APBRmetrics • Statistical Trivia

Page 1 of 24

Statistical Trivia

Posted: Thu Feb 12, 2015 1:00 pm

by Mike G

DeAndre Jordan was 12/26 from the FT line last night. The 12 FT is easily his highest for the year, and the .462 FT shooting boosts his season FT% from .408 to .414.
In fact, he's hit 51% over his last several games, to bring it up from .390

For the season, his FG% (and eFG%) is .725, and his TS% is .667 -- a difference of .058, all of course due to bad FT shooting.

DeAndre set a record of sorts last year, in how much his FT shooting dinged his TS%, relative to his eFG%.
Ranked by eFG% - TS%, minimum 300 FGA:

Code: Select all

eFG%-TS%                 Season    Tm   eFG%    FT%    TS%    FTr
.059   DeAndre Jordan   2014-15*  LAC   .725   .414   .667   .678
.046   DeAndre Jordan   2013-14   LAC   .676   .428   .630   .726
.045   Wilt Chamberlain 1966-67   PHI   .683   .441   .637   .761

.039   Wilt Chamberlain 1971-72   LAL   .649   .422   .610   .686
.038   DeAndre Jordan   2010-11   LAC   .686   .452   .648   .636
.038   DeAndre Jordan   2012-13   LAC   .643   .386   .606   .510
.038   Wilt Chamberlain 1972-73   LAL   .727   .510   .689   .776
.037   Wilt Chamberlain 1967-68   PHI   .595   .380   .557   .677

.032   Andre Drummond   2012-13   DET   .610   .371   .578   .465
.024   Shaquille O'Neal 2006-07   MIA   .591   .422   .567   .614
.023   Andre Drummond   2013-14   DET   .623   .418   .599   .427
.020   Eric Montross    1995-96   BOS   .566   .376   .546   .384

Drummond appears to be better at avoiding the foul line, but this year his FG% has tanked.
Jordan earlier was ahead of Wilt's alltime-best .727 FG%, currently just behind.

Has anyone applied a different TS% FTA-multiplier for the era when players got 3-to-make-2 ? Wilt's FTr (FTA/FGA) would be lower under modern rules, and his TS% would be higher.

Re: Trivia

Posted: Thu Feb 19, 2015 8:58 pm

by Mike G

Among players with at least 600 NBA minutes:

Code: Select all

Ast%  career lowest    Blk%  career highest   TReb%  career highest
0.6  Yinka Dare        10.3 Hassan Whiteside  25.5 Hassan Whiteside
1.0 Hassan Whiteside   10.2  Manute Bol       23.4  Dennis Rodman
1.5  Mamadou N'Diaye    8.8  Jim McIlvaine    22.5  Earl Williams
1.6  Hasheem Thabeet    7.8  Shawn Bradley    22.4  Andre Drummond
1.7  Jahidi White       7.7  Rudy Gobert      21.8  Reggie Evans
1.8  Manute Bol         7.6  Chuck Nevitt     21.8  Jamal Sampson

Whiteside is also 1st in career DReb% and 4th in OReb%.
Earl (The Twirl) Williams: played 1882 minutes, 1975-79; mostly in his rookie year.

EDIT: Thru age 25, highest career Blk%: Bol 10.7, Whiteside 10.3, McIlvaine 9.2, Bradley 8.3 ...
TReb% -- HW 25.5, Earl 22.7, Boerwinkle 22.6, Drummond 22.4, Moses 21.9 ...
DReb% -- HW 33.7, Walton 32.5, J Sampson 30.1, Love 29.8 ...
(Walton had just won MVP at age 25)

Re: Trivia

Posted: Fri Feb 20, 2015 11:30 am

by Mike G

DeAndre Jordan last night was 8-11 on FG and 10-28 from the FT line.
He's now at .408 on FT, .725 on FG (and eFG%), with .661 TS% -- a deficit/ding of .064
That is a deficit almost 50% greater than anyone who came before. Other than Wilt, the only player with half as much ding is Andre Drummond, .032 in 2013.

Re: Trivia

Posted: Fri Feb 27, 2015 1:01 am

by Mike G

The 3 summary player rate stats at b-r.com -- PER, WS/48, and BPM -- generally distinguish better players from weaker ones. At least, on a given team, they should.
The minutes a coach allots should also correspond to 'how good' his players are.
Therefore, we should expect that a player with better stats should get more minutes than his teammates with lesser stats.

Here are the correlations between MPG and the 3 stats, by team, for each team's top 12 players in total minutes this year.
xyz is just the avg of 3 correlations.

Code: Select all

xyz   Tm    PER  WS/48  BPM      xyz   Tm    PER  WS/48  BPM
.81   LAC   .84   .77   .83      .86   Cle   .89   .83   .87
.72   Mem   .66   .72   .77      .84   Atl   .76   .86   .89
.69   Por   .74   .55   .77      .68   Chi   .70   .57   .78
.68   Sac   .63   .56   .85      .66   Was   .75   .48   .76
.65   OKC   .73   .58   .62      .48   Orl   .54   .24   .67
.61   GSW   .47   .59   .78      .45   Ind   .45   .31   .58
.60   Hou   .62   .50   .68      .44   Cha   .55   .11   .65
.58   Phx   .61   .40   .74      .43   Det   .55   .22   .53
.51   Min   .59   .35   .57      .42   Tor   .70   .27   .29
.44   Uta   .55   .43   .35      .36   Mia   .43   .06   .59
.44   NOP   .41   .24   .68      .29   Mil   .33   .11   .43
.43   SAS   .56   .18   .56      .21   NYK   .35   .04   .24
.41   Den   .41   .40   .44      .08   Brk   .19  -.22   .27
.33   LAL   .50   .00   .49      .03   Bos   .07  -.21   .22
.08   Dal   .04  -.03   .23      .00   Phl  -.02  -.44   .46

.53  west   .56   .42   .62      .42  east   .48   .22   .55

In each conference, the better teams are giving their better players (according to PER, WS, and BPM) more minutes. The Mavs seem to be an exception, the Spurs are an anomaly; in the east, the Raps and Bucks are winging it.
The Kings are the worst team with high correlations. They just got a new coach.

Does this tell us more about the stats or about the coaching?

Re: Trivia

Posted: Fri Feb 27, 2015 1:57 am

by Crow

It would help to see if RPM runs counter (positive) for the teams low ranked with boxscore metrics. That could justify the coaching. Without that save, it would not look real good for the coaching or part of it. Though is you are winning that is the main thing. There are opportunities to coast to victory or throw in the bench on an night were you are down instead of fighting or on a back to back. A single test is not sufficient to totally answer the question.

Re: Trivia

Posted: Fri Feb 27, 2015 12:03 pm

by Mike G

RPM would be another useful entry. But I don't see a practical way to download it.
eWins (per minute) has always correlated better to minutes than PER and (especially) WS/48; now that BPM is in the mix, I'm curious how they stack up.

XYZ is now the avg of 4 metrics' correlation to MPG. Dispensing with the decimal points

Code: Select all

xyz  Tm   PER   WS  BPM   eW      xyz   Tm   PER   WS  BPM   eW
82   LAC   84   77   83   83       86   Cle   89   83   87   88
69   Mem   66   72   77   60       81   Atl   76   86   89   74
68   Por   74   55   77   68       69   Was   75   48   76   79
68   Sac   63   56   85   66       69   Chi   70   57   78   69
67   OKC   73   58   62   74       53   Tor   70   27   29   84
63   Phx   61   40   74   78       51   Cha   55   11   65   72
62   Hou   62   50   68   68       51   Orl   54   24   67   58
58   GSW   47   59   78   49       46   Det   55   22   53   54
53   Min   59   35   57   59       44   Ind   45   31   58   43
49   Uta   55   43   35   65       37   Mia   43    6   59   42
48   SAS   56   18   56   62       35   Mil   33   11   43   53
45   NOP   41   24   68   49       27   NYK   35    4   24   47
43   Den   41   40   44   47       14   Brk   19  -22   27   30
41   LAL   50    0   49   65        9   Bos    7  -21   22   29
12   Dal    4   -3   23   25        6   Phl   -2  -44   46   24
                                    
55  west   56   42   62   61       45  east   48   22   55   56

eWins is lowest for Mem and Atl -- also predicted their seasons badly.
BPR is lowest for Utah. PER lowest for GSW, highest on LAC and Cle.

Of course, a high correlation does not mean a 'better' stat. Your stat could be MPG/10 (say), and you'd have correlation of 100. eWins has a small mpg component, but it's also adjusted (upward) by % of games started.
Lacking those corrections, BPM looks even better.

Re: Trivia

Posted: Fri Feb 27, 2015 12:41 pm

by Mike G

Choosing lineups of an arbitrary number like 'top 12' can allow for some anomalous results here. The Sixers are a mess however you look at them. In the table above, correlations with MPG vary as much as 90 (between WS and BPM). Here's how they look, ranked by MPG:

Code: Select all

. Sixers           Min   mpg   PER   WS/48   BPM   e484
M Carter-Williams 1391   34   12.9  -.022   -1.4   1.08
Nerlens Noel      1605   30   12.7   .051   - .5    .63
Tony Wroten        895   30   14.9   .008   -1.6   1.12
Luc Mbah a Moute  1418   29   10.7   .025   -3.4    .51

Robert Covington  1334   28   14.9   .099   + .3   1.00
Hollis Thompson   1171   25   10.0   .059   -2.4    .34
K.J. McDaniels    1319   25   10.8   .018   -2.6    .49
Henry Sims        1145   20   15.1   .074   -3.5   1.03

Jerami Grant       781   19   11.4   .059   -2.4    .39
Brandon Davies     379   19   10.7   .034   -2.9    .51
Alexey Shved       285   17   20.0   .154   - .1   1.30
JaKarr Sampson     685   13    9.6   .023   -5.0    .37

WS thinks their worst player (of these 12) has gotten the most minutes. And near the bottom is their best looking player, the departed Shved. Only BPM is not fooled into thinking he's good or great.
Just stripping Shved from the lineup, correlations improve to .73 for BPM, -.27 for WS -- the difference has increased to 100!

Re: Trivia

Posted: Fri Feb 27, 2015 6:35 pm

by sndesai1

here's a really crude excel tool to grab rpm if anybody wants it:
http://sndesai.com/apbr/misc/Real%20Plu ... %2027.xlsb

Re: Trivia

Posted: Sun Mar 01, 2015 12:19 pm

by Mike G

I came across a video (5 min.) of "20 Greatest NBA Players of the 90's"
https://www.youtube.com/watch?v=mCw7U2S6y_s

Of course, there's no explanation of how they were selected; but the concept of 'best players in an era' is kind of attractive. And nobody can clearly define an 'era', except that we are all cognizant of decades as they come and go.

Now that BPM -- and its cumulative equivalent, VORP -- are available at b-r.com, we can grab 10-year chunks of player stats, for regular seasons and for playoffs.
http://bkref.com/tiny/N1weX
It looks as if VORP are smaller than WS by an avg factor of about 2.5, at least for top players. So, to keep WS and VORP on roughly equal footing, these lists will use that factor on VORP.
How to weight playoffs? The consensus has been that playoff minutes are about 3X as relevant as RS. But there's the issue of playoff opportunity. To alleviate that disparity among players, I'm just multiplying PO by 2.

Here's the current decade (2009-10 thru yesterday) top 20

Code: Select all

.      decade   regular season     playoffs
xyz     2010's    WS     VORP     WS    VORP
303    LeBron    90.7    48.4    21.4    9.6
212    Durant    83.3    32.3    12.1    4.7
168    Paul      70.4    29.9     5.3    2.5
156    Wade      50.7    23.5    10.9    5.0
144    Harden    58.6    23.7     6.2    2.8

141    Westbrook 47.6    24.0     6.9    3.9
130    M Gasol   49.1    22.5     5.3    2.9
123    Duncan    46.1    18.2     8.4    2.9
119    Curry     49.0    23.3     2.5    1.3
115    Dwight    53.7    18.3     4.1    1.4

115    P Gasol   48.3    18.7     6.0    1.5
112    Noah      44.3    21.2     4.0    1.3
112    Dirk      52.6    15.4     5.5    1.9
107    Griffin   48.7    19.7     2.5     .8
104    Ginobili  36.7    16.1     6.2    3.0

104    Pierce    43.5    17.5     5.6    1.1
99     Millsap   43.7    18.5     2.0    1.1
98     Bosh      47.2     9.7     8.4    2.0
96     Kobe      36.7    16.1     5.8    1.5
95     Love      48.6    18.4      .0     .0

No units are assigned to the xyz column, but if you divide it by 2.25, you get a sort of estimate of "wins" in season + postseason.
It's necessarily ugly when many players' careers are chopped in half by a decade change. Still we get an overview of the last 2/3 decade.

I don't know what the previous decade is called. It's 2000-2009. The "oughts"?

Code: Select all

.      decade   regular season     playoffs
xyz     2000's    WS     VORP     WS    VORP
363    Duncan   129.9    57.6    23.0    8.6
348    Kobe     123.2    53.0    21.7    9.7
344    Garnett  131.8    68.5     9.1    4.5
316    Dirk     137.1    48.4    17.0    4.8
290    Shaq     100.4    42.1    21.2    8.4

279    LeBron    84.8    50.7    12.4    8.6
268    Kidd      86.5    49.2    11.2    7.3
255    Billups   95.3    31.2    20.3    8.3
254    B Wallace 79.2    41.1    15.4    8.3
244    Pierce    99.4    41.8     9.5    4.3

238    Allen     96.0    40.6    10.2    4.1
234    Marion    99.3    42.9     7.5    2.4
226    McGrady   88.0    48.0     4.6    1.8
215    Carter    86.3    42.6     5.3    2.4
207    Rasheed   79.6    30.3    12.1    5.5

194    Nash      95.2    25.6     9.9    3.0
192    Iverson   78.0    34.0     6.2    3.3
186    Brand     82.8    37.7     2.4     .8
178    Ginobili  59.8    27.4    12.8    4.8
176    R Lewis   78.6    30.5     4.8    2.4
171    Wade      56.7    30.7     9.0    3.9

Wade is thrown in as extra. A relative latecomer with huge playoffs.
Nash is loved by WS and unloved by VORP -- ratio almost 4:1 . Others are less than 2:1 -- McGrady, Ben, Kidd, LeBron, KG -- versatile players.

Watching that '90s video, I was reminded that the decade's best players look a lot like the best players ever. Yet several in this list are not in the video.

Code: Select all

.      decade    regular season     playoffs
xyz     1990's     WS     VORP     WS    VORP
411    Jordan    131.1    58.3    32.0   14.0
378    Malone    146.9    65.8    17.9    6.2
357    Robinson  137.8    66.4    12.6    5.5
333    Pippen     98.0    55.5    20.2   11.1
319    Barkley   112.2    56.6    14.9    7.1

291    Olajuwon  100.3    49.4    14.1    7.7
279    Stockton  120.6    41.7    14.7    5.0
262    Drexler    82.9    44.7    13.9    7.9
257    Reggie    110.8    41.7    12.0    3.6
239    Ho Grant   88.1    36.8    15.9    5.4

235    Hornacek   88.5    37.0    12.9    5.6
203    Ewing      92.6    30.4    11.6    2.3
200    Payton     80.3    36.1     7.8    2.8
194    Hawkins    81.7    33.4     7.7    2.6
191    Shaq       75.5    30.8     9.5    3.9

187    Schrempf   91.1    31.1     5.6    1.3
186    Kemp       80.7    28.9     9.5    2.9
180    Rodman     71.5    29.0     8.8    3.7
175    Blaylock   65.1    36.4     2.6    2.8
173    Majerle    65.1    30.0     7.2    3.6

Just missing : T Hardaway, KJ, Divac
Horace, Horny, and Hersey don't make many "best of" lists; these guys, Det and Dan had very 90s-centric careers.
Then there's Mookie -- beneficiary of the short arc, for sure. And VORP loves him dearly -- he has playoff VORP/WS > 1.0

Finish up with the '80s. Note Dominique does not appear, as his career was half in the '90s, with not a lot of playoffs. Bad Boy Pistons also poorly represent.

Code: Select all

.      decade    regular season     playoffs
xyz     1980's     WS     VORP     WS    VORP
397.5  Magic     120.3    60.0    27.1   14.6
397.2  Bird      124.2    66.7    23.6   11.8
259    Kareem     97.4    35.4    17.7    7.6
233    Moses     115.7    31.1    11.5    3.3
221    Dantley   103.4    34.8     8.0    3.0

218    Erving     77.7    38.5    10.4    4.7
216    McHale     87.3    27.0    17.2    5.3
214    Jordan     73.4    41.4     7.7    4.3
207    Parish     91.0    30.5    11.9    3.2
202    Moncrief   88.5    33.5     7.7    2.9

197    Cheeks     82.8    29.1    11.0    3.9
193    Sikma      89.8    35.4     4.3    1.2
182    English    82.8    31.7     5.4    1.9
177    Isiah      60.9    28.8     8.9    5.3
176    Barkley    63.1    36.2     4.7    2.6

165    Worthy     55.6    20.4    13.5    6.3
164    Laimbeer   76.6    25.0     7.0    2.1
159    Nance      67.6    30.1     3.5    1.8
150    Olajuwon   53.7    25.2     8.3    3.3
137    Gilmore    71.1    23.9     2.3     .3
136    Drexler    52.6    30.8     1.6     .7

Poor Moses -- everyone knows he was The Center of the '80s. Michael Cooper has equal playoff VORP.
I hope this settles once and for all the Bird vs Magic debate: It's a draw.

Re: Trivia

Posted: Fri Mar 06, 2015 8:37 am

by Mike G

Players who have averaged, per36 minutes, at least 23 pts, 6 reb, and 7 ast for a season (min 1500 min.)

Code: Select all

per 36   yr     Pts   Reb   Ast   Ast%
Oscar   1965   24.0   7.1   9.1   37.8
Oscar   1966   24.5   6.0   8.7   34.5
Magic   1987   23.7   6.2  12.1   47.2
Jordan  1989   29.1   7.2   7.2   34.7
LeBron  2010   27.4   6.7   7.9   41.8
Westbr. 2015   29.3   7.6   8.9   48.1

Ast% is thrown in for standardization's sake. It seems 9 assists today is roughly equivalent to 12 ast in Magic's heyday.
Given that, Westbrook tops these guys in all 3 stats.

Players who have averaged 23 Pts/36, while shooting .600 TS% and getting 7 Ast/36 :

Code: Select all

per 36   yr     Pts    TS%    Ast   Ast%
Magic   1987   23.7   .602   12.1   47.2
Jordan  1989   29.1   .614    7.2   34.7
LeBron  2010   27.4   .604    8.4   41.8
Curry   2014   23.7   .610    8.4   39.9
Curry   2015   25.9   .628    8.5   38.5

Pretty good company, and to be here twice ...

Averaging 25-5-6, shooting .600 --

Code: Select all

per 36    yr     Pts   Reb    Ast   TS%
Jordan   1989   29.1   7.2   7.2   .614
LeBron   2010   27.4   6.7   7.9   .604
LeBron   2013   25.5   7.6   6.9   .640
LeBron   2014   25.9   6.6   6.1   .649
Harden   2015   26.4   5.7   6.9   .608

Only Jordan with all 4 marks better in one season.
Missing due to TS% qualification -- Bird with .585 TS%, Oscar .583, Wade .577, West .559, Westbrook .536 (this year).

With 23 Pts/36 on .590 TS%, plus Blk% > 5.0

Code: Select all

per 36    yr     Pts   Reb%   Blk%   TS%
Ewing    1988   23.4   15.2   6.0   .594
Ewing    1990   26.7   15.7   5.9   .599
D Rob    1990   23.9   18.5   6.5   .597
D Rob    1991   24.4   19.0   6.2   .615
D Rob    1995   26.2   16.2   5.9   .602
A Davis  2015   24.6   16.5   6.0   .602

Without the TS% cutoff: Shaq .583, Parish .579, Olajuwon .577. These are from 1974 to present.

http://bkref.com/tiny/nQ0XJ

Re: Trivia

Posted: Sat Mar 07, 2015 4:18 pm

by Mike G

DeAndre Jordan went 2-10 from the FT line in his last outing, and his eFG% - TS% is now
.713 - .647 = .066
This is more than twice as great as anyone not named Wilt.

Re: Trivia

Posted: Wed Mar 18, 2015 12:59 am

by Mike G

DeAndre's eFG% - TS% is now .071
Monthly splits:

Code: Select all

mo.    Min   eFG   FGA    FT  FTA    eFG%    FT%   FTr  TSA/36   TS%   ding
Nov    532    58    80    12   28    .725   .429   .35    6.2   .693   .032
Dec    592    75   108    29   69    .694   .420   .64    8.4   .647   .048
Jan    482    76    98    28   73    .776   .384   .75    9.7   .692   .084
Feb    395    64    93    48  116    .688   .414  1.25   13.1   .611   .077
Mar    304   32.5   47    27   73    .691   .370  1.55    9.4   .581   .110

tot   2305  305.5  426   144  359    .717   .401  .843    9.1   .646   .071

Nov line includes 2 Oct games.

Have the Clipps lost games because of his FT shooting? I can find just a few possibles:

Code: Select all

opp    MOV  FTA   FT     Win     OT
Brk    -2    12    2    .417    .333
Hou    -2     8    1    .500    .375
Por    -5    10    2    .800    .700
Cle    -5    12    5    .917    .833
Den    -3     7    3   1.000    .857

Showing the FT% needed to win (or not lose) the game. All else being equal.
Just 2 games this year that they would have won with even less-terrible FT%.
Median FT% for players 6-11 and taller (>100 FTA) in last 3 seasons is 74%

Re: Trivia

Posted: Fri Mar 20, 2015 10:54 pm

by Mike G

Players who have had Pts/G > 26, Ast% > 34, TReb% > 8, and Stl% > 2.5 : Jordan '89, Westbrook '15, Harden '15
http://bkref.com/tiny/eeC6Y

Lose the Stl% qualifier -- or relax it to 2.0 -- and LeBron is there 5 times. Apparently it's impossible to do the other 3 without at least 2.0 Stl%

Re: Trivia

Posted: Wed Apr 08, 2015 12:10 am

by Mike G

There are just 7 player-seasons of >2000 minutes with PER>20 and BPM < 0.
This year, there may be 3 more.

Code: Select all

Player             Season    Tm    PER   TRB%   AST%   TOV%   USG%   WS/48  BPM
Andre Drummond    2014-15   DET   21.3   23.9    3.7    9.6   21.8   .146   -.2
Brook Lopez       2014-15   BRK   22.5   14.3    4.6    8.9   26.7   .151   -.3
Jonas Valanciunas 2014-15   TOR   20.8   19.1    3.0   11.9   19.1   .192   -.4
Corey Maggette    2009-10   GSW   20.3   10.0   13.2   13.0   26.7   .144  -1.0
Zach Randolph     2006-07   POR   22.8   17.6   12.8   12.6   33.1   .112   -.1

Patrick Ewing     1995-96   NYK   20.9   17.1   11.2   11.8   31.5   .116   -.4
Ricky Pierce      1990-91   TOT   21.0    5.3   13.0    9.7   29.7   .174   -.2
Moses Malone      1985-86   PHI   20.6   18.0    5.1   14.1   27.6   .157   -.8
Dan Issel         1983-84   DEN   21.1   13.0   11.6    8.3   26.6   .141   -.3
Moses Malone      1977-78   HOU   21.2   22.3    2.1   17.7   23.5   .139  -1.1

Sub-zero BPM indicates a below avg player.
About 30% of AllStars have PER < 20; median is about 22.0

Re: Trivia

Posted: Wed Apr 08, 2015 12:30 am

by permaximum

That's the effect of ast*reb.