Side O’ Stats, Week 15

As you know, I sometimes like to spend my time on the Side O’ Stats challenging “common wisdom” regarding basketball often found in various dark corners of the internet.  This week, I’d like to take on a particular pet peeve of mine: ever since the Kings drafted Tyreke Evans, there has been some vocal opposition to an Evans/Martin backcourt for the Sacramento Kings.

Frankly, I don’t see why such a backcourt can’t work.  Evans and Martin seem fairly complementary to me: Evans spends a lot of time with the ball in his hands, excels at getting to the rim, driving and kicking, and has at least the potential to play good perimeter defense.  Martin does use a lot of possessions, but he does it very efficiently (coming into this season, he was about the 4th most efficient player in NBA history), and he doesn’t hold on to the ball and dirbble pointlessly for long periods of time.  When he gets the ball, he shoots, drives, or passes.  And he’s excellent at hitting threes, which is exactly what the Kings would need to help stretch the defense for Evans’ drives.  Seems like a good match to me, but people get ideas into their heads, and Kevin’s major struggles this season (very out of character, considering his very consistent career up to this point) certainly don’t help things.

Anyway, my least favorite of all the “trade Martin” arguments is the following:  No team with two 20-point scorers in the backcourt has ever won a championship.  Therefore a team with two 20-point scorers in the backcourt can’t win a championship.  Since we expect both Evans and Martin to be 20-point scorers, we should trade Martin because it can’t work (nobody in their right mind would trade Evans right now – insanity only goes so far).

Is this a reasonable argument?  Well, if there were hundreds of examples of double-20 backcourts, the argument might carry some weight.  But it turns out that since 1979 (30 years, and also the institution of the 3-point shot) there have been only 10 teams to have two guards both average 20 points and play at least 50 games each for the same team.  It’s such a short list, I’ll give it to you:

2008: Devin Harris, Vince Carter, NJN

2004: Larry Hughes, Gilbert Arenas, WAS

2002: Jerry Stackhouse, Michael Jordan, WAS

2001: Cuttino Mobley, Steve Francis, HOU

2001: Michael Jordan, Rip Hamilton, WAS

1996: Allen Iverson, Jerry Stackhouse, PHI

1990: Mitch Richmond, Tim Hardaway, GSW

1984: Otis Birdsong, Michael Ray Richardson, NJN

1979: Paul Westphal, Walter Davis, PHO

1979: Otis Birdsong, Ray Williams, NYK

That’s it.  So let’s assume for a null hypothesis, that there is no difference between the probability that a double-20 backcourt will win a championship and the probability that a “normal” backcourt will win a championship.  Given that, we can easily estimate the expected number of championships that a double-20 backcourt would win.  We do have to take into account the number of teams in the league, as this has changed over the years.  We can actually ignore all the years that there were no double-20 backcourts, because there was no probability that a double-20 backcourt could win it all.  Then, a simple Monte Carlo is faster than math, and with a 100,000 iteration simulation, the results are:

Championships	Probability
0	67.90%
1	27.10%
2	4.60%
3	0.40%
4+	0.03%

So there you are.  In fact, given the number of teams that have actually had two 20-point scorers in the backcourt since 1979, we would expect the most likely number of double-20 championships to be zero – exactly the value we find.  In other words, the argument that “nobody has ever won a championship with two 20-point scorers in the backcourt” is a really weak argument.  Given the number of times it has actually been tried, that’s exactly what we would expect.

 

Commissioner’s Crumpets, Week 15

Well, ’tis the week of the Super Bowl, which basically means that nothing notable was bound to happen in the basketball world.  I suppose the big news is always the news that goes under the radar, though, and this week the big news is that the NBA fired off their first proposal to the Players’ Union regarding the Collective Bargaining Agreement, which needs to be renegotiated in the summer of 2011.  The proposal included the following:  Significantly lower percentage of Basketball Related Income going to players; Significantly lower maximum salaries; Shorter maximum contracts; A hard salary cap.

Yeah, that’s going to fly.

Things are brewing for a big-time labor stoppage after next season, so enjoy your DST&CC while you can…

In brighter news, there are new Crumpets!  Read below for details!

The Slamson Memorial Virtual Round Robin

Team	VRR Record	VRR %	Actual Record	Actual %	Difference
Dry Heat	106-33-4	75.5 %	7-5-1	57.7 %	-17.8 %
Bodacious D	92-47-4	65.7 %	9-4-0	69.2 %	3.5 %
Slamson	92-49-2	65.0 %	5-7-1	42.3 %	-22.7 %
Sweet Sassy Molassy	79-63-1	55.6 %	7-6-0	53.8 %	-1.8 %
Lipsticked Pigs	74-66-3	52.8 %	4-9-0	30.8 %	-22.0 %
The Bombay Runners	70-67-6	51.0 %	9-4-0	69.2 %	18.2 %
nutria on viagra	63-75-5	45.8 %	8-5-0	61.5 %	15.7 %
The Very Bad Team	59-80-4	42.7 %	9-4-0	69.2 %	26.5 %
Hiphopapotamus	58-82-3	41.6 %	6-7-0	46.2 %	4.6 %
Smash	52-88-3	37.4 %	4-9-0	30.8 %	-6.6 %
Baker’s Poop Makers	51-92-0	35.7 %	5-8-0	38.5 %	2.8 %
Mandookie	44-98-1	31.1 %	4-9-0	30.8 %	-0.3 %

The Objective Stud

Stud	SDAAs	Min	FGM	FG%	FTM	FT%	3PTM	3PT%	OReb	DReb	Ast	St	Bl	A/T
LeBron James	19.01	121	33	54.1	31	88.6	8	40.0	4	17	31	8	1	2.38
Rajon Rondo	17.92	160	25	58.1	19	61.3	4	80.0	7	14	46	11	1	3.83
Dwyane Wade	12.20	155	36	43.9	21	77.8	4	26.7	7	8	36	11	2	1.89
Stephen Jackson	12.14	121	28	53.8	15	93.8	8	53.3	3	12	14	8	1	1.27
Kevin Durant	11.86	122	26	44.1	35	94.6	5	45.5	1	21	9	6	3	0.64
Chris Bosh	11.46	119	35	58.3	21	77.8	0	NaN	9	26	12	2	5	1.50
Steve Nash	11.44	102	21	56.8	10	83.3	6	50.0	4	7	32	3	2	5.33
Andrei Kirilenko	10.53	113	22	81.5	12	85.7	1	25.0	5	13	14	11	3	2.33
Russell Westbrook	10.50	107	24	47.1	11	78.6	0	0.0	8	16	29	10	1	5.80
Chauncey Billups	8.88	116	23	39.0	14	82.4	13	50.0	0	6	21	2	0	5.25

The Objective Turd

Turd	SDAAs	Min	FGM	FG%	FTM	FT%	3PTM	3PT%	OReb	DReb	Ast	St	Bl	A/T
Antawn Jamison	-5.81	120	10	25.6	7	70.0	1	10.0	2	21	4	5	1	1.33
Mike Conley	-5.42	109	12	35.3	2	33.3	2	20.0	1	6	15	6	1	1.88
Kendrick Perkins	-4.79	101	12	50.0	7	53.8	0	NaN	3	14	3	0	4	0.38
O.J. Mayo	-3.11	135	18	33.3	6	85.7	4	25.0	1	10	10	2	2	1.43
Boris Diaw	-3.02	110	4	23.5	6	100.0	1	14.3	6	10	14	1	2	1.56
Brad Miller	-2.98	100	8	36.4	8	80.0	2	25.0	4	16	3	5	0	0.60
Eric Gordon	-2.69	105	20	43.5	15	71.4	6	33.3	1	10	5	2	0	0.50
Taj Gibson	-2.24	103	11	40.7	3	42.9	0	NaN	4	17	3	1	8	0.60
Rashard Lewis	-1.37	104	15	38.5	3	75.0	6	50.0	1	10	2	4	0	0.67
Trevor Ariza	-1.00	108	16	39.0	8	72.7	4	28.6	1	14	7	5	2	1.75

Game Capsules

Bodacious D def. Lipsticked Pigs, 9-4-0: Bo D nearly played the Pigs to a standstill, but pulled out a few close categories in the end to win by what looks at first glance to be a comfortable margin.  However, a closer look reveals that three categories were won by Bo D by only 2 – DRebs (thanks to Michael Beasley, 25); Assists (thanks to Brandon Jennings, 27 – with an A/T of 5.4); Steals (again thanks the Beasley and Jennings, 7 and 5 respectively).  The Pigs had some nice performances, including Bosh with 35 FGM and 5 blocks, and rookie Darren Collison with 150 minutes, 27 FGM, and 31 assists filling in for Chris Paul.  However, maybe the most surprising came from Carlos Delfino: 12-13 from the line, 7 3PTM, 29 rebounds, 18 assists, and 12 steals.  Who is that masked Argentine?

The Very Bad Team def. Smash, 9-4-0: The Very Bad Team had a very easy time with the Very Worse Team this week, and it didn’t hurt that his opponent’s lineup didn’t change all week.  The VBT was sitting on dead red this week, with huge numbers coming out of the United Center: Derrick Rose 35 FGM, 18 FTM, 21 assists; LaMarcus Aldridge 35 FGM, 18 FTM, 38 rebounds.  Oh, wait a second…the Bulls TRADED Aldridge for Tyrus Thomas (9 FGM, 19 rebounds, one team suspension).  Never mind…  Despite the easy week, the VBT did lose blocks as only Bogut (6) had more than 2 for the squad (and the 2 was Shane Battier of all people).  Smash had 8 players go this week, and 6 of them were basically awesome – Wade, Stojakovic, Bryant, Millsap, Howard and Kirilenko.  Now if only that bench got rotated in…

The Bombay Runners def. Mandookie, 9-4-0: The Bombay Runners took to Mandookie this week like a cockroach to crap.  Kevin Durant had yet more all-star numbers with 26 FGM, 35 FTM, and 22 rebounds.  How long before Greg Oden becomes Sam Bowie 2?  Surprisingly, the Runners were convincingly led in assists by Josh Smith – in fact, if one takes seriously the D’Antoni proclamation that Nate Robinson is a shooting guard, only one of their top five assist guys was a PG (Jose Calderon…remember that guy?  8.9 assists last year, 5.8 assists and falling this year.  Eep.)  I’m thinking that Mandookie would probably be more competitive with Danny Granger and Joe Johnson in the lineup instead of say, Bobby Simmons, but what do I know?  They managed to win all four percentage categories, so that’s something.

nutria on viagra def. Baker’s Poop Makers, 8-5-0: Baker continues his slow slide into senescence, this time losing to the Pfizer Hamsters. I’m not sure what happened to the former powerhouse, but Baker’s biggest win this year has been a 9-4 defeat of Smash.  Nutria needed every one of Amare Stoudemire’s 28 FGM to win the category by one, and I’m sure they loved the 38 rebounds, bested only by Kenyon Martin’s 47.  Boston Celtics tradebait Ray Allen (it’s a lie! says Danny Ainge, and we don’t believe him) had 26 FGM – tough to complain about.  What about Baker?  One clue to his decline might be found in his players’ Yahoo ranks – they aren’t perfect but they do carry some information.  In a 180-player league, Baker has 4 players not even in the top 200 of Y-rank over the past month.  And none in the top 40.  Wowsers, that’s not good.

Dry Heat def. Slamson, 7-5-1: The Dry Heat-Slamson Faculty Feud came just in time…for Bo D.  Two rivals beating up on each other is always worthwhile – especially when they’re #1 and #3 in the VRR.  For the Dry Heat, Russell Westbrook had a great week with 24 FGM, 24 rebounds (3rd on the team), 29 assists and 10 steals (both tops).  Monta Ellis, on the other hand, was nice with 36 FGM but his performance from the free throw line (15-24) had to be painful considering the FTM category tied.  The Slamson LeBrons got a triple-30 from their mascot (33 FGM, 31 FTM, 31 assists) and added 32 assists and 12 steals from Jason Kidd.  Marcus Camby’s 41 rebounds, 7 steals, and 7 blocks may have been the only other really notable performance, though – just goes to show what a bunch of slightly-above-average guys can do for you.

Sweet Sassy Molassy def. Hiphopapotamus, 7-6-0: The final battle of the week is always a close one, because the Matlab Auto-Crumpeter orders them that way.  (Now if it could only write the text…)  This time it was a little too much sass and not enough hip-hop, but it was a close call.  Sweet Sassy had one of those Slamsonesque solid weeks, with 11 players beating 10 FGM, and 11 players beating 10 DRebs.  Maybe the only real standouts were Dalembert’s 39 rebounds and 7 blocks and Deron Williams’ 37 assists.  The ‘potamus had a lot more variability, with big numbers from Rajon Rondo (46 assists, 25 FGM) but small numbers from players like D.J. Augustin (17 minutes, 2 assists and one turnover and nothing else).

 

Side O’ Stats, Week 14

Basketball, like any other professional sport, is about winning.  It doesn’t have to be pretty, it doesn’t have to be a blow-out, and you don’t have to lead until the very end of the game.  All that matters at the end of the day is which team had the lead at the end – which team got a mark in the win column in newspaper standings.  From there is determined playoff contenders, and the results of the playoffs themselves are determined in the same fashion.  (OK, by “any other professional sport” I really mean the NBA, MLB, NFL, and to some extent the NHL, which grants three points for a win and one for a tie, so it’s not exactly winning percentage that counts, but close enough.  I’m not worried about the ranking system in professional tennis, or whatever obscure system cricket uses, or professional hog-calling, or whatever.)

Anyway.  Wins matter.  And it would be nice if we (or our favorite sports franchises) could estimate the number of wins a player contributes, because that would help identify which players to sign, which players to trade for, which players not to overpay, which players to draft, and so on.  You know, if it could all be that simple.

Well, maybe it can be.  There’s a nifty little stat called Win Shares that is calculated by basketball-reference.com (pretty much my favorite stats site, if you haven’t figured that out).  I can’t go into the deep details of Win Shares for two reasons.  First, I’m already pressed for time (it’s 20 minutes to midnight and in the flurry of work required for my split-pea soup I’ve yet to take the garbage out, feed the cats, clean out the litterbox, etc.  World’s smallest violin, anyone?) and second, it’s complicated and the details aren’t widely available on the web.  They are apparently available in a book called “Basktetball On Paper” by Dean Oliver, but that’s on my list of books to get at this point.

What I can give right now is a quick little rundown.  Win Shares are stats that are calculated for individual players.  Win Shares are split into offensive and defensive components, each of which are calculated on the basis of boxscore numbers for the individual, and to some extent, for the team (there are also some corrections based on team pace and overall league scoring).  The interpretation of Win Shares is simple – a player with 5.1 defensive Win Shares on a season contributed about 5 wins for his team, by himself, on the defensive end of the floor.  Win Shares are also additive, which means that you can take a player’s offensive and defensive Win Shares to estimate an overall Win Share total.  Win Shares do not take into account the actual number of games won by the team a player played for.

Not so fast, you might say.  It’s well known common wisdom that a team is more than the sum of its parts.  Do you mean to suggest that you can use nothing but these boxscore stats, and tell me how valuable a player is?  And if you do claim that, how do I know I can believe you?

Well, perhaps the best test is suggested by the fact that Win Shares are supposed to be additive.  This means that we should be able to add together number of Win Shares (which can be negative) that each player contributed to his team and get a number close to the actual number of wins the team got.  Right?  Let’s check it out:

First, consider the blue data, which are Win Shares summed across all of a team’s players for all 30 teams in the 08-09 season, plotted against actual wins.  Holy smokes!  It actually works!  In fact, you can explain 97.7% of the variance of a team’s win total by simply adding up the Win Shares of the individual players.  Compare this to the Pythagorean estimate (which we’ve talked about earlier) in red.  The data are so close that I actually had to offset the Pythagorean data by 10 on the y-axis so you could see all the data points.  The Pythagorean, which takes into account the actual scoring, is only a little better, at 98.4% of the variance explained.

In the end, this probably shouldn’t be too terribly surprising.  All of the complex math behind Win Shares is designed to calculate how many points a player is responsible for creating, and how many points a player is responsible for allowing.  As long as these estimates are close to reality, they should jive pretty well actual points scored/points allowed, which is what goes into calculating the Pythagorean anyway.  So as long as the assignment of credit is reasonable, it looks like Win Shares will do a good job of telling us how valuable a player has been.

The intuitive nature of Win Shares, along with their seemingly pinpoint accuracy, makes them a stat I’m quite enamored of.  Hopefully in the coming weeks I can investigate them some more…but for now, time for bed!

 

Commissioner’s Crumpets, Week 14

Some weeks I have to pore through the NBA events of the past week to come up with an introduction.  Some weeks it’s easy.  Like this week.  I mean, I could talk about All-Star invites and the (league-falsified?) last minute vote surge that kept the benched T-Mac OUT of the game.  I could talk about Chris Paul’s injury and who might fill his slot at the All-Star game (or even whether Roy will be healthy enough to make it – there may be two open slots in the Western Conference…)

But really, all I really have to say is two words: Greg Oden.  And his enormous spike in Google searches.

For something a little tastier, it’s time for the Crumpets.  (And later, it’s time for homemade split pea soup…the hamhock stock is smelling just so tremendous right now…)

The Slamson Memorial Virtual Round Robin

Team	VRR Record	VRR %	Actual Record	Actual %	Difference
The Bombay Runners	109-31-3	77.3 %	11-2-0	84.6 %	7.3 %
Sweet Sassy Molassy	109-33-1	76.6 %	9-3-1	73.1 %	-3.5 %
Slamson	105-36-2	74.1 %	11-2-0	84.6 %	10.5 %
The Very Bad Team	79-60-4	56.6 %	8-5-0	61.5 %	4.9 %
Bodacious D	76-64-3	54.2 %	3-9-1	26.9 %	-27.3 %
Lipsticked Pigs	74-64-5	53.5 %	2-11-0	15.4 %	-38.1 %
nutria on viagra	59-78-6	43.4 %	7-6-0	53.8 %	10.4 %
Baker’s Poop Makers	59-83-1	41.6 %	5-8-0	38.5 %	-3.1 %
Dry Heat	50-88-5	36.7 %	6-7-0	46.2 %	9.5 %
Smash	44-98-1	31.1 %	2-11-0	15.4 %	-15.7 %
Mandookie	42-97-4	30.8 %	5-7-1	42.3 %	11.5 %
Hiphopapotamus	33-107-3	24.1 %	7-5-1	57.7 %	33.6 %

The Objective Stud

Stud	SDAAs	Min	FGM	FG%	FTM	FT%	3PTM	3PT%	OReb	DReb	Ast	St	Bl	A/T
Steve Nash	18.44	155	22	40.0	12	100.0	12	54.5	2	16	51	4	0	4.25
Chris Paul	16.40	128	29	59.2	15	100.0	7	50.0	1	15	33	10	0	2.75
Gerald Wallace	15.95	178	33	57.9	38	76.0	5	62.5	8	32	9	5	5	0.90
Chauncey Billups	15.60	152	25	46.3	35	92.1	7	46.7	0	21	33	8	0	1.74
Marc Gasol	14.26	156	26	57.8	25	80.6	0	0.0	11	31	13	10	7	1.44
LeBron James	13.92	146	28	40.6	34	70.8	8	36.4	1	26	39	5	5	3.00
Kevin Durant	13.38	118	35	59.3	28	87.5	5	50.0	7	19	11	5	4	1.10
Stephen Curry	13.37	163	29	48.3	12	92.3	8	42.1	3	16	28	10	0	2.33
Zach Randolph	12.75	166	38	51.4	13	81.2	2	50.0	18	39	12	3	2	1.50
Tim Duncan	12.18	139	24	38.1	23	95.8	0	NaN	20	31	16	0	6	5.33

The Objective Turd

Turd	SDAAs	Min	FGM	FG%	FTM	FT%	3PTM	3PT%	OReb	DReb	Ast	St	Bl	A/T
Eric Gordon	-6.79	106	15	34.1	8	53.3	3	18.8	3	9	10	2	2	1.67
Vince Carter	-6.40	102	11	25.6	5	71.4	2	18.2	5	8	6	4	1	2.00
Matt Barnes	-5.37	100	6	35.3	3	60.0	0	0.0	4	13	6	4	0	1.00
Richard Jefferson	-4.54	130	15	37.5	5	50.0	1	11.1	4	12	8	5	1	2.00
Tayshaun Prince	-4.01	101	15	42.9	3	75.0	1	50.0	4	12	3	1	0	1.50
Kendrick Perkins	-3.38	101	10	47.6	9	50.0	0	NaN	6	34	0	0	3	0.00
Rudy Gay	-2.91	166	18	34.0	12	63.2	2	18.2	6	13	7	4	3	0.88
Jason Richardson	-2.85	148	20	39.2	1	25.0	4	22.2	4	21	8	3	2	1.33
Martell Webster	-2.36	109	10	30.3	8	100.0	6	30.0	3	9	3	3	2	3.00
Al Horford	-2.35	127	13	37.1	2	66.7	0	NaN	7	24	6	0	2	1.50

Game Capsules

The Bombay Runners def. Smash, 11-2-0: In Soviet Russia, you do not smash cockroach, in Soviet Russia cockroach smash YOU! (I have now officially fulfilled my lifetime quota of Yakov Smirnoff jokes…not that that’s going to stop me.)  Tim Duncan had a nice quad-20 for the Runners with 24 FGM, 23 FTM, 20 ORebs and 31 DRebs, but his 38.1 FG% knocked him down to #10 on the Stud list.  Dwight Howard SHOULD have had a better week – he had a triple-30 (31 FGM, 31 FTM, 48 DRebs) and added in 11 ORebs and 18 blocks – but his horrible 54.4% shooting from the line sunk him.  I guess with 59 rebounds and 18 blocks, you take the loss in FT% and don’t look back, though.

Slamson def. Lipsticked Pigs, 11-2-0: I think it’s no secret that kitties like a nice pulled pork sandwich, and Slamson savored his way all the way into a tie for first in the league.  (Speaking of tasty, tasty pork, you really should smell that stock!)  LeBron “Ho-Hum” James had another quad-20 for Slamson (28 FGM, 34 FTM, 26 DRebs, and an impressive 39 assists) while Marc Gasol had 26 FGM, 25 FTM, and 42 rebounds in an effort to eclipse his older brother (it might be working, shockingly!)  The Savory Soup Bases got their last 29 FGM and 33 assists out of Chris Paul for a few months, or the season…they say it’s a torn meniscus and he’s out 1-2 months, but with the Hornets likely to fade out of the playoff picture (they’re in the 8 slot as of now) I wonder if they’ll activate the franchise for a few meaningless weeks in April, or just let him heal.

Sweet Sassy Molassy def. Bodacious D, 9-3-1: Sweet Sassy is on a surge, and did his own part to help Slamson move into a tie for first by beating the snot out of Bo D – and solidified third place as a result.  Sweet Sassy was led in assists by…Rodney Stuckey?  Yup, Stuckey’s 20 assists was tops, partially due to Deron being a bit injured and Baron only getting 98 minutes on the week.  Playing against the Sacramento Kings helped Gerald Wallace score a triple-30, with 33 FGM, 38 FTM, and 32 DRebs.  For Bo D, Corey Maggette somewhat recovered from his Monday slump (3 of 22 shooting) to avoid the Turd list with 25 FGM at 38.5% and 34 FTM.  Pau Gasol had 28 FGM, 21 FTM, and 44 rebounds…pretty much mirroring his brother (but with fewer blocks).

The Very Bad Team def. Baker’s Poop Makers, 8-5-0: The VBT outplayed the hapless Poopmakers this week, thanks in large part to All-Star snub Chauncey Billups.  Billups had a quad-20 with 25 FGM, 35 FTM, 21 DRebs (to 0 ORebs!) and 33 assists.  A performance like that might give Billups a shot at replacing Paul in the All-Star game.  Baker’s best player this week was a bit of a surprise – Aawon Bwooks!  (Sorry, had a bit of peanut butter in my mouth there…)  Brooks led the team in 5 categories – 148 minutes, 33 FGM, 14 3PTM, 18 assists, and 4 steals (tied with 2 others).  Taj Gibson (pushing 6′9″ and the #26 pick) has come out of nowhere this season, and led the team with 6 blocks.

Hiphopapotamus def. Mandookie, 7-5-1: Rajon Rondo was the DJ for Hiphopapotamus this week, leading the team with 40 assists, 24 total rebounds, and seven steals.  That no other player beat 24 boards (Dirk? Nene? Kaman? Biedrins? Hawes?) is kinda sad, but at least the last three had limited minutes.  Mandookie’s David Lee has spent most of the season trying to earn his money in the upcoming offseason – this week he had 31 FGM and 38 rebounds in that quest.

nutria on viagra def. Dry Heat, 7-6-0: Fortunately the ABRs do not have Greg Oden on their team – he and Viagra would be a bad mix.  But the nutria won both rebounding categories even without 2007’s number one pick, and Antawn Jamison’s 14 ORebs can’t be discounted, as the margin was only one.  Steve Nash had 51 assists in yet another big week – his 12 threes was similarly outstanding.  Zach Randolph, the man with one of the least explicable nicknames in the game (Z-Bo?  Really?) had another huge week for the Dry Heat with 38 FGM and 57 rebounds, while Paul Pierce went 32 of 33 from the free throw line.

 

Commissioner’s Crumpets, Week 13

The big news in the NBA this week…well, not a lot.  There’s Javaris “The Other Guy In The Arenas Incident” Crittenton getting a year of probation in D.C., there’s the huge pending trade between New Orleans and Chicago (Devin Brown for Aaron Gray, and if you don’t know who’s going which way…don’t worry), and there’s the blockbuster announcement from Glen Davis that he no longer wants the be called “Big Baby”…more on that in an upcoming poll…but for now, it’s Crumpet Time.

The Slamson Memorial Virtual Round Robin

Team	VRR Record	VRR %	Actual Record	Actual %	Difference
The Very Bad Team	109-34-0	76.2 %	10-3-0	76.9 %	0.7 %
Bodacious D	102-38-3	72.4 %	12-1-0	92.3 %	19.9 %
Slamson	97-45-1	68.2 %	10-3-0	76.9 %	8.7 %
Sweet Sassy Molassy	80-59-4	57.3 %	8-3-2	69.2 %	11.9 %
The Bombay Runners	77-64-2	54.5 %	10-3-0	76.9 %	22.4 %
Hiphopapotamus	74-66-3	52.8 %	10-3-0	76.9 %	24.1 %
nutria on viagra	72-71-0	50.3 %	3-10-0	23.1 %	-27.2 %
Dry Heat	69-73-1	48.6 %	3-10-0	23.1 %	-25.5 %
Lipsticked Pigs	55-85-3	39.5 %	3-8-2	30.8 %	-8.7 %
Baker’s Poop Makers	46-96-1	32.5 %	3-10-0	23.1 %	-9.4 %
Smash	40-101-2	28.7 %	3-10-0	23.1 %	-5.6 %
Mandookie	27-116-0	18.9 %	1-12-0	7.7 %	-11.2 %

The Objective Stud

Stud	SDAAs	Min	FGM	FG%	FTM	FT%	3PTM	3PT%	OReb	DReb	Ast	St	Bl	A/T
Kevin Durant	16.29	165	36	42.9	47	92.2	5	50.0	7	26	11	5	2	1.22
Stephen Curry	15.44	180	36	44.4	11	91.7	15	46.9	2	18	21	6	1	1.91
Corey Maggette	14.52	152	40	58.0	39	92.9	2	33.3	7	18	20	3	0	1.54
Chauncey Billups	13.85	111	22	52.4	20	90.9	13	59.1	1	15	23	0	0	3.29
Dwight Howard	13.07	155	26	57.8	33	71.7	0	0.0	17	41	4	9	12	0.22
Andrew Bogut	12.91	138	33	67.3	15	75.0	0	NaN	7	34	10	1	16	1.43
Chris Paul	12.81	161	31	50.0	22	91.7	4	30.8	1	16	41	3	0	2.93
LeBron James	12.12	124	30	48.4	35	76.1	7	26.9	0	23	32	5	5	2.46
Chris Bosh	11.52	162	39	54.2	22	88.0	0	0.0	10	32	12	5	1	1.33
Brook Lopez	11.05	135	33	61.1	18	75.0	0	NaN	16	16	10	4	8	1.67

The Objective Turd

Turd	SDAAs	Min	FGM	FG%	FTM	FT%	3PTM	3PT%	OReb	DReb	Ast	St	Bl	A/T
Kevin Martin	-5.04	128	13	31.7	14	63.6	2	14.3	3	13	13	4	0	1.44
Ray Allen	-4.28	119	15	38.5	5	71.4	3	23.1	2	9	9	1	1	1.50
Trevor Ariza	-4.22	112	12	35.3	5	71.4	3	20.0	2	13	6	5	0	1.20
Jason Thompson	-3.41	101	13	34.2	10	71.4	0	0.0	9	17	1	0	4	0.33
Charlie Villanueva	-2.48	107	22	39.3	7	87.5	4	21.1	5	14	1	2	3	0.33
Chris Douglas-Roberts	-2.12	105	10	47.6	4	100.0	0	0.0	3	6	6	6	0	1.50
Jason Richardson	-1.24	109	24	45.3	4	80.0	2	16.7	5	20	4	2	1	4.00
Danilo Gallinari	-1.17	100	14	40.0	8	88.9	6	35.3	2	10	4	2	3	0.57
Wilson Chandler	-1.13	104	23	53.5	7	70.0	1	20.0	2	10	6	1	4	1.50
Rasual Butler	-0.61	141	11	31.4	16	100.0	4	22.2	2	5	7	2	3	1.75

Game Capsules

Bodacious D def. Mandookie, 12-1-0: Bo D had an easy time this week with Bye-dookie.  Corey Maggette led the team with 40 FGM, 39 FTM, and was only one short on the assist list with 20, while Stephen Curry parlayed his 180-minute week (only one of those games without Monta Ellis) into 36 FGM, 15 threes, and a team-leading 21 assists.  For the Dook, Joakim Noah continued his shockingly solid free throw shooting with an 11 of 12 outing to help win FT% (team = .909), the only category the Dook managed this week.  Meanwhile Danny Granger turned the ball over 14 times, and the Dook lost A/T by 0.01.

The Very Bad Team def. Dry Heat, 10-3-0: All that rain this week took the Dry out of the Heat, and allowed the VBT to slide in for an easy win (category-wise).  Tyrus Thomas had 12 blocks and 26 rebounds (1 offensive?!?) in a nice week, but couldn’t hold a candle to Andrew Bogut’s 33 FGM, 41 rebounds, and 16 blocks.  The Dry Heat lost assists by one, and might have benefitted from having a backup point guard not named Jrue Holiday (who managed 11 minutes and 1 assist in the lineup this week).  Sure, he’s young.  Sure, he’s cheaply keepable.  But unless he turns it around soon, he’s not going to be worth keeping anyway – he’s on track to be available in the last round of the draft at best as it stands now.

Slamson def. nutria on viagra, 10-3-0: The ABRs got the minutes, but Slamson swept the back categories to pull out a close victory that looks like a landslide in the summary.  LeBron’s quad-20 certainly helped, with 30 FGM, 35 FTM, 23 DRebs, and 32 assists.  Marcus Camby? 64 rebounds.  For the ABRs, Matt Barnes, Leandro Barbosa, and Ray Allen all went 3 of 13 from three – horrible numbers for any of the three – and only four more made threes would have turned the category (and would have turned FG% and 3PT% as well!).  Kenyon Martin went 2 of 8 from the free throw line – just 50% on the week would have won FTM.  Those are two woulda-coulda-shouldas that make the difference between a 3-10 loss and a 7-6 win.

The Bombay Runners def. Baker’s Poop Makers, 10-3-0:  Apparently roaches cannot be destroyed by either nuclear fallout or poop.  Not only did the Runners take care of Poopy this week, but they also overcame Kevin Martin playing like a Turd – the #1 Turd of the week, in fact.  One FG every ten minutes, 32% shooting, 64% from the line, and 14% from deep…makes a Kings fan cry.  It was a rough road trip and from Sacramento to Bombay, just about everybody is going to hope that Martin gets his legs under him sooner rather than later.  Baker lost two games of Eric Gordon to an ouchie toe (now three counting tonight’s game), which is probably the last thing he needed, seeing as Gordon is top-3 in five counting categories (average stats) for the team.

Hiphopapotamus def. Smash, 10-3-0:  It’s a 10-3 party, with our final “stop transmitting, good buddy” coming courtesy of that grandma from The Wedding Singer.  It’s a good thing that Dirk and Carmelo are really, really awesome (35 FGM, 28 FTM and 25 rebounds for Germany, 23 FGM, 32 FTM, and 7 steals for Syracuse) because T.J. Ford has apparently played his way out of favor in Indiana.  It appears that T.J.’s welcome lasts about two years (Milwaukee: two seasons.  Toronto: two seasons.  Indiana: 1.5 seasons and counting.) Considering that Indiana stuck with Jamaal “Guns Are For Night Clubs, Not Clubhouses” Tinsley for 7 years, that’s got to be saying something.  It didn’t help that Smash didn’t change his lineup last week (a mistake that has already been rectified) and left 41 FGM-worth of Dwyane Wade and other friends like Andrei Kirilenko on the bench.

Sweet Sassy Molassy def. Lipsticked Pigs, 8-3-2: And, in the closest matchup of the week, Sweet Sassy sent the Pigs off to their worst loss of the season.  My absolute favorite thing about this matchup is that the two teams dead tied in both assists AND A/T (meaning obviously that they had the same number of turnovers, too).  Never seen anything like that.  Al Jefferson (35 FGM, 45 rebounds) was a star for SSM, while the Pigs were led by Chris Bosh (39 FGM, 42 rebounds).

 

Side O’ Stats, Week 12

While watching the Kings game on this lovely MLK day afternoon (between two work stints), I came to the conclusion that the Kings seem to be giving up a ridiculous number of transition points this season.  Now, I don’t know of any site that carries any statistics on number of transition points allowed, so it’s not even clear that I can even ask a question about transition defense.  However, I have noticed that the majority of transition points scored against the Kings seem to come off of three-point misses.  There, I thought, is an idea.

So what I’m going to do this week is look into the question of whether three-point misses result in more transition points.  It certainly seems that way to the eye, but is it true?  I can’t ask the question directly, because I don’t know the number of transition points, but I can try to sniff around the answer and see what comes out.

I’ve got a nice graph for you in a moment, but let me explain what I did.  Over the 10 NBA seasons between ‘99-’00 and ‘08-’09, I gathered each team’s three-point misses, estimated possessions, and defensive rating (points allowed per 100 possessions).  I then plotted percentage of possessions that ended in a missed three-point shot against defensive rating.  The results?

Now that looks like a pretty noisy cloud, and the number of possessions that end in a missed three only explain about 2% of the variance in DRTG.  Still, the relationship is significant, with a p-value of 0.013.

Despite the significant p-value, it’s important to make sure that this relationship is independent of missed shots in and of themselves – do missed two-point shots correlate with DRTG as well?  The answer, it turns out, is yes.

So…the answer is yes (this relationship is also significant, p = 0.030), but the correlation goes in the opposite direction!  This tells us that teams that miss a lot of two-point shots (and thus, miss a lot of shots overall) actually play BETTER defense.  This probably simply reflects a tendency for good offensive teams to be poor defensive teams, and vice versa.  What it does tell us is that the relationship between missed THREES and poor defense is not a simple factor of just missing shots in general – there’s something else going on causing teams that miss a lot of threes to allow their opponents to score more points – and that may just be transition points.  (It’s certainly the most logical explanation that comes to mind.)

So what else can we learn about the relationship between missed threes and defensive rating?  Well, if we look at the equation for the line, we estimate that each missed three per 100 possessions results in 0.21 points per 100 possessions.  We can just knock out that whole “per 100 possessions” thing and see that each missed three results in 0.21 points.  Well, an average possession results in something close to 1.00 points (it varies from year to year, but that’s a reasonable-enough value that I don’t want to pull out the data) and what I would call an “easy” transition attempt probably results in about 2 points per possession (for the purposes of this back-of-the-envelope discussion, missed layups probably cancel out three-point plays).  A little simple algebra (and the conveniently simple numbers chosen help here) suggests that something like 21% of missed threes result in a fast-break opportunity.  One in five!  One in five times your team misses a three, you get the ball shoved down your throat.  Kind of makes you care about good looks from three, don’t it?

Furthermore, traditionally we think of 33.3% as the break-even percentage from three.  The average possession results in about 1 point, so if you shoot 33.3% from deep you’ll score 3 points on 3 shots – no harm, no foul.  Well, except for that little bugaboo about -0.21 points per miss – this is merely the “offensive” break-even point.  Again, a little algebra suggests something fairly remarkable – to break even at shooting threes on both offense and defense, a player doesn’t need to shoot 33.3%, he needs to shoot 37.7%!  This season, exactly 70 players have attempted at least 20 threes and are above this new “net” break-even point.  In comparison, 143 have shot worse than the break-even point, and the league three-point percentage is 35.2%!  Now, the league instituted the three-point shot because they wanted to increase scoring – and it worked.  I’ll bet my last dollar that they figured that the additional scoring would be a result of the three-point shot itself.  To be sure, the three-point shot is increasing offensive scoring (35.2% is greater than the offensive break-even point of 33.3%), but it appears that the gain for the three-point defending team (off of easy transition points, presumably) is actually greater!  Who’da thunk it!?!

With a conclusion like that I could (and should) finish there, but there’s at least one more interesting and related question: what about turnovers?  Shouldn’t a high percentage of turnovers result in more (putative) transition scoring as well?  I’ll spare you the details, but this result may be even more remarkable – it turns out that there is no statistical relationship between the two (and if the small relationship that is present were significant, it would in fact be a negative relationship – the more turnovers you have, the better a defensive team you are).  Again, this is probably related to the tendency for good offensive teams to be poor defensive teams, but it’s somewhat surprising that things like breakaway steals get completely swallowed up.  But, if we think about the fact that many turnovers result in a dead ball – offensive fouls and any out-of-bounds turnover – this may not be quite so surprising.

And that’s about that for transition points scored off three-point misses.  I just love how a simple question can sometimes lead to such a counterintuitive result!

 

Commissioner’s Crumpets, Week 12

This week, it’s Imaginary Press Release time!  We’ll start with this one:

June 26, 2002 – New York City

Today the Chinese government traded the rights of 21-year-old center Yao Ming of the Shanghai Sharks to the NBA for the right to ignore all future international global warming treaties and a player to be named later.  Yao, who is 7′6″, was selected with the first overall pick in the NBA draft by the Houston Rockets.  Wang Fei, coach of the Chinese National Basketball Team, explained the trade: “For glory and pride of indomitable Chinese National Team, greatest world basketball player Yao Ming will practice 82 games year during Chinese National offseason with short inferior American basketball team to ensure Chinese victory in all FIBA and Olympic championships of coming century.”  Yao has an unprecedented combination of size and skill, and pundits expect that the player to be named later in the trade will likely be a current star.  Given the Chinese need for young athletic point guards who can both score and pass, players like Baron Davis and Stephon Marbury are probably on the top of the list.

And follow up with this one:

January 18 2010 – Shanxi, China

Today the Chinese government announced that the player to be named later in 2002’s trade of Yao Ming to the NBA is 32-year-old NBA washout Stephon Marbury.  Marbury, who played for 5 NBA teams in his 13-year career, was best known in recent years for his large, untradeable contract, his locker-room issues, getting an advertising tattoo on the side of his head, and crying during a nervous breakdown on a 24/7 webcam.  Marbury will play with the Shanxi Zhongyu Brave Dragons of the Chinese Basketball Association.  Shanxi boss Wang Xingjiang told reporters, “We knew infinite carbon emissions for benefit of glorious Chinese people would cost more trouble than losing only Yao Ming.  But why me?  Why me?” and proceeded to sob uncontrollably.  After Marbury’s departure for Shanxi, U.S. Customs officials quietly placed Marbury’s name on the U.S. “No-Fly” List, but cautioned that Marbury might still reenter the country by sneaking across the Mexican border or stowing aboard a container ship.

Hmm.  OK, maybe it didn’t quite go down that way, but I still think we win.  And now, the Crumpets!

The Slamson Memorial Virtual Round Robin

Team	VRR Record	VRR %	Actual Record	Actual %	Difference
Sweet Sassy Molassy	109-33-1	76.6 %	8-5-0	61.5 %	-15.1 %
Slamson	103-39-1	72.4 %	5-8-0	38.5 %	-33.9 %
The Bombay Runners	93-49-1	65.4 %	9-4-0	69.2 %	3.8 %
Dry Heat	92-50-1	64.7 %	4-9-0	30.8 %	-33.9 %
nutria on viagra	86-53-4	61.5 %	9-4-0	69.2 %	7.7 %
Lipsticked Pigs	70-72-1	49.3 %	10-3-0	76.9 %	27.6 %
Baker’s Poop Makers	64-76-3	45.8 %	7-6-0	53.8 %	8.0 %
Bodacious D	60-81-2	42.7 %	10-3-0	76.9 %	34.2 %
The Very Bad Team	55-88-0	38.5 %	4-9-0	30.8 %	-7.7 %
Hiphopapotamus	50-90-3	36.0 %	6-7-0	46.2 %	10.2 %
Mandookie	38-105-0	26.6 %	3-10-0	23.1 %	-3.5 %
Smash	28-112-3	20.6 %	3-10-0	23.1 %	2.5 %

The Objective Stud

Stud	SDAAs	Min	FGM	FG%	FTM	FT%	3PTM	3PT%	OReb	DReb	Ast	St	Bl	A/T
Chris Paul	21.57	165	27	47.4	15	88.2	5	41.7	1	19	54	12	1	9.00
Kevin Durant	17.24	155	42	51.9	38	86.4	9	69.2	3	30	9	2	3	0.45
LeBron James	17.09	121	36	57.1	25	75.8	8	50.0	4	17	21	9	5	1.75
Stephen Jackson	13.73	107	32	60.4	15	75.0	9	52.9	3	17	14	10	0	2.33
Corey Maggette	12.19	120	24	61.5	40	95.2	1	100.0	6	16	13	2	0	1.44
Antawn Jamison	12.14	187	38	45.8	20	64.5	8	36.4	13	43	10	2	1	5.00
Deron Williams	10.90	129	25	52.1	15	88.2	6	46.2	5	6	34	5	0	2.12
Danny Granger	10.36	146	33	45.8	17	89.5	13	44.8	2	15	9	3	2	0.90
Gerald Wallace	9.03	122	22	56.4	13	76.5	2	50.0	7	21	7	7	7	1.17
Carmelo Anthony	9.00	108	30	55.6	23	95.8	5	41.7	5	12	9	6	1	0.60

The Objective Turd

Turd	SDAAs	Min	FGM	FG%	FTM	FT%	3PTM	3PT%	OReb	DReb	Ast	St	Bl	A/T
Jonny Flynn	-7.56	100	9	22.0	5	71.4	2	22.2	0	6	11	3	0	1.83
Spencer Hawes	-3.52	102	14	40.0	3	50.0	3	33.3	5	11	3	1	5	0.38
Mo Williams	-3.40	107	13	35.1	8	100.0	2	15.4	1	5	17	3	0	2.12
Kenyon Martin	-3.33	105	11	37.9	5	41.7	0	NaN	7	27	4	1	2	2.00
Richard Jefferson	-3.21	119	13	39.4	6	66.7	4	33.3	1	12	5	3	1	0.71
Ray Allen	-2.37	104	14	51.9	7	100.0	4	36.4	1	7	4	0	0	0.67
Shawn Marion	-1.86	100	15	55.6	1	25.0	0	NaN	6	9	5	2	2	5.00
Hedo Turkoglu	-1.64	108	8	28.6	10	66.7	5	29.4	1	18	14	4	1	2.33
Channing Frye	-0.56	100	5	20.0	7	87.5	3	25.0	5	15	6	3	8	2.00
Trevor Ariza	-0.43	131	19	38.8	11	64.7	4	26.7	4	15	14	5	1	1.27

Game Capsules

Bodacious D def. Smash, 10-3-0: A lackluster Bo D took a Smash! brand energy drink and got a big boost in the VRR this week.  Trade-target Corey Maggette went 40 of 42 from the charity stripe for Bo D this week to likely secure this week’s FTA title.  Formerly hot rookies Brandon Jennings and Omri Casspi combined for 20 of 65 (31%) from the floor this week.  For Smash, Andrei Kirilenko had 1 rebound and no blocks in his active game (and had 10.5 pts, 2.75 rebs, 0 blocks per game in 4 games this week).  Not great for a guy making…$34M+ over this season and next.  Smash’s top FTA came from Dwight Howard (35).

Lipsticked Pigs def. Mandookie, 10-3-0: The Pigs clobbered the Dooks this week, with Chris Paul’s 54 assists singlehandedly destroying the TEAM total of 39 from his opponent.  Also of note: Steals – Chris Paul 12, Dooks 9.  Top FTA = Chris Bosh (32).  The Dooks still aren’t changing their lineup, and that hurts when it includes players like the out-for-the-season Joel Przybilla and the DNP-CD machine Bobby Simmons.  On the bright side, Joakim Noah and his trademarked Ugliest Free Throw Stroke In NBA History went 22/23 from the line this week to rack up the most FTAs for Dook.

nutria on viagra def. The Very Bad Team, 9-4-0: Nutria took care of the VBT this week, assisted no doubt by Antawn Jamison’s 38 FGM and 56 total rebounds.  Carlos Boozer added 49 rebounds and 10 steals of his own.  Jamison also led the ABRs in FTA (31).  Andrew Bogut had 51 rebounds and 8 blocks for the VBT in a team-leading 125 minutes.  Vince Carter took 7 shots on the week (6 from three) and only hit one.  Chauncey Billups had 35 FTAs this week, so halfway through the teams, Maggette still leads the way.

The Bombay Runners def. Dry Heat, 9-4-0:  This week the DST&CC demonstrated the rare circumstances under which a wet heat beats a Dry Heat.  The Monsoon Marathoners were led by Kevin Durant’s triple-30 (42 FGM, 38 FTM, and 30 DRebs).  Durant also eclipsed Maggette in FTA with 44, though he made fewer.  The Dry Heat had a balanced attack with 6 players with 20+ FGM, 6 players with 14+ FTM, and 7 players with 10+ assists.  Zach Randolph led the way with 23 FTA.

Sweet Sassy Molassy def. Slamson, 8-5-0: Sweet Sassy is on a charge, with 8 consecutive winning weeks, the victim Slamson this time around.  If you thought that Dry Heat’s scoring attack was balanced (and I sure did), then you’ll like seeing that 8 players for SSM had 20+ FGM.  The leader in FTA was Emeka Okafor with 26.  Kendrick Perkins had 11 blocks on the week for Slamson, and was no slouch with 27 rebounds in 3 games.  Still, Slamson’s got better board machines than that, and Troy Murphy (53), Marcus Camby (41), Ben Wallace (36) and Marc Gasol (34) all outshined Perkins there.  LeBron’s 33 FTA was tops for the LeBrons…I mean, Slamson.

Baker’s Poop Makers def. Hiphopapotamus, 7-6-0: Go Deuterostomes! was the chant as the Poopmakers eked out a win against the Hip-hop Potty Mouths.  Eric Gordon led the Poopmakers with 27 FGM, and 5 players had at least 8 threes as Baker easily doubled up the Hippos in 3PTA.  Kevin Love led Poopy in FTA with 19.  Andre Miller only played in two games this week, but his 14 FGM and 15 assists were inexplicably benched on both days for Larry Hughes, who didn’t end up playing in either contest as he continues to battle a case of terminal egotism.  Dirk Nowitzki led the way with 26 FTA, leaving Kevin Durant at the top of the list.

 

Side O’ Stats, Week 11

PER2ENG

This week, I’d like to take a look at the PER (Player Efficiency Rating) statistic, and what it all means in English, rather than in its native mathese.  PER was originally devised by John Hollinger, who has probably invented about 25% of the basketball statistics out there.  The idea is to be able to take all of the information available in a box score (and essentially, nothing else) and to create one number that encapsulates the value of each player in a per-minute fashion.  (Whether the name “PER” is intended to mimic the idea of a rate stat, i.e. PER-minute, I have no idea.)

PER is one of the stats calculated and presented at the fabulous site www.basketball-reference.com, and the mathematical formula for it is presented here.  For your own benefit, I won’t rewrite it – rather I’m going to slog through it myself and translate it to conceptual English so that you have an idea of what it’s really measuring.  Lest you think this is a futile exercise, keep in mind that PER is possibly the most popular “advanced” statistic.  Anybody who cares to look at/think about advanced basketball stats in any way needs to have an idea of what PER is – it’s the starting point.

The Strategy

What Hollinger wanted to do was to essentially provide a “point value” for every possible stat – determining approximately how many actual scoreboard points each event was worth.  Obviously some stats are worth positive points, like…field goals.  Other stats cost a team points, like turnovers.  So the first step is to go through the whole box score, and estimate a player’s complete contribution to the scoreboard.  After that, Hollinger does some normalization that I’ll also talk about.  But first, there are a few factors that will be used from time to time, and which we should get out of the way first.

The “Factors”

If you note, I said that PER could be calculated from essentially the box score and nothing else.  The “factors” are where the caveat comes in, because they all rely on league average data – something which cannot be obtained from the box score stats of a single player.

The most important factor is known as “Value of Possession”, or VOP.  This factor is absolutely crucial, because it is an estimate of how many points an average team is going to score when it has the ball – it’s a very important reference.  It has the obvious definition of League Points / League Possessions.  Now, estimating possessions is inexact from the box score, but we can get pretty close by summing up the number of times a possession ends.  For the purposes of PER, a possession ends in any of three ways: 1) A turnover; 2) A field goal that does not result in an offensive rebound; 3) A foul that results in free throw attempts but no field goal attempt.  Calculating 1) is obvious.  We have calculated 3) many times before by multiplying 0.44 by FTAs.  Finally, 2) is easy as well, by just subtracting offensive rebounds from field goal attempts.  (Note that this definition considers an offensive rebound to be a continuation of the same possession.  My own personal bias is to consider an offensive rebound as resultling in a new possession, but in the end it shouldn’t make a big difference in the calculation.)

The second factor is League Defensive Rebounding Percentage (DRB%), which is nothing more than the percentage of rebounds grabbed by the defense.

The third factor is curiously called “factor”, and damn if it isn’t goofy.  I’m actually going to skip it for now and hand-wave it later.

The Positives

As I said above, PER essentially adds together positive contributions in terms of scoreboard points and the subtracts out negative contributions.  Here are the positive contributions:

1) For each assist, credit 2/3 of a point

2) For each field goal, credit 2 points if unassisted and credit 1 1/3 points if assisted** (the other 2/3 goes to the assisting player).  The number of assisted field goals is estimated by simply looking at team assists divided by team FGs.

3) For each three-point shot made, credit 1 additional point (do not try to account for assists here)

4) For each free throw made, credit 1 point**.

For a team, these first four elements should add to the total number of points scored…but see the note** below.

5) For each steal, credit the VOP – this is the number of points your team is expected to score now that you have possession.

6) For each block, credit the VOP times the DRB% – this is estimating the probablility that your team will recover the block.  (Note that I think this is not a very good estimate – said loose ball is simply not equivalent to a missed shot.)

7) For each offensive rebound, credit the VOP times the DRB% – this gives credit for the new possession, but tries to account for the percentage of the time your team would have gotten the rebound anyway.

8) For each defensive rebound, credit the VOP times (1 – DRB%) – the logic here is identical to the logic in 7).

These second four elements estimate the number of points your team should score on possessions you specifically created.

**My description of these two elements is not precise. Hollinger essentially takes a small amount of credit away from each free throw made (about 1/12 of a point) and adds a small amount of credit to each assisted field goal (about 1/16 of a point, this is “factor”).  Why he does this, I can’t even begin to guess.  And I’m not buying his book to figure it out.  It’s small, at any rate.  To the best of my ability to tell, these two adjustments do not mathematically offset, so the total number of “scoring” credits may not be equal to the total number of points scored.

The Negatives

Now, a quick look at the negative components, or how you can “cost” your team points.

1) For each turnover, subtract the VOP – this estimates the number of points the opponent gets on possessions after your turnovers.

2) For each missed field goal, subtract the VOP times the DRB% – this estimates the number of points the opponent gets on possessions resulting from your missed shots.

3) For each missed free throw that might be rebounded by the opponent, subtract the VOP times the DRB%*** – this estimates the number of points the opponent gets on possessions resulting from your missed free throws.

4)Determine the probability that a foul results in free throw attempts.  For each foul that results in free throw attempts, subtract the number of extra points the opponent is expected to score, relative to a possession on which they are not fouled†††.

Altogether, these maybe kind of sum up to the number of points the opponent is expected to score because you screwed up.  Where screwing up includes missing shots (but oddly not MAKING shots, which always results in a new possession for the opponent).

***The actual formula for this looks like a dog chewed up the right numbers, took some extra ones, and vomited them all out on the stock page of a newspaper.  So I’m not exactly sure that this is what Hollinger is trying to estimate here.  It would be far easier to do it, correctly, with far fewer numbers and operations.

†††The actual formula for this is completely inscrutable.  It may have been written by a blind scribe who spoke only Aramaic mistakenly trying to translate acoustical ceiling tiles as ancient braille.  But it’s really the only thing Hollinger could have been going for.  I think.

Crap, What Now?

OK, well we’ve got a heck of a number so far – it’s essentially the number of points you’re directly responsible for scoring, plus the number of points you’re indirectly responsible for scoring, minus the number of points you’re indirectly responsible for allowing.  So if you indirectly give the opponents more points than you indirectly give to your own team, your total points will be modified downwards.  Otherwise, your total points will be modified upwards.

Now, divide your “point total” by your total minutes.  That’s reasonable.

Then, since some teams play “fast” and some teams play “slow”, adjust this value for league pace (by simply multiplying by league pace and dividing by your team pace).

Then, to completely baffle the uninitiated, normalize this number so that a player with league-average PER will get a value of 15.  Because 15 was John Hollinger’s number on his high school basketball teamˆˆ.  This normalization is most useful because it allows PER to be compared across seasons.

ˆˆNote: this is a completely unfounded accusation, but it makes as much sense as any other explanation.  Normalize to 15…because…it’s 15!  Duh!

What PER Does

Well, if you’ve been paying attention…you get a gold star!  And you also probably have a reasonable idea of what PER does – it is a single number that tries to summarize a player’s scoring/possession value from only box score numbers.  It allows comparison across seasons, and is not dependent on team pace.  These are all pretty good things.  But PER is far from perfect.

What PER Does Not Do

There are quite a few things that PER does not do, or does not do well.

1) It does not actually measure defensive value.  It uses steals and blocks as a proxy for defensive value, but even then it gives a basically arbitrary (and probably low) weight to the defensive proxies.  Good defensive players should have PERs that underestimate their actual value.  Poor defensive players (for Kings fans, Kevin Martin) should have PERs that overestimate their actual value.

2) It does not measure “intangibles” that are not in the box score.  Notice that “intangibles” may include vague terms like “hustle” and “leadership” (ugh) but may also include uncounted “tangibles” like saving a ball from going out of bounds, or knocking a ball off of an opponent.  Such plays could in theory be counted in a box score but are not, and thus can’t weigh in to PER even if they do result in possessions.

3) It does not account for “garbage-time” play – players at the back of the bench who only get in at garbage time typically face back-of-bench opponents as well, and this can make really bad players look pretty darn average.

4) It has been argued (by people who pay more attention to Hollinger’s scribble-vomit than, say, Hollinger) that PER rewards inefficient volume shooting.  This contention is probably correct, but I’m not going to try to evaluate it.

5) PER does not save drowning puppies.  Boo, PER!

Should I Use PER?

No.  (Normally I would be a bit more charitable, but it’s now 1 in the morning and I’m getting downright surly.)  But other people will.  So you should know what it does, and what it doesn’t do.  Or don’t.  I don’t care.  (See? Surly.)  OK, going to bed.  Aaaaah.

 

Commissioner’s Crumpets, Week 11

Perhaps in order to deflect the sports nation’s attention from the fact that Javaris Crittenton has yet to be formally suspended in the Gilbert Arenas Showdown Incident, Mark McGwire has just admitted he used steroids. (For Marty’s benefit, “Mark McGwire” was a really famous baseball player who hit a lot of home runs and destroyed Roger Maris’ almost 40-year-old season home run record before non-admitted steroid user Barry Bonds turned around and broke the record again.  Mark McGwire was initially “outed” when a quasi-legal enhancement substance was found in his locker by a reporter, but maintained until today that he had not used steroids.) Estimated news cycle time for this tidbit of really obvious information: 10 hours.

Fortunately for us basketball fans, nobody in the NBA has ever, ever used steroids, and the worst thing you can find in a basketball player’s locker is a loaded gun.  Or, back in the ’80s, coke.  Lots and lots of coke.  Speaking of basketball role models, what about J.R. Smith, who somehow avoided indictment on a vehicular manslaughter charge after killing one of his own passengers when he ran a stop sign in 2007?  Well, as J.R. knows, you can’t spell “manslaughter” without “laughter”.  Here’s a video, and maybe it’s best to let that and the wise man say it all: “Sometimes all you can do is smash a plastic guitar.”

The Slamson Memorial Virtual Round Robin

Team	VRR Record	VRR %	Actual Record	Actual %	Difference
nutria on viagra	106-35-2	74.8 %	9-4-0	69.2 %	-5.6 %
Lipsticked Pigs	101-42-0	70.6 %	9-4-0	69.2 %	-1.4 %
Dry Heat	92-49-2	65.0 %	9-4-0	69.2 %	4.2 %
Slamson	89-53-1	62.6 %	12-1-0	92.3 %	29.7 %
The Bombay Runners	85-56-2	60.1 %	4-9-0	30.8 %	-29.3 %
Sweet Sassy Molassy	81-61-1	57.0 %	12-1-0	92.3 %	35.3 %
Bodacious D	71-71-1	50.0 %	10-3-0	76.9 %	26.9 %
Smash	68-72-3	48.6 %	4-9-0	30.8 %	-17.8 %
Hiphopapotamus	64-77-2	45.5 %	4-9-0	30.8 %	-14.7 %
Baker’s Poop Makers	40-101-2	28.7 %	3-10-0	23.1 %	-5.6 %
The Very Bad Team	38-102-3	27.6 %	1-12-0	7.7 %	-19.9 %
Mandookie	13-129-1	9.4 %	1-12-0	7.7 %	-1.7 %

The Objective Stud

Stud	SDAAs	Min	FGM	FG%	FTM	FT%	3PTM	3PT%	OReb	DReb	Ast	St	Bl	A/T
Lamar Odom	16.79	159	18	52.9	7	70.0	3	37.5	16	49	28	6	2	7.00
Chris Paul	15.41	160	24	40.7	12	70.6	3	25.0	6	12	54	12	0	6.00
Dwyane Wade	14.93	152	48	47.1	27	77.1	6	28.6	12	12	22	9	2	2.20
Antawn Jamison	13.27	156	43	55.8	25	80.6	7	43.8	4	31	5	4	0	0.83
David West	11.84	146	36	66.7	13	86.7	1	50.0	7	22	6	5	5	1.00
Brandon Roy	11.66	167	35	54.7	24	75.0	5	31.2	1	15	24	5	1	3.43
Brook Lopez	11.16	138	23	48.9	19	90.5	0	NaN	8	27	12	5	6	2.00
Kevin Durant	11.15	114	28	59.6	31	93.9	5	50.0	2	23	9	3	0	0.75
LeBron James	10.96	117	32	54.2	29	76.3	6	37.5	4	19	23	2	2	1.92
Andre Iguodala	10.83	120	17	56.7	16	94.1	3	30.0	6	17	22	8	2	2.75

The Objective Turd

Turd	SDAAs	Min	FGM	FG%	FTM	FT%	3PTM	3PT%	OReb	DReb	Ast	St	Bl	A/T
Courtney Lee	-6.22	105	9	28.1	0	NaN	1	14.3	3	9	7	3	0	3.50
Omri Casspi	-2.12	111	16	39.0	7	70.0	5	35.7	5	16	3	0	1	0.75
Michael Beasley	-1.92	101	20	42.6	7	77.8	1	25.0	5	18	4	2	2	0.50
Hedo Turkoglu	-1.03	101	11	39.3	4	50.0	5	45.5	0	8	13	2	1	4.33
Stephen Jackson	-0.63	133	26	38.8	12	63.2	2	15.4	5	13	11	7	2	1.38
Matt Barnes	-0.07	104	19	63.3	3	50.0	3	50.0	4	20	3	2	0	0.50
Nene Hilario	-0.06	103	14	50.0	9	69.2	0	NaN	6	16	8	4	2	0.89
Marc Gasol	0.37	150	14	48.3	8	47.1	0	NaN	6	19	8	2	9	0.62
Ron Artest	0.91	122	12	41.4	0	NaN	6	60.0	5	13	3	3	1	0.43
Luol Deng	0.91	140	21	38.2	12	63.2	0	0.0	10	19	6	3	5	0.75

Game Capsules

Slamson def. The Very Bad Team, 12-1-0: Well, once the teacher, always the teacher.  Once the student, always the student.  Looks like Slamson spent this week marking his territory all over the VBT.  All Over. (VBT knows the hand signal…)  Lebron only had three games this week, so I should probably point out somebody else.  How about Marcus Camby, who had 49 rebounds, 13 assists, and only 2 turnovers?  6 steals, 6 blocks…that’s a solid week.  Meanwhile Team Living Down To Expectations shot free throws well, and that’s about it.  Derrick Rose rocked 33 FGs and 32 assists, just the latest week in his attempt to reconvert the VBT after his initial disappointment in the Sorcerer Kid this year.

Sweet Sassy Molassy def. Mandookie, 12-1-0: Not to take anything away from Sweet Sassy, but the Dookster didn’t exactly, how do we say, change his lineup this week, and that 9.4% in the VRR would be pretty easy for anybody to beat.  All things considered, this may be the biggest whooping I can remember.  Baron Davis’ 33 assists were 12 more than Dook’s total!  (Looks like Deron Williams picked a good week to skip two games.)  Jonas Jerebko had a nice spot, but is beginning to look like he’s not the greatest thing to come out of Sweden since the Volvo 122S after all.  On the other side of the ledger, if it weren’t for Joakim Noah (25 FGM, 49 rebounds), the Dook might have actually had negative numbers in a few categories.

Bodacious D def. Baker’s Poop Makers, 10-3-0: Bo D went all Man-Dook on the Poopmakers and totally forgot to set his lineup Sunday – costing one category (O-Rebs).  So don’t say I never did anything for you, eh Baker?  Former Rookie-Of-The-Year candidate Brandon Jennings went 4 of 17 from the field and 1 of 6 from three as he continues to hit the rookie “wall”.  With Michael Redd out and defenses keying entirely on him, will his shooting percentages go down?  When can we declare Jennings the worst player to ever go for 55?  Does anybody remember when the basketball-headed Eric Gordon declared predraft that he was a PG in the NBA?  (No seriously, look at his head, it’s shaped exactly like a basketball.)  Yeah, well I sure do, and I was a bit skeptical as he had never been anything but a slightly undersized 2-guard.  Well, to this date he’s averaging 2.8 assists per game, and managed a whole 3 last week in nearly 100 minutes.  I’m thinking that was a little bit of a fib, Eric.  You score.  It’s what you do.  Duh.

Dry Heat def. Hiphopapotamus, 9-4-0: Now we finally see what happens when the Heat get to the Hip-Hop.  I guess it doesn’t take a nation of millions after all – it just takes Zach Randolph: 33 FGM, 25 FTM, 48 rebounds, and apparently 2 heat-check threes (missed them both).  Still, Monta Ellis’ 1 of 7 from three hurt the Dry Heat more.  The Hip-Hop-Potty-Mouth had a rough week: Dirk shot .400, Nene only grabbed 22 rebounds, Deng shot .382 (and .632 from the foul line), and Michael Redd shot .345 before tearing two of his CLs and going down for the season, with (premature?) speculation that his career might be done.  Oh, and Carmelo was out with a bruised knee.  That’s not a winning combo.  (And after all that, they won FG% anyways…)

Lipsticked Pigs def. Smash, 9-4-0: Conventional wisdom says that Pigs like to eat trash.  Apparently they like to eat things that rhyme with “trash” as well.  Chris Paul had only 54 assists this week.  Compare this to Joel “The White Hole” Przybilla, who has recorded 54 assists since December 12th.  Of 2007.  I know, I know, he’s out for the season, but it’s still OK to kick a guy in the patellar tendon when he’s down, right?  D-Wade made 48 FGs this week, which is quite a few.  Still, Smash’s best player was Lamar Odom, who ended up at #1 in the Stud rankings for the first time that I remember.  Hard not to when you shoot 53% from the field, grab 65 rebounds, and dish out 28 assists.

nutria on viagra def. The Bombay Runners, 9-4-0:  Open up the Bombay, here come the ABRs!  Kenyon Martin, a career 23.4% three-point shooter, went 1 for 1 this week – and it was a good thing for the ABRs that he hit it, as they won the category by 0.001!  In what is either a somewhat remarkable coincidence or a deliberate design, the ABRs have the top FOUR free throw shooters on the season:  Randy Foye (95.1%), Steve Nash (94.3%), Manu Ginobili (90.1%), and Ray Allen (90.1%).  Those four went 54 of 57 combined this week, for 94.7%.  For the Runners, David West went 36 of 54 from the field this week (66.7%) – his 29 boards and 5 blocks were nice as well.  And now that the Hornets have traded away Hilton Armstrong for nothing, those numbers can only go up, right?  (Kidding!  Armstrong hasn’t even seen the court since mid-December.)

 

Side O’ Stats, Week 10

Are three-point shooting big men by definition “soft”?  Does the presence of a three-point shooting big man preclude a team from winning a championship?

These two statements are certainly held true by some people, and I have had my share of internet arguments over the validity of these supposed truisms.  One of my favorite arguments to shoot down is this one, paraphrased: “With the possible exception of Bill Laimbeer, no team with a three-point shooting center has ever won a championship.  Therefore, three-point shooting centers preclude teams from winning championships.”  It’s a fairly easy argument to counter.  First off, there have been a grand total of ten centers (at least, centers as defined by databasebasketball.com, whose Ultimate Stat Search makes arguments like this easy to win) who have played at least 100 games in their career and averaged — get this — 0.4 three-point attempts per game or more.

As you can see, three-point shooting centers are quite a rare commodity.  If you were to make an argument that three-point shooting centers, by virtue of being three-point shooting centers, actually hurt their teams’ chances of winning it all you might want to be able to point to more than 10 examples of failures — especially when one of those ten is Bill Laimbeer himself.

But, let’s be a bit more fair.  We can actually look up season-by-season stats for the same data, and find that, for centers playing at least 20 games in the season, exactly 26 seasons have been recorded with an average of 1.0 three-point attempts per game.  Bill Laimbeer has four of those 26, and in two of those seasons…you guessed it, Laimbeer’s Pistons won the championship.

So that means that of all teams that have boasted a three-point shooting center, 2/26 = 1/13 = 7.7% have won the championship.  Of course, in any given year, only 1 in 30, or 3.3% of all teams win a championship.  Does this, then, argue that teams that have a center who shoots at least one three a game have a BETTER chance of winning a championship?  Well, I wouldn’t argue that (rather, I would guess that three-point shooting from a center is such a small factor that it plays almost no role in determining championships, and the advantage we see above is just a small-sample-size blip), but it certainly doesn’t provide a whole lot of evidence for the prior claim.

If only shooting down such arguments meant they didn’t pop back up again a month later.  Alas…

An interesting twist on this one came by me just today, though in this case it referred to a player traditionally cast as a power forward rather than a center.  I’ll quote in the interest of thoroughness:

The funny thing is that as Dirk himself has gotten more and more serious about winning, he has taken fewer and fewer threes. When he was young and dumb he was chucking nearly 5 a game . Its been years since that number was even 3 (JJ: well, 2.9 in ‘07-’08 probably counts), and this year its down to fewer than 2 a game. And that’s from the PF most often pointed to in all of basketball as the example of three point shooting working for a big.

In addition to the statistical clarification, I’d like to note the interesting psychological analysis made by the individual who wrote this.  When Dirk was young, he shot more threes “because he was a happy-go-lucky chucker” but as he got older, he shot fewer threes “because he became more serious about winning”.  That’s an argument you just can’t lose.  Look!  Dirk got serious about winning, so he took fewer threes.  How do I know he got serious about winning?  Because he took fewer threes!  (When you get down to it, this is a version of “begging the question”, folks, and it’s a classic logical fallacy.)

One thing is true - Dirk has taken fewer threes lately

(Note: in the interest of time I’m not fixing it, but the Y-axis should be “3-point Attempts per 36 minutes“, not “per game”.

But this does provide us with another point of attack – if Dirk’s threes were hurting his team, at least from an individual point of view, we should be able to find a correlation between the number of threes that he took and other statistics.  That is, when he took more threes, he should have suffered in other categories.

Unfortunately, there isn’t a lot of data here, and it’s hard to get significant correlations with only 12 points in data as noisy as these.  We’ll have to go with trends for the most part.  There is, in fact, only one comparison which is significantly correlated (and highly so at that):

As you can see, there’s a strong negative correlation between the percentage of shots Dirk took from the outside and the percentage of team possessions he used while on the floor.  This is a bit weird and I don’t quite have an explanation.  As Dirk’s role in the offense increased, the percentage of shots he took from outside went down.  There could be a lot of explanations for this, but someof them somewhat resemble that psychological argument made above.

On the other hand, there are no other stats that correlate with three-point attempts in the fashion we might expect if three-point attempts were an accurate proxy for “non-seriousness about winning”.   Defensive rebounding?  Trends higher when shooting more threes.  Offensive rebounding?  Trends higher when shooting more threes.  (This is actually an interesting comparison, because one argument that can be made against a post player shooting threes is that taking position outside will reduce the chance of the player getting an offensive rebound.  That is clearly not the case here.)  Offensive efficiency (TS%)?  Trends higher when shooting more threes.  Offensive rating (team points scored per possession while player is on the floor)?  Trends higher with shooting more threes.  Defensive rating (opponent points scored per possession while player is on the floor)?  Trends lower with shooting more threes.  Steals?  Positive correlation again.  Blocks?  No correlation at all.

Every single one of those statistics show trends that indicate that Dirk was a MORE effective player when he was shooting more threes, not a LESS effective player.  OK, blocks were flat. And I didn’t mention, assists did in fact have a slightly negative trend.  So it’s not unanimous.  But it’s pretty obvious Dirk did not become a more effective player when he started shooting fewer threes.  The threes weren’t hurting him, but rather the more threes he shot, the more effective of a player he was.

Why did Dirk stop shooting so many threes?  I don’t really know.  The change doesn’t correspond to the coaching changes in Dallas.  Some of it might have had to do with his poor shooting year in ‘03-’04.  That year he shot a career-low .341 from deep, 18 points worse than he has ever shot, and 38 points worse than his prior career low (excepting his very spotty and low-minutes rookie year). That same year, he took 1.1 fewer threes per game, perhaps because he was shooting poorly, and the subsequent year he dropped another 0.9 threes per game, perhaps in another compensation for his one bad year.  But that’s just speculation.  It’s interesting to note that for the ‘02 and ‘03 seasons (see graph above) Dirk was a teammate of Raef LaFrentz, the single-greatest volume three-point shooting center ever.  Was there an internal battle between the two?  Who knows.  The one thing that appears certain to me is that Dirk was just as effective (if not more) in his “happy-go-lucky chucking” days as he is now in his “serious about winning a championship” days.  The threes didn’t hurt him.

And with that, my rant about three-point shooting big men, and whether they kill a team, is over.