Trading for a Championship

    Listening to Talking about Practice, Episode 11 this morning, I heard Jared Wade and Rob Mahoney discussing some of the deadline trades from this past NBA season. The context of the discussion was how dissappointing the additions of Antawn Jamison for the Cavs, and Caron Butler and Brendan Haywood for the Mavs, were from a short-term playoff success perspective. One of them wondered aloud if anyone had ever looked at how often the mid-season acquistion of that “big, missing piece” has worked out. I looked around the interwebs this morning, and not finding anything on the topic, decided to look at it myself.

    The table below shows every NBA champion since 1987. I went with 1987 because, as near as I can tell, that is the first season that the NBA had a trade deadline. For each team I looked at players that were acquired by trade, during the season, who played at least 15 minutes per game for their new team over the rest of the season. For each player I included their Minutes per Game, PER, and Win Shares for their new team, broken down by regular season and playoffs.

Team	Year	Players Acquired through Trade	MPG/R	PER/R	WS/R	MPG/P	PER/P	WS/P
Los Angeles Lakers	1987	MyChal Thompson	20.6	12.6	1.0	22.3	11.1	0.7
Los Angeles Lakers	1988	None	–	–	–	–	–	–
Detroit Pistons	1989	Mark Aguirre	29.7	15.1	2.8	27.2	15.2	1.3
Detroit Pistons	1990	None	–	–	–	–	–	–
Chicago Bulls	1991	None	–	–	–	–	–	–
Chicago Bulls	1992	None	–	–	–	–	–	–
Chicago Bulls	1993	None	–	–	–	–	–	–
Houston Rockets	1994	None	–	–	–	–	–	–
Houston Rockets	1995	Clyde Drexler	37.1	22.1	5.2	38.6	21.1	3.0
Chicago Bulls	1996	None	–	–	–	–	–	–
Chicago Bulls	1997	None	–	–	–	–	–	–
Chicago Bulls	1998	None	–	–	–	–	–	–
San Antonio Spurs	1999	Steve Kerr	16.7	9.9	1.6	8.8	6.2	0.0
Los Angeles Lakers	2000	None	–	–	–	–	–	–
Los Angeles Lakers	2001	None	–	–	–	–	–	–
Los Angeles Lakers	2002	None	–	–	–	–	–	–
San Antonio Spurs	2003	None	–	–	–	–	–	–
Detroit Pistons	2004	Mike James	19.7	14.3	1.4	8.9	10.2	0.3
		Rasheed Wallace	30.6	18.8	2.4	35.0	15.3	2.3
San Antonio Spurs	2005	Nazr Mohammed	18.1	14.2	0.8	23.0	16.3	1.7
Miami Heat	2006	Derek Anderson	20.2	8.7	0.6	8.3	8.8	0.1
San Antonio Spurs	2007	None	–	–	–	–	–	–
Boston Celtics	2008	None	–	–	–	–	–	–
Los Angeles Lakers	2009	None	–	–	–	–	–	–

     I was pretty surprised to see that only seven of the last 23 NBA champions had even added a rotation player, through a trade, during the season. Drexler and Rasheed Wallace are the ones who stand out as being most successful. Although Mike James did not play often or well for the Pistons during their 2004 playoff run, he was a significant contributor for them finishing out the regular season. Looking at this table really highlights what an impressive coup it was for the Pistons to add those two players mid-season.

     This is obviously a small sample, and doesn’t really give all the information I was looking for. Sometime next week I will try to dig in a little deeper, and setting aside championships, see if mid-season acquistions have historically had a certain type of impact. Stay tuned . . .

Effective Height and Defending the Paint, Part 2

     Our first look at correlations between Effective Height and defending the paint was earlier this week. That post was inspired by another post from Tom Haberstroh at Hardwood Paroxysm comparing Effective Height to team rebounding rates. To recap, Haberstroh created the metric of Effective Height, which is the average height of a team with the heights of respective players weighted by the number of minutes they played. Haberstroh found moderate correlations between eHeight and Total and Offensive Rebound rates. I then took his eHeight numbers and compared them to a few team defensive categories: Block Rate, Opponents At Rim FG%, and the percentage of an opponents field goal attempts which came at the rim. I was surprised to find essentially no correlation between any of those categories and a team’s eHeight.

    After seeing those results I wondered if the height contributions of a team’s backcourt might be muddying the correlation data. I then calculated the eHeight of each team using just their frontcourt players. To do this, I used the position designations from Basketball-Reference.  There were a lot of teams I didn’t see much of this season, and I didn’t feel comfortable using my own observations and opinions to assign positions to different players. I am sure some of these position designations are not entirely consistent with a player’s role on his team, but for the sake of consistency I stuck with them.

     The first table shows the Frontcourt eHeight compared to Haberstroh’s numbers for each team’s eHeight. I also calculated the difference between the two. The higher the eHeight difference, than the smaller a backcourt the team played with this season. I also included the percentage of a team’s minutes which were played by frontcourt players.

    I want to return the frontcourt minute percentage for each team for just a moment. A traditional basketball lineup features 2 forwards and 1 center, or 3 out of the 5 players. Therefore a minute distribution based on a traditional lineup would have exactly 60% of the minutes being played by frontcourt players. Looking at the numbers here it is easy to see the teams ascribe to the small ball movement. Teams like Houston and Orlando, who have true small forwards (Vince Carter and Trevor Ariza) playing a majority of their minutes at shooting guard, stand out as well.

     The next table shows my Frontcourt eHeight calculations compared to team rebound rates. 

     In each case the correlations which much weaker than they were Haberstroh’s team eHeight. I found this to be extremely surprising. Apparently, I have underestimated the rebounding contributions of a team’s backcourt. I also think it’s interesting that you can almost see the influences of individual players. Orlando and San Antonio don’t have particularly big front courts, but they do have Dwight Howard and Dejuan Blair.

    The last table compares Frontcourt eHeight to those defensive categories I discussed earlier.

     There was an increased correlation with Block Rate, but other than that the correlations were weaker than when using the eHeight of the entire team. Again this would seem to indicate that controlling dribble penetration and challenging shots, have just as much impact as length and height, in defending the paint. I am sure there is much more to this equation than my meager analysis has provided. With an eye towards the NBA Draft it might be wise for teams to focus on the skill set and motor of a player as opposed to falling in love with a physical profile.

Crazy about Colors and Correlations, Pt. 1

      This post came dangerously close to being the second installment of my Random Number Generator series. In the end the numbers weren’t so much random as they were predictable and led to extremely obvious conclusions. I was very intrigued by Tom Haberstroh’s effective height post at Hardwood Paroxysm last week. One bit of minutiae that I loved was the way he color-coded his tables. Now that I have discovered this feature on Excel you can expect to see a lot more of it! I also discovered how to calculate statistical correlations with Excel, and have obviously gone a little overboard.

     With my new found correlating and color-coding skills, I set out to examine the idea of Offensive Efficiency a little bit more. The three most efficient areas on the floor to score from are: at the rim, from the free throw line, and on three pointers. The two former increase efficiency because they present relatively easy scoring opportunities, while three pointers offer the bonus of an additional point per shot.

     I created this table to see which of these areas has the greatest correlation with Offensive Efficiency. I included the FG% for each team at the rim, as well as what percentage of their shots were taken at the rim. The same data is included for three pointers. For free throws I included the percentage made and the average number of attempts per game.

Strong Correlations –

• Correlation between At Rim FG% and Offensive Efficiency:  0.663
• Correlation between 3PT% and Offensive Efficiency:  0.773

Low/Moderate Correlations –

• Correlation between 3PTA% and Offensive Efficiency:  0.399
• Correlation between FTA/g and Offensive Efficiency: 0.390

(Essentially) No Correlation –

• Correlation between FGA% At the Rim and Offensive Efficiency:  -0.075
• Correlation between FT% and Offensive Efficiency:  0.099

     Obviously, At the Rim FG% and 3PT% had very strong correlations with Offensive Efficiency. I am not sure I really need to state this fact. If you make more of your shots, you will average more points per possession. Not sure I needed to state that one either. The categories with low or no correlations were much more surprising. You definitely need to make a high percentage of your shots at the rim to be efficient offensively, but apparently it doesn’t matter much how many shots you actually from that area of the floor. Also making a high percentage of your free throws isn’t nearly as important as attempting a lot of free throws. Good news for Dwight Howard! Stay tuned for the total lack of surprises in Part 2 of this post, when I put my coloring and correlating skills to the test with Defensive Efficiency.

Effective Height and Defending the Paint

     Last week, Tom Haberstroh had an excellent post at Hardwood Paroxysm, examining the relationship between a team’s collective height and their rebounding performance. He gathered the height measurements for each player and then weighted it by the minutes they played for their team. The result was a metric he calls “effective height.” Haberstroh found their to be a decent correlation (.33) between effective height and total rebounding percentage. When looking at rebounding by category there was almost no correlation (-0.03) between effective height and defensive rebounding percentage, but a reasonably strong correlation (0.42) between effective height and offensive rebounding percentage. It’s a terrific post and definitely worth a read.

    I was really intrigued by his analysis, especially by the idea of effective height. I have been baffled recently by a few questions which seem related. One is, How do you predict which college players will become effective NBA shot blockers? The other is, How do you explain the Pacers being an average defensive team with an apparent lack of shot blocking, but still holding their opponents to the sixth lowest At Rim FG% in the league? I decided to use Haberstroh’s effective height data and compare it to three statistics. The first is each team’s block rate, or what percentage of an opponent’s shots they blocked. The second is At Rim FG%, or the FG% each team held their opponents to on shots at the rim. The third is Opponents At Rim%, or what percentage of an opponents shots were taken at the rim.

Team	eHeight	BLKR	At Rim FG%	O At Rim%
SAC	79.85	4.6 (24th)	64.3% (27th)	32.2% (12th)
TOR	79.66	4.9 (22nd)	60.3% (15th)	32.6% (15th)
MEM	79.47	5.0 (19th)	62.5% (22nd)	36.0% (28th)
PHO	79.43	5.2 (15th)	59.8% (11th)	30.4% (8th)
CLE	79.40	5.5 (9th)	58.3% (7th)	29.3% (2nd)
ATL	79.36	5.4 (10th)	60.7% (16th)	33.7% (22nd)
OKC	79.25	6.1 (1st)	60.1% (13th)	36.1% (29th)
NJN	79.17	5.1 (17th)	61.7% (20th)	37.2% (30th)
MIN	79.11	3.8 (28th)	66.0% (30th)	31.5% (10th)
LAL	79.09	5.1 (16th)	59.8% (10th)	32.3% (13th)
LAC	79.09	6.0 (3rd)	61.7% (19th)	33.2% (20th)
WAS	79.04	5.4 (12th)	65.9% (29th)	29.1% (1st)
IND	79.01	5.4 (11th)	58.3% (6th)	29.8% (5th)
POR	78.99	4.7 (23rd)	61.9% (21st)	33.5% (21st)
UTH	78.96	5.1 (18th)	60.1% (14th)	33.0% (17th)
DET	78.95	4.2 (25th)	62.8% (24th)	33.1% (18th)
NYK	78.95	3.8 (30th)	63.5% (25th)	34.2% (24th)
CHI	78.87	6.0 (2nd)	56.8% (1st)	33.2% (19th)
ORL	78.76	5.9 (5th)	57.4% (2nd)	29.4% (3rd)
DEN	78.70	5.2 (14th)	60.1% (12th)	32.5% (14th)
DAL	78.56	5.7 (7th)	61.1% (17th)	30.0% (7th)
SAS	78.44	4.9 (21st)	58.1% (5th)	31.1% (9th)
CHA	78.44	5.8 (6th)	59.0% (9th)	34.2% (25th)
NOR	78.44	3.8 (29th)	64.9% (28th)	34.5% (26th)
PHI	78.44	5.7 (8th)	61.1% (18th)	34.6% (27th)
MIL	78.30	4.9 (20th)	58.1% (4th)	33.0% (16th)
MIA	78.28	6.0 (4th)	58.1% (3rd)	29.8% (4th)
BOS	78.25	5.2 (13th)	58.5% (8th)	29.9% (6th)
HOU	78.21	4.0 (27th)	62.7% (23rd)	33.8% (23rd)
GSW	77.91	4.0 (26th)	64.3% (26th)	31.6% (11th)
Average	78.88	5.1	61.0%	32.5%

     I was surprised to find in each case there was almost no correlation whatsoever.

• Correlation between eHeight and BLKR:  0.096
• Correlation between eHeight and At Rim FG%:  0.133
• Correlation between eHeight an O At Rim %:  0.128

     Obviously, it takes more than height to defend the basket well. These numbers would seem to emphasize the importance of denying dribble penetration and challenging shots, even if you can’t block them. In the comments section of the Hardwood Paroxysm post someone mentioned the idea of calculating the effective height of just the front court players of each team, as having tall backcourt players can greatly increase a team’s effective height but may not make a difference in rebounding number,s since they spend much of their time away from the basket. Hopefully, I can crunch those effective height numbers for just frontcourt players soon and try running them against these three statistical categories. Perhaps we will see a stronger correlation. Stay tuned . . .

Random Number Generator Series – Forecasting Shotblocking

A random number generator (often abbreviated as RNG) is a computational or physical device designed to generate a sequence of numbers or symbols that lack any pattern, i.e. appear random.

     This is how I end up feeling about a lot of my posts here at Hickory High. I often begin with a question and start assembling data that I think may help me find an answer. With my underdeveloped math and logical reasoning skills, this often leads me to assembling several tables or charts with no actual connection to the answer I was seeking. Usually these failed inquiries end up in the trash, but this one took a lot of work. Maybe someone can take what I have started and finish it up for me.

    This will be the first of my Random Number Generator Series; a, hopefully, sporadic series of posts where I pose a question and then assemble a bunch of numbers which don’t really answer said query.

    Today’s question: How do you forecast shotblocking? My favorite team, the Pacers, is in search of shot-blocking in the NBA Draft this year. There are several players with gaudy block numbers available in the draft this year. How can you predict who will be able to translate this skill to the NBA?

     Since the 2001/2002 season, there have been 409 individual player seasons in which a player averaged 3.0 or more blocks per 40 minutes in Division I college basketball. If we subtract duplicate seasons by the same player, and players who are either still in college or entered in the draft this year we are left with 209 players. Of those 209, only 43 played even 1 minute of NBA action, and only 11 (Hasheem Thabeet, Sean Williams, Emeka Okafor, Javale McGee, Joakim Noah, Danny Granger, Joel Anthony, Roy Hibbert, Chris Kaman, Jason Maxiell, and Taj Gibson) have ever totaled 82 blocks in a single season. That would be an average of one block per game across an entire season. Robin Lopez and Greg Oden probably would have made it this season if not for injuries. Channing Frye and Jason Thompson have each had seasons in the 70s. Even with those players included that’s only 15 of the 209 best college shot-blockers over the past nine years, or 7.1%, who have become a shot blocking presence in the NBA. Obviously, this number could go up somewhat when this year’s draft class is included.

      The statistic of blocked shots in college doesn’t seem to be an adequate stand alone indicator of who will be a productive shot blocker in the NBA. Injuries are a huge variable to this equation, one that is largely out of the hands of the individual players. As strange as it sounds, scoring and rebounding turn out to be variables as well. If you can only block shots, and can’t help out in other areas of the game then you’re a long shot to get drafted, let alone see any playing time. All 11 of the players I mentioned above averaged at least 8.5 points and rebounds per 40 minutes in the season they totaled 82 blocks. Size of the player and size of the college would also be factors. Players at smaller schools, playing against less talented competition can rack up blocks even if their own physical and athletic shortcomings would prevent them from doing the same things in the NBA.

     35 players in the NBA blocked at least 82 shots this season. Of those 35, 12 either came from Europe or directly from high school and never played college basketball. I couldn’t find college statistics for Chris Andersen or Ben Wallace, but I am going to assume their numbers were astronomical considering the competition level they were playing against. Below is a table showing the 23 players with at least one season of college experience, who also blocked at least 82 shots this season. Included is their Blk/40 from their last year in college, and their Blk/40 from this season.

Name	Total Blocks ’09-‘10	Blk/40 ’09-‘10	Blk/40 College
Andrew Bogut	175	3.1	2.1
Brendan Haywood	158	2.7	5.3
Samuel Dalembert	151	2.8	6.7
Marcus Camby	146	2.5	5.1
Chris Andersen	143	3.3	–
Brook Lopez	139	1.8	2.7
Roy Hibbert	131	2.6	3.4
Emeka Okafor	127	2.1	5.0
Tim Duncan	117	1.9	3.6
Joel Anthony	109	3.3	6.5
Taj Gibson	104	1.9	3.4
Javale McGee	101	4.2	4.1
Joakim Noah	100	2.1	2.8
Paul Millsap	99	1.7	2.7
Chris Kaman	94	1.4	3.7
Al Horford	91	1.3	2.6
Hasheem Thabeet	89	4.0	6.2
Tyrus Thomas	85	2.8	4.8
Kevin Durant	84	1.0	2.1
Ben Wallace	84	1.7	–
Spencer Hawes	83	1.7	2.4
Gerald Wallace	83	1.0	1.8
Dwayne Wade	82	1.2	2.0

     

     So except for Andrew Bogut, the outlier, no one seems to block more shots in the NBA; not a gigantic surprise given the increased talent level. But when looking at players like Hassan Whiteside, Ekpe Udoh, Ed Davis and Jarvis Varnado who are available in this year’s draft, how do we project if they will end up on this list? Is this just a question of being talented enough in other areas to make a roster and earn minutes? Is there anything besides traditional scouting to aid in this evaluation? Are there any statistics which can be used to project their ability to bring this skill set to the NBA? I don’t know, I’m just a Random Number Generator for today. Anyone else have any ideas?

Franchising Defense

     The NBA announced their All-Defense teams this week, and once again, none of the Indiana Pacers made the list. I remembered Artest making the team as a Pacer, but other than that I couldn’t remembered anyone else from the team ever making the list. I checked it out and came up with a list of All Defensive selections by franchise. The list is by franchise, so all of the Seattle Supersonics’ selections count for Oklahoma City, etc.

Franchise	All Defensive Selections
Los Angeles (Lakers)	38
Chicago	37
Boston	35
San Antonio	34
Detroit	24
Atlanta	22
New York	22
Philadelphia	18
Oklahoma City	17
Phoenix	17
Utah	17
Milwaukee	16
Houston	15
Portland	14
Denver	11
Golden State	11
Cleveland	10
Indiana	8
Miami	8
Minnesota	8
Washington	8
New Jersey	7
New Orleans	7
Sacramento	7
Orlando	5
Dallas	2
Charlotte	1
Los Angeles (Clippers)	0
Memphis	0
Toronto	0

     Turns out the Pacers weren’t as bad as I thought. Artest had three selections in a row, ’02-’06. Derrick McKey also made two in the early 90s. Another interesting thing, is how much the identity of a franchise changes from era to era. Phoenix and Golden State are two of the fastest run and gun teams of the last decade, but apparently had a much bigger defensive focus during the 80s and early 90s.

     Congratulations to Gerald Wallace, who this year earned the first selection for the Charlotte Bobcats. Now it’s just the Clippers, Grizzlies and Raptors who have never had an All-Defense selection. Hedo Turkoglu, Zack Randolph and Baron Davis, next year is your year! Remember to get low and move your feet.

Kodachromatic Genius

How much do your sneakers cost?

     Yesterday, Ball Don’t Lie began their list of the Top-20 Bloggable Photos of all time with photos #20 – #16. It only took me a few minutes, but it was one of the most entertaining posts I have seen in a long time. Of course I disagree, I am not sure how they could top the above photo (Kirby only had it ranked as #20!). Have a look and a laugh and stay tuned for the rest of the list.

 

Another Pony for the Carousel

He's got intensity and terrific hair!

 The NBA coaching carousel has begun it’s slow revolutions with Chicago, New Orleans, Philadelphia, New Jersey and the Los Angeles Clippers are all looking for replacements. The usual suspects have been rounded up for interviews. There is one name out there I haven’t heard mentioned; a name that I think certainly deserves some consideration: Quin Snyder Snyder has no experience as an NBA head coach, and that is part of the reason he is, and will remain an under the radar candidate. He did spend 1 year as an assistant coach with the Los Angeles Clippers, and 4 years as an assistant with Duke. He then spent 7 season as the Head Basketball Coach at the University of Missouri. During his time there he amassed a .581 winning percentage, a .528 winning percentage in the Big 12, and led his team to the NCAA Tournament 4 straight seasons. He did resign from the job at Missouri after two difficult seasons, a recruiting scandal, and a strained relationship with his athletic director.

For the past three seasons, Snyder has been the Head Coach of the Austin Toros, the San Antonio Spurs’ D-League affiliate. During his tenure there he has amassed a 94-56 record for a .627 winning percentage. Many will scoff at this accomplishment, but each year the talent gap between the NBA and D-League has gotten smaller. Over the past three seasons Alonzo Gee, Curtis Jerells, Dwayne Jones, DeMarcus Nelson, Pops Mensah-Bonsu, Malik Hairston, DeMarr Johnson, Marcus Williams, Andre Barett and Keith Langford have all received call-ups from Austin to the NBA. The Toros are regarded as one of the model D-League programs for developing players and getting them to play within their system.

Being a Head Coach in the D-League seems like an especially challenging proposition. You have a roster full of players, most of whom were the stars of their college teams. They have recently had their egos bruised, finding out the NBA doesn’t think as much of them as they think of themselves. You have to discourage these players for going out and shooting every time they touch the ball, getting them to focus on details and development, as well as learning NBA style systems. You have to get them to buy into the fact that their best chance to make it to the NBA is to find the niche skill they can excel at, and carve out a career as a role player at the next level. This is a difficult task but Snyder has been among the best over the past few years.

What he lacks in top-level head coaching experience, Snyder makes up for in intensity. He seems to connect with his players, and is able to get them to buy in and to play hard and smart every night. He may not be the best candidate out there, but certainly some team with a young roster, and an eye towards player development should be giving him a call for an interview. Of course if it doesn’t happen this season, I will be happy cut and past this post next summer when the Pacers began their search for a new Head Coach.

Stats Glossary Update: 5/4/2010

     I haven’t been posting much lately, as I’ve been spending quite a bit of time working on the Stats Glossary page here at Hickory High. The page is up and running now! It links to my descriptions of PER, eFG%, TS%, Hoopdata’s Shot Location Database, and Queen City Hoops Player Swap Tool. I will continue to link additional pages as I finish them.

     I hope these pages can be a resource for readers here, to learn more about some of the statistics and tools I use in my writing. I am certain that I have made errors of omission, vagueness, analysis and outright understanding. I would love to hear thoughts, feedback, corrections or disagreements from anyone, so I can improve this feature and make it as useful as possible.

     You can link to the parent page on the menu bar above, or by clicking here. Thanks!