either take a shot himself or feed a "potential assist" pass to a teammate.

Also important to note is that we didn't detect an obvious trend of the points per possession getting better with more touches, although there are some very strong patterns in terms of foul drawing and turnovers both declining with more touches.

Of course there are notable variations on these numbers -- both frequency and efficiency -- from team to team. Let's move on to another view on the touches equation though.

Touches per Second

Since there's quite a difference between a quick shot possession and a more drawn out attempt to break down a defense, let's put it on a time scale.

Touches/Sec	Freq.	FG%	FTA/FGA	TO/FGA	PtsPoss
.01 to .14	25%	.442	.31	.15	92.8
.15 to .24	32%	.441	.27	.13	92.9
.25 to .34	23%	.505	.23	.12	105.6
.35 to .44	7%	.519	.28	.09	114.1
.45+	12%	.585	.55	.14	122.5

PtsPoss = Points per 100 possessions (offensive rebounds not considered)

Now we're getting somewhere! It's always nice to find undeniable trends, in this case the more touches per second the higher the points per possession (along with FG%). There's a little bias at work in that the quick put backs on offensive rebounds, which we know to be very effective situations for the offensive at close range, tweak the upper categories some but even below the highest touch per second category we see the same evidence at work.

So Ball Movement Advocates -- here's your proof: the faster the ball is moving from one player to another, the better the results overall!

George Karl is one of a number of coaches who preach that once you get the ball you need to do something right away otherwise you're just allowing the defense to reset...and he's right! In the NBA the players are so fast and so good, that any momentary advantage the offense creates can easily be nullified if the defense gets any time to react.

One interesting aspect is that turnovers didn't increase with a higher number of touches per second, so the idea that "frantic" movement and passing is dangerous seems misplaced. Again, at the NBA level players are so skilled they can handle a very fast pace indeed.

There's an obvious dividing line between the <.25 and .25+ touches per second (these were just splits we conceived as reasonable breaks, they are not fitted to the data), which at a glance suggests that when a player is holding the ball for more than four seconds he's costing his team. In truth since in the beginning of a possession the initial ball handler may well use a lot of time, it's likely that holding the ball for more than a second without some kind of action (dribble, pass, shot) is a bad thing!

Dribbling and Possession Effectiveness

For now we're only looking at the dribbling of the player who actually was involved in the possession event (shot, draw foul, turnover) and not the players touching the ball before him on a possession.

Dribbles	Freq.	FG%	FTA/FGA	TO/FGA	Ast'd	Blk'd	PtsPoss
0	42%	.484	.19	.10	86%	3%	105.7
1	17%	.475	.42	.16	59%	7%	96.7
2	11%	.500	.52	.14	20%	9%	102.6
3+	19%	.457	.37	.18	3%	9%	93.4

Ast'd = percentage of FGM that were assisted
Blk'd = percentage of FGA that were blocked
PtsPoss = Points per 100 possessions (offensive rebounds not considered)

The reality of NBA basketball is that over half the shots are taken without a dribble, and over 40% of possession events are coming from "catch and shoot" type situations. Moreover, the highest effectiveness in terms of points per possession is seen on zero dribble plays (which again also includes some putbacks after offensive rebounds).

So rather than lament when a player takes a pass and jacks up a shot, we need to evaluate whether the player had his feet set, was well balanced, and could get the shot off relatively uncontested -- if so, that might have been the best available option. Certain players make a fine living by moving to open spots and waiting for a pass!

Ah, but a team of five catch and shoot types would struggle mightily if it doesn't have a ball handler who can shake up a defense off the dribble, which is why most of your NBA stars and superstars have a knack for getting past defenders while bouncing the ball!

The numbers though show that the worst efficiency comes with the most dribbles, so over-dribbling can be a real problem: on 6+ dribbles the league as a whole shoots a woeful .398 FG% and averages a meek 81.5 points per 100 possessions. What you're really looking for is a guy with a great handle who also makes smart decisions, ideally being able to pass and shoot with the best...wondering why Steve Nash and Chauncey Billups get so much praise and attention? Look no further!

There's a lot of reason to take individual player stats with the above dribble splits seriously -- finding the guys who can create off the dribble is very important for overall team offense, and sometimes you'll watch a game and realize there's only one player on the floor for each team that really is capable of doing this.

Possession Results by Touches & Dribbles

There's a whole range of different play types that go into each touch number or dribble number, so doing the touch/dribble cross index can help us explore things a little deeper.

Touches	Dribbles	Freq.	FG%	FTA/FGA	TO/FGA	Ast'd	Blk'd	PtsPoss
1	0	6%	.540	.39	.24	10%	5%	99.3
1	1	3%	.431	.88	.13	3%	13%	104.5
1	2	2%	.509	.85	.18	4%	7%	114.6
1	3+	12%	.476	.40	.23	1%	7%	96.1
2	0	16%	.508	.18	.10	97%	3%	111.8
2	1	6%	.515	.50	.17	70%	4%	104.3
2	2	4%	.443	.56	.16	41%	8%	91.6
2	3+	4%	.468	.39	.14	7%	8%	98.6
3	0	12%	.484	.19	.09	97%	3%	105.7
3	1	5%	.452	.29	.16	60%	8%	88.4
3	2	3%	.543	.42	.16	18%	11%	98.7
3	3+	3%	.418	.39	.13	6%	15%	83.8
4+	0	13%	.437	.11	.06	94%	2%	101.3
4+	1	5%	.476	.26	.16	69%	6%	92.6
4+	2	3%	.523	.41	.08	10%	11%	112.3
4+	3+	2%	.389	.15	.13	7%	10%	79.6

The initial ball handlers on the offensive set do far better on multi-dribble possessions than the "late touch" guys and there's a fair correlation between the initital two touchers being the stars and the later touchers being "third/fourth" option types.

Of course players who are excellent dribblers and penetrators also tend to want the ball in their hands a lot, which is why we may find in our "chemistry experiments" series that having too many ace dribblers in a five man unit may be overkill, just as the Knicks are having trouble adjusting to a Marbury-Francis backcourt.

Are NBA players ball hogs?

So let's get down to the touchy subject -- are NBA players a little too interested in "getting their own" and not playing team ball as some college hoops fanatics would claim? Well, Bob Chaikin has for many years pioneered the idea of tracking player touches and charting what he calls the "Possession Factor" -- how often a guy passes, shoots, turns the ball over, or draws a foul.

On a league wide basis for all touches that occur in a "reasonable attacking position" you'll find this:

61% - Pass
29% - Shoot
 6% - Draw Foul
 4% - Turnover

So to paraphrase Mark Twain, "the rumors of the demise of team play in the NBA have been greatly exaggerated."

Players on the whole overwhelmingly look to pass first, and with the most efficient shot option being a zero dribble finish/open look jumper, it's with good reason! Now that's not to say there aren't a couple of exceptions, and then the truly talented mega stars of the NBA could on occasion be chided for being too unselfish and letting their teammates be the possession users when the best option is just to keep going back to the well.

Looking for NBA coverage at the "every touch" level? You won't have to look far: stop back in these parts during the playoffs!

Also see:
Game Charting: Contested Rebounds (04/06)
Game Charting: Floor Locations (03/27)
Game Charting: The Value of a Good Pass (03/20)

Next up: rotations, switches, and double teams!

Game Charters for the "Touches/Dribbles" Project:

Adrian Lawhorn, Andy Shiffman, Anthony Cerminaro, BJ Colby, Brad Burnett, Brendan Gildea, Brett Steele, Brian Ganster, Cameron Tana, Chad Casarotto, Charles Floyd, Chris Goudey, Curtis Chody, Dana Henderson, Daniel Kelly, Dave Daniels, David Mintz, Derek Adesso, Dmitri Salcedo, Don Marconi, Eric Patten, Eric Wallace, Greg Humphries, Husamettin Erciyes, Jared Wade, Jason Blaze, Jeff Miller, Jeremy Killey, Joe Malloy, John Magee, Jon Walker, Kevin Bartlett, Matt Kolsky, Michael Boehm, Miguel Cuaron, Mike St. Pierre, Mike Wolf, Patrick Sheehy, Phil Edwards, Raj Kannan, Ray Maze, Rich Schmidt, Rob Ireland, Rob Stewart, Sachin Gupta, Sandy Weil, Schuyler Sheaffer, Shawn Krest, Susan Nelson, Tom Lore, Tony Williams, Zach Ellin

Thanks as always to the noble efforts of the charters -- they are the ones who will be pushing NBA statistical analysis to new heights!

Interested in doing some game charting? Send a message to: charting@82games.com