The Atlanta Hawks Defy the Percentages

NOTE: All shot location data cited comes from Hoopdata. League averages and rankings cited come from Basketball-Reference.com

Prelude

The Atlanta Hawks have an above average offense but they're below average at two of the four factors (offensive rebounding, free throw rate) and their turnover rate is almost exactly at the league average. Only when shooting the ball do the Hawks exceed an average team's performance. If you were going to pick just one of the four factors at which to excel, that would be the one but the Hawks are a good, rather than an exceptional field goal shooting team. They rank 11th in the league in eFG%, as many percentage points behind league-leader Boston as they are ahead of 26th and 27th ranked Washington and New Jersey.

The interesting stuff

The Hawks aren't an above average shooting team because they take or make a lot of three-point shots. The average NBA team takes 22.4% of their field goal attempts from beyond the arc. The Hawks take 23.1% of their field goal attempts from beyond the arc. The league averages 36% on three-point shots. The Hawks make 36.2% of their three-point shots. The Hawks pick up some ground on the league with their three-point shooting, but not a lot.

The Hawks aren't an above average shooting team because they take or convert a high percentage of shots at the rim. The average NBA team takes 27.2% of their field goal attempts at the rim and make 62.9% of those shots. The Hawks take 23.2% of their field goal attempts at the rim and make just 60.6% of those shots.

No, the Hawks are an above average shooting team because they make low-percentage shots far more often than an average team.

Here are a couple of graphs comparing the shot distribution of the league as a whole to the shot distribution of the Hawks:

Look at the size of that green piece of pie the Hawks make. Only the Wizards and the Heat use a larger percentage of their field goal attempts on 16-23' jump shots. However, only the Dallas Mavericks better Atlanta's 44.4% shooting from that range. The league, as a whole, shoots just 39.3% from that range.

One possible explanation for Atlanta's success: The Hawks earn assists on 69% of their made field goals from that range. The league as a whole earns an assist on just 61.2% of such shots. It's quite possible the Hawks create higher-quality long two-point jump shots than most teams.

Looking at the success of individual Atlanta players from that range this season lends some support to this hypothesis. Joe Johnson (36.4% from 16-23') and Mike Bibby (37.6% from 16-23') both create a lot of their own long, two-point jump shots with just 30.4% and 52.3%, respectively, coming off assists. Al Horford (57.7% from 16-23') earns a teammate an assist on 92% of his long two-point jump shots.

There's conflicting evidence, as well. Jamal Crawford has made 45.5% of his long two-point jump shots but just 36.9% of them are assisted.

Another way the Hawks thrive despite taking so many (generally speaking) low-percentage shots is by taking fewer of the next lowest-percentage shots. NBA teams typically use 25.5% of their field goal attempts on shots outside of the basket area but inside of 16 feet. The Hawks take just 24.4% of their shots from that vast mid-range. The Hawks best the league average field goal percentage from that range, as well, almost entirely on the strength of Joe Johnson and Al Horford combining to make 129 of 243 shots (53.1%) inside of 10 feet but not at the rim so far this season. The league shoots just 43.4% from that range.

Sustainability

Is the unusual nature of Atlanta's shooting success sustainable come the playoffs? The Hawks have played just 15 of their 44 games this season against teams with winning records and they've struggled horribly on the offensive end in those games, averaging less than a point per possession.

Given the small sample size and the across-the-board offensive declines the team has suffered against better opposition, it's difficult, if not unreasonable to proffer an answer with any certainty.

Based on what's happened, though, jump shooting has been the least of Atlanta's offensive problems against teams with winning records. The Hawks have shot slightly more frequently and slightly less accurately outside of 16 feet in those games, but the serious problem has been in converting shots inside of 15 feet:

Opp	eFG% (16+')	%FGA (16+')
all	48.8	52.4
over .500	47.2	54.2

Opp	eFG% (inside 15')	%FGA (inside 15')
all	52.7	47.6
over .500	47.1	45.8

Whether this is indicative of a larger inability to create high-percentage shots* or the nature of sub-dividing performance into smaller and smaller segments (especially considering said segment includes an historically poor performance) is a fair question. Additional evidence (the Hawks are scheduled to play 10 straight games against teams with winning records in February and March) should help us come to an understanding as to whether the Hawks are forging a distinct path of offensive success or have just discovered a way to leverage their strengths and limit exposure of their weaknesses against less talented and successful teams.

*Supporting this notion, Atlanta's below average free throw rate plummets from 22.2 (19th in the league) to 17.0 (would rank dead last, and is less than three-quarters of the league average) when they've played teams with winning records this season.

Atlanta Hawks Pace and Efficiency Graphs

Looking at the point scored and allowed per 100 possessions from each game, we can see that Atlanta's best defensive results of the season have been concentrated in the last 16 games:

click to enlarge

Here's the game-by-game season average efficiency graph for the Hawks this season which shows both the team's improving defensive efficiency and the team's declining offensive efficiency:

click to enlarge

Sine I've done the work, here's a graph of the team's offensive possessions from each game:

click to enlarge

And the game-by-game season average of offensive possessions:

click to enlarge

The Hawks are again 27th in the league in pace this season.

Josh Smith, Jump Shooter? A Graphical Update

Previously: 12.8.10, 11.24.10, 11.9.10

The green line shows Smith's eFG% on all shot attempts outside of sixteen feet over the past four seasons. The red line shows Smith's eFG% on long (16-23 feet) two-point jumpers over the past four seasons. As you can see there was little difference in his results over that time.

The yellow and blue lines represent Smith's game-by-game season averages this season.

click to enlarge

Further updates for chart fiends...

Josh Smith	3PTA/FGA
2005-09	11.2%
2009-10	0.7%
2010-11	13.1%

Smith's hot start (12-52) from beyond the arc has lifted his career 3PTFG% to 28.1%.

Smith is making a significantly higher percentage of his long two-point jump shots so far this season.

Josh Smith	eFG% (16-23')
2006-07	32.7
2007-08	30.9
2008-09	33.7
2009-10	28.7
2010-11	41.1

He's both taking and making a greater percentage of his long jump shots so far this season.

Josh Smith	%FGA (16+')	eFG% (16+')
2006-07	40.3	34.5
2007-08	30.4	32.9
2008-09	30.8	37.4
2009-10	24.7	27.9
2010-11	41.3	50.6

Smith has continued to take a lower percentage of his shot attempts at the rim than in previous seasons. Some of that may be due to a data collection change at Hoopdata. Last season, 32.5% of shots in the NBA were categorized as at the rim. This season, 27.3% of shots in the NBA have been so categorized.

More of it has to do with shot selection. Smith has made just 10 and attempted just 23 shots at the rim over the last seven games, a stretch during which he's taken 41 jump shots outside of 16 feet. He's eFG% on those jump shots is 40.2%. All of which goes a fair way toward explaining why Smith has seen improved results as a jump shooter yet his eFG% this season is slightly lower than last season's.

Josh Smith	%FGA (at rim)	eFG% (at rim)
2006-07	43	63.5
2007-08	42.9	66.3
2008-09	43	69
2009-10	54.3	65.6
2010-11	30.7	63.2

Updating Jeff Teague

Yesterday's post on Jeff Teague's simultaneous improvement and diminishing playing time deserves, in the wake of his 18 point, 3 assist (and 4 turnover), 3 steal, 25 minute performance last night in Boston, an immediate update.

Teague is now scoring more and scoring more efficiently than he did in his rookie season.

Jeff Teague	Pts/36	TS%	eFG%	2PTFG%	3PTFG%	FT Rate
2009-10	11.4	45.9	41.2	42.6	21.9	18.1
2010-11	12.1	54.9	48.9	49.4	30.0	27.6

Teague's turnover rate rose last night, but so did his assist and defensive rebound rates.

Jeff Teague	Ast%	TO%	OR%	DR%
2009-10	25.1	16.7	0.8	10.2
2010-11	25.5	17.5	1.1	13.1

Teague remains second on the team in steal rate and fourth in block rate.

Teague's PER this season is up to 14.9. The league average is 15.0.

The graph of his (game-by-game) minutes played this season look like this:

click to enlarge

Examining the Playing Time of the Draft Picks

Jeff Teague

When he's gotten to play, Jeff Teague has had an encouraging sophomore season. Especially considering the lack of development he received in his rookie campaign. Teague's scoring rate is down slightly this season but he's been much more efficient both by shooting better from the field and by increasing his free throw rate.

Jeff Teague	Pts/36	TS%	eFG%	2PTFG%	3PTFG%	FT Rate
2009-10	11.4	45.9	41.2	42.6	21.9	18.1
2010-11	11.0	52.3	45.4	45.5	30.0	28.9

Small improvements in Teague's rebound, assist, and turnover rates have helped also:

Jeff Teague	Ast%	TO%	OR%	DR%
2009-10	25.1	16.7	0.8	10.2
2010-11	25.3	16.3	1.2	12.5

Teague's defensive numbers are more of a mixed bag. He's second* on the team in steal rate and fourth on the team in blocked shot rate (behind only Josh Smith, Al Horford, and Zaza Pachulia) but the team is allowing 4.8 more points per 100 possessions with him on the court than with him on the bench. How much of that is due to the impact of garbage time or playing the bulk of his minutes when Josh Powell and/or Jamal Crawford are also on the court is an open question. The Hawks were slightly better defensively with Teague on the floor last season.

*Third if you include Damien Wilkins.

None of the above are great numbers, but it's hard to knock demonstrable improvement in irregular minutes. Teague's PER so far this season is 13.6. Below average, but better than Mike Bibby has managed so far this season. Or last season, for that matter.

All in all, I don't think Teague has been so untrustworthy in his play as to have a graph of his (game-by-game) minutes played this season look like this:

click to enlarge

Jordan Crawford

Jordan Crawford hasn't played a single minute since Joe Johnson had elbow surgery. Etan Thomas has dressed instead of Crawford the Younger four of the last five games. Crawford showed flashes of NBA talent in the 100 minutes he played over the first 18 games of the season, primarily through a decent scoring rate and high usage rate (though his scoring efficiency was poor). That's the sort of profile a young player could build on. And he needn't build on it in the NBA. If he's not a part of the team's immediate plans, Crawford seems an ideal candidate for a D-League stint. He won't struggle to get shots at any level of basketball.

Pape Sy

The Hawks went to the bother of buying out Pape Sy's contract this Summer but he still hasn't played a single minute of professional basketball in North America. Sy missed 18 games due to a back injury and hasn't dressed for the other eight games the Hawks have played. I seriously doubt Sy is an NBA player right now so with the warm, comforting veteran-ness of Damien Wilkins safely ensconced in the rotation, wouldn't it make sense to familiarize Sy with Provo, UT and get a better idea of what exactly the Hawks have gotten themselves into?

SBNation: Ziller: Mapping the NBA's Best Teams

Tom Ziller has a visual representation of the 30 NBA teams based on their offensive and defensive efficiency. The good news: the Hawks are one of just seven teams in the good/good quadrant. The bad news: their closest neighbor in that quadrant is Portland or Utah rather than Dallas, Boston, Miami, San Antonio, or the Lakers.

Josh Smith, Jump Shooter? A Graphical Look

Previously: Josh Smith: Jump Shooter? and Redux

The third bi-weekly update of Josh Smith's improvement as a jump shooter, this one coming after two consecutive hot shooting performances which leave him close to setting a career high in scoring rate and on pace for a career-best TS%, both despite a career-low FT Rate.

The graph below demonstrates both how much more often Smith is making jump shots in general and how much of the value of that better shooting is wrapped up in Smith making 46.9% of his three-point attempts rather than in making 42.4% of his long two-point jumpers.

click to enlarge

The green line shows Smith's eFG% on all shot attempts outside of sixteen feet over the past four seasons. The red line shows Smith's eFG% on long (16-23 feet) two-point jumpers over the past four seasons. As you can see there was little difference in his results over that time.

The yellow and blue lines represent Smith's game-by-game season averages.

The blue line shows that Smith has been shooting below 40% on his long (16-23 feet) two-point jumpers much of this season, before spiking over the last two games, when he's made 9 of 12 shots from that range.

The yellow line shows Smith's eFG% on all shot attempts outside of sixteen feet this season. Here, the value of the third point is clear. Smith is roughly as effective a jump shooter at 23 feet as he is at 20 feet so it wouldn't be a bad idea, if he's going to shoot this volume of jump shots, for his three-point attempts to increase, for him to take a step back on those 20- or 21-foot two-point jumpers and receive a greater reward for the risk of the jumper.

The 16- or 17-foot jumper has some value, regardless of whether or not it goes in, in setting up Smith's pump fake and should never disappear completely from Smith's game.

Now, the rest of the charts that have appeared in the previous installments of this series...

Josh Smith	3PTA/FGA
2005-09	11.2%
2009-10	0.7%
2010-11	10.8%

Smith's hot start (15-32) from beyond the arc has lifted his career 3PTFG% to 27.8%.

Smith is making a significantly higher percentage of his long two-point jump shots so far this season.

Josh Smith	eFG% (16-23')
2006-07	32.7
2007-08	30.9
2008-09	33.7
2009-10	28.7
2010-11	42.4

He's both taking and making a greater percentage of his long jump shots so far this season.

Josh Smith	%FGA (16+')	eFG% (16+')
2006-07	40.3	34.5
2007-08	30.4	32.9
2008-09	30.8	37.4
2009-10	24.7	27.9
2010-11	39.4	53.7

Unsurprisingly, Smith has continued to be devastatingly effective at the rim even the percentage of his shot attempts taken there have declined though some of that may be due to a data collection change at Hoopdata. Last season, 32.5% of shots in the NBA were categorized as at the rim. This season, 27.1% of shots in the NBA have been so categorized.

Josh Smith	%FGA (at rim)	eFG% (at rim)
2006-07	43	63.5
2007-08	42.9	66.3
2008-09	43	69
2009-10	54.3	65.6
2010-11	32	67.4

Looking Back To Looking Ahead: The Four Factors

Before the season started, I made some predictions about how the four factors would change for the Atlanta Hawks this season. I'm not interested in evaluating those predictions this early* in the season but, in the first of a recurring series, let's track how the Hawks are performing relative to last season's 53-win team.

*Not least of which because, according to John Hollinger's power rankings, the Hawks have played the league's fifth-easiest schedule thus far. Granted, the end of season power rankings at the Basketball-Reference.com Blog in April ranked the 2009-10 Hawks schedule the sixth-easiest in the league.

First, the offense:

Name	Off Eff	eFG%	FT Rate	OR%	TO%
2009-10	113.4	50.6	21.3	28.2	13.3
2010-11	109.8	51.6	23.2	25.0	15.1

Unsurprisingly, the Hawks haven't been able to match last season's offensive efficiency. Then again, they don't have to do so to remain a good team. They've declined only to fifth in the league as improved shooting from the field and more frequent trips to the foul line have mostly compensated for a decline in offensive rebound rate and an increase in turnover rate.

The team's turnover rate, though far higher than last season's, has improved significantly over the last two-and-a-half weeks.

click to enlarge

Now, for the defense:

Name	Def Eff	eFG%	FT Rate	OR%	TO%
2009-10	108.3	49.7	28.0	27.2	15.4
2010-11	106.3	48.8	27.6	26.8	14.4

The Hawks have been slightly better defensively (so far) this season. The changes are less dramatic than cumulative. Opponents are missing slightly more shots and the Hawks are rebounding slightly more of those misses. Most importantly, the Hawks haven't seen a spike in their opponents' free throw rate despite their opponents making 78% (league average (76.5%) of their free throws this season compared with 74.5% (league average: 75.9%) last season.

Premature conclusions: the Hawks are 1.6 points per 100 possessions to the worse so far this season compared to last but are, as expected, a fundamentally similar team.

State of the Team, Or, A Four Factors Graph Dump With Commentary

The Hawks have been treading water for 30 games, though, I guess, in terms of this metaphor, it should be made clear they're not treading in dangerous water, winning a little more than 63% of their games and almost certainly assuring themselves a top-four seed in the East. As this graph of the team's game-by-game winning percentage demonstrates, even while the level of accomplishment is clear, the lack of progress is not simply a trick of the mind.

click on graph to enlarge

The flattening out of the team's offensive and defensive efficiency is something of a result of the method employed (The variance in the team's season average will diminish as each individual game has less impact on the total.) but it's, again, useful to have some demonstrative proof that, outside even the organization's willingness to avoid change of roles and personnel whenever reasonably possible, this season's team both is what it is and has been that for more than half of the season-to-date.

click on graph to enlarge

For a different look at (essentially) the same information, here's the graph of the team's point differential per 100 possessions.

click on graph to enlarge

Since the big dip between games 27 (the blowout win in Minnesota) and 36 (the first loss in Orlando), the differential has stayed between 4.5 and 6.2 points per 100 possessions. Over the last 17 games, the differential has stayed between 4.5 and 5.3 points per 100 possessions as the Hawks have gone 10-8.

Is there anything to be learned from the shape of the Hawks' performance in different areas of the game? Let's look at the four factors.

First, the most important factor: eFG%.

click on graph to enlarge

The hot shooting of the opening night game (Atlanta shot 56.9%; Indiana shot 60.3%) extends the graph a touch higher than would be ideal to see the smaller, but real, changes in performance later in the season but the fact remains the Hawks' season eFG% has stayed between 49.8% (reached twice) and 52% (reached once) since the second game of the season. Their opponents have been slightly more variable but one can restrict the range to between 49.4% and 52.2% since the fourth game of the season.

The difference, though, between Atlanta's eFG% and that of their opponents is currently 0* which is the single greatest reason for the constancy of the team's winning percentage and the narrowing of their scoring differential. After Atlanta's 11-2 start, their eFG% was two percentage points higher than their opponents. At their subsequent peak for the season (19-6, 76 Win%) the difference was a percentage point and-a-half.

*Actual difference: 0.00065 in favor the Hawks. Fist pump?

Both Atlanta's offensive rebounding and offensive rebounding prevention (aka "defensive rebounding") have improved since the last time I graphed them.

click on graph to enlarge

That the recent hot rebounding streak has coincided with the Hawks winning no more than five of eight games (62.5%, essentially (or less than depending on your preferred degree of exactitude) the treading water rate referenced above) does little to make me re-think the mea culpa I posted a couple days after 2.7.10 rebounding graph was first seen in this space.

A low turnover rate has been this team's defining characteristic for much of the year. Not turning the ball over isn't especially sexy but the team's remarkable ability to maintain such a low turnover rate is, I contend, mildly titillating.

click on graph to enlarge

That's a turnover rate of 14% or below since game 7, a turnover rate of 13.6% or below since game 19, and a turnover rate of 13.3% or below since game 46.

Here's where a free throw rate graph would go had I column in my game-by-game spreadsheet for the free throw rate of Atlanta's opponents ((oppFTM*100)/oppFGA) rather than or addition to the column which exists to track Atlanta's opponents' FT Rate ((oppFTA*100)/oppFGA) which is there in the interests of comparing foul rates between teams more than comparing opponents' free throw percentage--a vestige of the spreadsheet's beginnings as a tool to compare college basketball teams and their far more diverse ability to shoot free throws.

Let it be said that, as bad as they are at getting to the line, the Hawks are better at keeping their opponents off the line. The difference (on the season) in (like-for-like) FT Rate between the Hawks and their opponents is 1.7 made free throws per 100 field goal attempts. At the rate the Hawks and their opponents attempt field goals, that comes out (on average) to a 1.4 point advantage to the Hawks per game.

Some of that difference is down to free throw shooting. The Hawks make 76.6% of their free throws. Their opponents make 74.7%. The league average is 75.8%.

Above, I declared the diminishing of the difference between the Hawks' eFG% and their opponents' eFG% to be "the single greatest reason for the constancy of the team's winning percentage and the narrowing of their scoring differential." The next greatest reason (and it's next greatest because it only applies strongly to the issue of scoring differential) is the decline in the number of possessions in games as the season has progressed.

click on graph to enlarge

When the Hawks were 11-2 (84.6 Win%), they averaged 92.2 possessions per game*. When they were 19-6 (76 Win%), they averaged 90.9 possessions per game. At 40-23 (63.5 Win%), they average 89.5 possessions per game.

*Both here and in the graph, "per game" means "per 48 minutes" so as to eliminate the impact of overtime games.

Most of the time, the Hawks are better than their opponents. The fewer possessions in a game, the fewer opportunities the better team has to exert its authority. That's generally true. Specifically for the Hawks, higher possession games are also indicative of better defensive rebounding, a commitment to transition offense, and a half-court offense more reliant on the (relatively) quicker-strike virtues of player and ball movement rather than dribbling in place isolated from teammates. Basically all the things that warm us when present and inflame us when absent.

Hoopdata: Shot Location in Motion

Tom Haberstroth explains:

It's a customizable toy displaying each player's shot location data on a per game basis. All the data from the Hoopdata player shot location page is at your fingertips. And yes, it has a play button.

Fun to play around with once you get the hang of all your options. Look at Josh Smith at the rim and from 16-23 feet (or his three-point eFG%) and you'll see how his improvement has been in shot selection far more than any change in the rate at which he makes any particular category of shot.

Josh Smith FG% By Shot Type

Pretty self-explanatory...

click on graph to enlarge

Season-Long Theme: Rebounding

Previously: 11/9/09

Here's the game-by-game season average (i.e. the season average at the end of each game) offensive rebound rates for the Hawks and their opponents for games 1 through 48.

click on graph to enlarge

That's not an encouraging trend.

The Least Inspiring Post-Players-Only Meeting Quotes Ever?

A players-and-one-objective-observer-only meeting might have produced, if not better, more relevant quotes. We'll have to wait and see with regard to on-court results but full credit to the player or players who chose to hold this meeting right before the Hawks host the Nets. Someone(s) doesn't miss a trick.

Joe Johnson on the need to:

"...get back to playing with intensity, having fun. That's pretty much the difference. It seems as if we're just kind of out there, just going through the motions."

Marvin Williams:

"It is disappointing sometimes when you know you can do better and you're not."

Al Horford:

"It's just a lack of effort. People have to make the extra rotation, the extra offense. It has something to do with offense. If guys are not making shots, they tend to get a little discouraged. We still have to defend and rebound to give ourselves a chance to be in the game.

You've got to get back to being more consistent on defense. It's something we've been harping on all year that we've been doing a good job, but we have to put a whole game together."

The Hawks defense has been consistent this year (at least by my definition of consistent). They've slipped a bit (now at 1.073 points allowed per possession for the season) in losing five of the last six but stood at a season-tying best 1.057 points allowed per possession after the win in Minnesota.

Click on graph to enlarge.

The offense, however, peaked a couple of games earlier (1.16 points per possession after the Memphis win) and has been consistently dropping (now at 1.135 points per possession) despite long stretches of good offensive play in every game but the last.

Click on graph to enlarge.

Marvin Williams: Scoring and Efficiency

click to enlarge graph

Marvin Williams continues to have a disappointing season offensively, both in terms of volume and efficiency but the shape of his production has changed a bit over the past few weeks.

Williams began the year in a terrible shooting slump, both from beyond the three-point line (making just 4 of 17 shots, 23.5%) and inside the arc (26 of 61, 42.6%) through 11 games. Williams was scoring at 72% of the rate (per possession) he averaged in the 2008-09 season and only managing that output thanks to a great performance at the free throw line (88.2 FT%, 38.5 FT Rate).

That was the low point of Williams' season. His scoring rate, though still well below last season's, has risen from 15.9 to 18.6 points per 100 possessions. He has made 14 of his last 29 three-pointers to raise his 3PTFG% to 39.1. His 2PTFG% has improved a touch, up to 44.3% on the season (though it got as high as 48.5% following his excellent back-to-back scoring nights against Houston and @New Orleans).

It's not all good news. As his field goal shooting has improved, Williams has failed to get to the free throw line at anywhere near the rate one would expect of him. He's attempted just 17 free throws over the last 9 games which has dropped his FT Rate* from 38.5 to 23.7 and kept his True Shooting Percentage (TS%) from equaling last season's mark.

I suspect the lesson to draw from this (beyond the sample size caveats) is that Marvin Williams is a young player still finding both his place in the league, in terms of what he can and cannot do against defenses, and his place in his team's offense. Expect variable performances.

*League average FT Rate in 2009-10: 23.1. Marvin Williams' FT Rate, 2008-09: 35.4

Offensive/Defensive Efficiency Graph

As the fog that results from mainlining cold medicine for the better part of a week dissipates, sample-sizes grow, and the Hawks have a couple of off days, it's graph time. First up, the game-by-game season average of points scored and allowed per possession.

click graph to enlarge

The sixth game of the year, represented by that sharp offensive dip was the loss in Charlotte.

More graphs to come. Suggestions, questions, and criticism always welcome in the comments.

Josh Smith: 2009-10 Breakout Candidate

Now that Josh Smith has renounced the three-point shot, the discussion about what this means for 2009-10 centers on whether the relatively small number of three-pointers Smith attempts will be replaced by shots he's more likely to make.

First let's look at Smith's overall shooting percentages. The graph below shows Smith's career eFG% and TS% at the conclusion of each of his five seasons.

(click on graph to enlarge)

Without eliminating, either by pronouncement or action, the three-point field goal Smith has steadily improved both his eFG% and TS% over the last three seasons. How so? By reducing the percentage of his field goal attempts which are two-point jump shots.

(click on graph to enlarge)

Reducing the two-point jump shots helps because, though Smith makes a higher percentage of those than his three-point attempts, the difference is not enough to overcome the difference in value between the two shots.

(click on graph to enlarge)

For his career, Smith has made 36.2% of 1748 two-point jump shots, providing the Hawks with 0.72 points per possession (before accounting for offensive rebounding*), and 27% of his three-point shots, providing the Hawks with 0.81 points per possession (before accounting for offensive rebounding). Over five seasons, Smith attempting a three-point jumper has been the lesser of two evils for the Hawks.

*The value of all the offensive rebounding opportunities created by Smith's missed jumpers is debatable, though likely marginal. Over the last two seasons the Hawks have had a higher offensive rebounding rate when Smith is off the court than when he is on the court which would, I believe, support the argument against Smith taking low-percentage shots.

In the interests of thoroughness, let me digress by showing that the upward trend in Smith's True Shooting Percentage (TS%) is not fueled by his free throw shooting. Smith's FT Rate has been remarkably consistent throughout his career...

(click on graph to enlarge)

...despite fluctuations in his FT%.

(click on graph to enlarge)

Yes, Smith's 58.8% shooting from the foul line last season reduced his career FT% by almost two-and-a-half percentage points. A return to competence at the free throw line would, by itself, significantly increase Smith's offensive production and efficiency, though not as much as turning as many of his jump shots into closer two-point attempts. Smith made a career-high 64.4% of non-jump shot two-point attempts last year. Smith's increases in eFG% and TS% stems directly from simultaneously taking and making more non-jump shots over the last three years.

(click on graph to enlarge)

Sharp-eyed readers might notice that the lower bounder of this graph is 52% where the upper boundary of the two- and three-point field goal percentage graph is 40%.

Smith has used more than half of his career field goal attempts on shots he's made (collectively) 34% of time and less than half of his career field goal attempts on shots he's made 57% of the time. Shot selection that poor provides Smith with ample opportunity for improvements in both scoring and efficiency, as does last season's inexplicably poor free throw shooting. Furthermore, the ankle injury that cost Smith 13 games likely contributed to Smith posting career lows as an offensive rebounder and shot blocker (both per minute and per opportunity) and failing to match his career average on the defensive glass.

It's difficult to identify an area of his game where Smith is more likely to regress than he is to improve in 2009-10. If he's really beginning to the see the light regarding his own strengths and weaknesses he could make the leap from talented but maddening player to legitimate All-Star.

Sources: 82games, Basketball-Reference.com

A Bad Prediction Rendered Visually

BallHype reviews the blog previews.

34 wins. Oops.

Series Preview: Hawks vs. Heat

NOTE: If you prefer this content in briefer, bullet point form, scroll to the bottom section marked "CONCLUSIONS"

First, the four factors (plus points per 100 possessions leading things off) for the Atlanta offense (red) and the Miami defense (black) from the 2008-09 season.

As always, click on the graphs to make them bigger.
The particulars of that graph, in table form.

Team	Poss	PPP	eFG%	FT Rate	OR%	TO%
ATL (off)	89.6	1.093	50.4	23.8	26	12.5
MIA (def)	89.9	1.076	50.1	25.1	27.1	14.5

The teams play at almost identical paces* and the only real rate difference in any of the four factors is turnover percentage. The Heat are (slightly) below average in three of the four factors but maintain a (slightly) above average defense through their ability to force turnovers.

*You'll see below in the head-to-head graph/table that both prefer to slow the pace and their games against each other have been even slower than their respective (slower than league average) 82-game average paces.

Next, the same as above for the Miami offense (still black) and the Atlanta defense (still red).
The particulars of that graph, in table form.

Team	Poss	PPP	eFG%	FT Rate	OR%	TO%
ATL (def)	89.6	1.076	49.4	21	28.4	13.2
MIA (off)	89.9	1.078	50	21.2	24.6	12.1

Offensively, as defensively, Miami's greatest strength this season was their turnover percentage. However, Miami's ability to hold onto the basketball wasn't enough to overcome three other below average factors and maintain an above average offensive efficiency. Most interesting (to me, at least) is that Miami, despite Dwyane Wade playing 38.6 minutes of each of 79 games, had the fourth-lowest free throw rate in the league.

The reason why is fairly obvious. No one else on the team gets to the line.

	FT Rate
Wade	33.9
Rest of Team	16.7

Now we look at the numbers from the three head-to-head matchups (two games in Miami, one in Atlanta) between the two teams.
The particulars of that graph, in table form.

Team	Poss	Off Eff	eFG%	FT Rate	OR%	TO%
ATL	88.8	0.965	42.8	30.8	29.4	17.7
MIA	88.8	0.941	45.3	17.2	22.1	14.6

This is only three games* worth of data so consider the following conclusions to be drawn in pencil**. The bulk of the difference in eFG%, relative to their season averages, for both teams comes from terrible three-point shooting and lots of it. Head-to-head both teams took more three-point attempts as a percentage of their field goal attempts than they averaged on the season as a whole despite neither team forcing their opponents to shoot a lot of three-pointers on the season.

*For obvious reasons I ignored the dead rubber Tuesday night.

**Lightly and in a sloppy hand, as well.

Here's their 3PTA/FGA broken down three ways.

Team	Head-to-Head	Season Avg	Opp Season Avg
MIA	27%	24.5%	23.7%
ATL	26.2%	25.3%	22.6%

Combine the extra three-point attempts with Atlanta making just 22.4% of their attempts (compared to a season average of 36.6%) and Miami making just 25.8% of their attempts (compared to a season average of 35.7%) and we're a ways toward explaining the offensive struggles of both teams before even mentioning all the turnovers that occurred in the head-to-head matchups.

Both teams had success in getting guys on the other team they wanted to shoot three-pointers to attempt a good number of them: Josh Smith was 1-7 for Atlanta; Yakubu Diawara was 1-10 and Dwyane Wade 1-6 for Miami. Both teams also got lucky with certain opposing players shooting well below their season averages from beyond the arc: Mo Evans (39.5%) and Flip Murray (36%) both went 0-5 against Miami, Marvin Williams (35.5%) was 3-11, and Mike Bibby (39%) were 4-15; Mario Chalmers (36.7%) was 3-13 and Daequan Cook (38.7%) 9-27 against the Hawks.

Even though both teams forced more turnovers on the season than the other coughed up, both teams forced turnovers at a much higher rate in their competitive head-to-head matchups. This is a bit of a red herring as the second matchup between the teams, Miami's 95-79 win on January 26th wherein Atlanta turned the ball over on 22.4% of their possessions and Miami on 19% of their possessions, serves as significant outliers.

Turnover percentage by game:

Team	Dec 12th	Jan 26th	Feb 27th
MIA	11.7%	19%	13.1%
ATL	15.2%	22.4%	15.3%

Two more encouraging (for Hawks fans*) nuggets found in the head-to-head matchups are the atypical dominance Atlanta maintained on both the offensive and defensive glass and their (relative) success in keeping Dwyane Wade off the free throw line.

As noted above, Wade's FT Rate for the season was 33.9. In the three games against the Hawks it was 23.3. In Atlanta's two wins it was 12.2.

*though, again, of questionable significance

CONCLUSIONS

6 plausible reasons for Atlanta optimism/keys to victory

1. Home-court advantage
2. Out-rebounding Miami is a bonus for a generally poor rebounding team.
3. Keeping Dwayne Wade off the free throw line (at least relative to how much Dwyane Wade wants to be at the free throw line).
4. Force/allow Wade and Diawara to shoot more than their share of three-pointers.
5. Five Hawks made more than 35% of their three-point attempts this season should Miami persist in forcing/allowing Atlanta to shoot a higher than normal number of three-pointers.
6. Chris Quinn

5 plausible reasons for Atlanta pessimism/causes of defeat

1. Miami forces enough Atlanta turnovers to negate any Atlanta gains on the offensive glass.
2. Dwyane Wade career regular season free throw rate: 39.3. Dwyane Wade career playoff free throw rate: 49.8.
3. As the series progresses Mario Chalmers realizes he can beat Mike Bibby off the dribble and becomes an inside/outside threat.
4. Daequan Cook's size turns Flip Murray circa 2008-09 into Flip Murray circa 2004-08.
5. Lingering, limiting back injuries.

Now, what have I got wrong?

Hawks' Efficiency Home vs. Road

The contents of this graph will come as a surprise to no one, I suspect but 1) It's in 3-D and 2) It took me absurdly long time to add even passable yet thoroughly homemade-looking labels. So I'll share.

And I'll summarize: The Hawks are 10.1 points/100 possessions better at home than they are on the road. At the Hawks' average pace (89.8 possessions per game) that converts to 9.1 points per game.

As always, click the image to make it bigger.

I have not expanded my study to figure out how much better the average NBA team is at home. Perhaps another day.

Team 3PTFG%

I think this one is pretty self-explanatory. As always, click on the image to make it bigger.