APBRmetrics

Posted: **Tue Dec 15, 2015 6:06 pm**

Mike G wrote: As I mentioned, players of all eras wind up averaging about 7-7.5% of their career minutes in playoffs. There are no meaningless playoff wins. If you win a playoff series, you get another series of 4-7 games, and a chance at yet more.

OK, maybe you didn't fully understand my point on meaningless playoff wins. The Cavs won 14 playoff games last season, but fell short of the championship by 2 games.

If one were to count every playoff win in history the same - then the Cavs get more playoff wins this last season than almost every NBA CHAMPIONSHIP team pre 1984.

The 2013 Pacers won 11 playoff games, didn't make the finals. Almost every NBA champion pre 1984 took 12 wins MAX to win the championship. That's quite a few more wins than Bill Russell would get credit for in 8 of his 11 championships. In 7 of Bill Russell's championships, the Celtics won 8 playoff games. In 1 other, they won 7. Should Pacer players in 2013 who didn't even make the finals get 37.5% MORE credit for playoff production than any Celtics' players get on any of those 8 championship teams?

I understand the concept of no meaningless playoff wins - BUT winning 8 playoff games in today's game is pretty worthless (relatively speaking) compared to winning 8 playoff games (& winning the championship) in, say, 1966.

Again, if we count all playoff wins the same - the career player great lists will be DOMINATED by modern players when compiled, many of which never having won a championship.

Posted: **Tue Dec 15, 2015 6:50 pm**

Yeah, one solution, which I use in rankings is to credit members of championship teams with another 10 games.
Every playoff series but the Finals leads to another playoff series, for the winner. If the 1960 Celtics could play another round, they'd get another 4-7 games, plus a 50% chance at another 4-7, etc.
If a series averages 5.5 games, an infinite series (5.5 + 2.75 + 1.375 + ...) adds up to 11.
Since there won't ever be infinite playoffs, I round down to 10.

But there's a reasonable argument that winning the title in 1960 is no greater an accomplishment than reaching the conference finals today. There are 4 times as many teams (almost), presumably because there's 4 times as much talent, competition, etc.

From 1968 to 1976, the NBA champs did not have to play the ABA champs. After 1976, the ABA was in the mix. The NBA title was then a bigger prize. The NFL winner met the AFL winner in the early Super Bowls. It's just better to be the only champ, rather than one of two.

... if we count all playoff wins the same - the career player great lists will be DOMINATED by modern players ...

Why is that? On average, players got more playoff minutes (relative to RS minutes) in the '50s and '60s.

In 1960, there were 25 playoff games after 300 RS games. PO games were therefore 7.7% of the total. Same-ish as ever.

Posted: **Tue Dec 15, 2015 10:22 pm**

Mike G wrote:
... if we count all playoff wins the same - the career player great lists will be DOMINATED by modern players ...
Why is that? On average, players got more playoff minutes (relative to RS minutes) in the '50s and '60s.

In 1960, there were 25 playoff games after 300 RS games. PO games were therefore 7.7% of the total. Same-ish as ever.

I JUST explained it - why do I need to explain it again? Treating every playoff game the same devalues greatly championships won by older era players. This last season, four teams teams had as many or more playoff wins than Bill Russell's Celtics had in 8 of their championships. So - these current players on these four teams, playing against the top 500 or so players in the world, deserve more or much more credit for their playoff performance this last season than those Celtic players did playing against the top 130 or so players in the world in each of their championships?

If you disagree with me - why in the world do you give a championship "bonus" in your own work to get the results to "look" right. Don't give that championship bonus, treat every playoff win as exactly the same - see how the career results compile.

I starting to think I have never made a post that you haven't argued with, & you just like to troll me. Seriously.

Posted: **Wed Dec 16, 2015 1:09 am**

There are games won, and there are Pythagorean wins. These are known, so we can compare your adjusted playoff wins to some multiples of pWins. A factor of 2 makes some similarities:

Code: Select all

15po  pW     x2    aW
GSW  15.8   31.6   42
Cle  12.8   25.6   25
Atl   7.9   15.8   12
LAC   7.6   15.2    8
Chi   7.5   15.0    7
Hou   7.2   14.4   13
Was   6.3   12.6    7
Mem   5.0   10.0    7
SAS   3.6    7.2    3
Brk   1.9    3.8    2
Dal   1.9    3.8    1
Por   1.4    2.8    1
NOP   1.0    2.0    0
Bos    .9    1.8    0
Mil    .8    1.6    2
Tor    .5    1.0    0

Note that if we add 10 to the Warriors pWx2, it's the same as your aW.

To clarify, the 10-game championship bonus is to approximate the benefit of winning a series -- which is to get more playoff games / win opportunities. It's certainly possible to give players of all eras their due, without assuming one era has an inherent competitive edge.

Posted: **Wed Dec 16, 2015 1:10 pm**

Taking the focus off the playoff debate for a moment, I originally set the criteria to 15,000 career RS MP, 15.0+ PER, 1.0+ BPM, and a WS/48 of .125+. Then, I decided to include anyone who met at least one of the 3 PER, BPM, WS/48 benchmarks, but now as I filter through the career RS MP list, I'm debating just including EVERYONE over 15,000 minutes played. Players like R.Mahorn, Avery Johnson, Sean Elliott, PJ Brown, Cuttino Mobley, Boris Diaw, Vernon Maxwell, Allan Houston, Bruce Bowen aren't included. I'm debating this because it will essentially turn it into a sort of "longevity" list, but we're all aware that better players play more minutes, better play produces more minutes, and by merely playing a lot of minutes, a player is assumed to have a decent skill set. Im leaning towards including all of them. I think it would make it a more comprehensive list.

Posted: **Wed Dec 16, 2015 2:09 pm**

That makes all kinds of sense and probably isn't any more work.
You might Also include players with shorter careers, who are good enough to meet some standards of WS or VORP; that 15,000 minutes cutoff misses Blake Griffin, Sabonis, DeAndre, Ibaka, Bill Walton; and George Mikan, along with other old-timers who date back before 1952.
http://bkref.com/tiny/f6Uxe

Posted: **Wed Dec 16, 2015 2:58 pm**

Here's my updated Hall Ranking, as it currently stands. I created an "Equivalent BPM" from WS/48 and MPG, and then for seasons with BPM available I averaged the two.

I also adjusted all playoffs and seasons to the current length. (This helps older players).

The actual Hall Rating is Positive Value over Average + Positive Value over Finals Average (which is +2).

https://docs.google.com/spreadsheets/d/ ... sp=sharing

Posted: **Wed Dec 16, 2015 6:12 pm**

More is good. Here's the streamlined Top 25; Hall Rating, Regular Season and Playoff component:

Code: Select all

rk   player           HR    RS   PO
1 Kareem AbdulJabbar 232   149   83
2   Wilt Chamberlain 224   141   84
3   Michael Jordan   223   133   90
4   LeBron James     207   127   80
5   Tim Duncan       168   100   68
6   Magic Johnson    164    90   74
7   Bill Russell     158    78   80
8   Karl Malone      157   121   36
9   Julius Erving    154    93   61
10   Charles Barkley 150   108   42
11   Jerry West      149    81   68
12  Shaquille O'Neal 146    88   59
13   David Robinson  142   104   38
14   Larry Bird      140    88   52
15   Oscar Robertson 131    93   38
16   Hakeem Olajuwon 126    77   49
17   Kobe Bryant     122    76   45
18   Kevin Garnett   121    99   22
19   Artis Gilmore   113    81   32
20   Dirk Nowitzki   112    76   36
21   John Stockton   110    76   34
22   Scottie Pippen  107    59   48
23   Bob Pettit      104    71   33
24   Clyde Drexler   103    72   31
25   Chris Paul      101    77   24

Erving (9) and Gilmore (19) are surprisingly high. I guess you didn't make any ABA adjustment?

How did you scale playoffs such that Wilt has a higher PO-HR than Kareem or Duncan?
http://bkref.com/tiny/fHTT3
Some playoff numbers:

Code: Select all

player   G    Min    PER   WS/48   WS    BPM   VORP   POHR
Duncan  241  9152   24.6   .197   37.5   6.0   18.4   67.6
Kareem  222  8500   23.5   .199   35.3   5.9   13.3   83.4
Wilt    160  7559   22.7   .200   31.5                83.8

It can't be championships that moves Wilt up; and that doesn't seem to help the others much here.

Posted: **Wed Dec 16, 2015 7:57 pm**

Mike G wrote: Erving (9) and Gilmore (19) are surprisingly high. I guess you didn't make any ABA adjustment?

How did you scale playoffs such that Wilt has a higher PO-HR than Kareem or Duncan?
http://bkref.com/tiny/fHTT3
Some playoff numbers:
Code: Select all
player   G    Min    PER   WS/48   WS    BPM   VORP   POHR
Duncan  241  9152   24.6   .197   37.5   6.0   18.4   67.6
Kareem  222  8500   23.5   .199   35.3   5.9   13.3   83.4
Wilt    160  7559   22.7   .200   31.5                83.8
It can't be championships that moves Wilt up; and that doesn't seem to help the others much here.

Correct, no ABA adjustment... yet.

I scaled the playoffs based on the length of them--since the playoffs weren't as long in Wilt's day, his numbers will be bumped up accordingly to be equivalent to the current length of playoffs.

Posted: **Wed Dec 16, 2015 8:03 pm**

Not sure if anyone's done this yet, but in my opinion the way regular season should be scaled is how it correlates to increasing odds of winning a title.

I.e. Delta % Odds Of Winning Title ~ x1*Regular Season VORP + x2*Playoff VORP

Posted: **Wed Dec 16, 2015 8:20 pm**

bbstats wrote:Not sure if anyone's done this yet, but in my opinion the way regular season should be scaled is how it correlates to increasing odds of winning a title.

I.e. Delta % Odds Of Winning Title ~ x1*Regular Season VORP + x2*Playoff VORP

I use a factor of 2 for playoffs in the ratings above.

Posted: **Wed Dec 16, 2015 11:01 pm**

...since the playoffs weren't as long in Wilt's day, his numbers will be bumped up accordingly to be equivalent to the current length of playoffs.

Wilt's teams lost in the playoffs every time but one. If the playoffs were longer, would his other 12 forays have been any longer?

From the top 100 Hall Ratings, here are some anomalously high PO/HR -- the fraction of players' HR which are from playoffs:

Code: Select all

rk   player          HR     RS     PO    %PO
83   Robert Horry   43.9   13.7   30.2   .69
70   Roger Brown    48.5   19.1   29.5   .61
60   Bob Cousy      54.5   23.3   31.2   .57
48   Sam Jones      63.2   27.8   35.4   .56
43   Bill Sharman   66.6   29.8   36.8   .55
72   Isiah Thomas   47.7   21.6   26.1   .55
93   James Worthy   41.6   19.9   21.6   .52
52   Cliff Hagan    59.3   28.9   30.4   .51
47   John Havlicek  64.7   31.8   32.8   .51
7    Bill Russell  157.5   77.8   79.7   .51
61   George Mikan   54.3   27.0   27.4   .50
33   Elgin Baylor   74.7   38.6   36.1   .48
88   George Yardley 41.9   22.3   19.6   .47
44   Zelmo Beaty    66.6   36.0   30.5   .46
11   Jerry West    149.2   81.3   67.9   .46

Seems there is a preponderance of old-timers from precisely the era that is also historically high in avg PO/RS minutes. Why do they need an artificial boost?

These guys are known playoff stepper-uppers, for sure. Well, Zelmo got a lot of ABA playoff minutes.

Posted: **Thu Dec 17, 2015 10:31 am**

And the other side of the coin: From the top 100 RS HR, the lowest % of HR from playoffs:

Code: Select all

rk     player          HR     RS    PO    %PO
110 Dominique Wilkins 36.8   34.4   2.4   .07
109  Andrei Kirilenko 36.8   33.2   3.6   .10
100  Grant Hill       39.8   35.9   4.0   .10
50   Neil Johnston    62.0   54.4   7.6   .12
115  Bob McAdoo       35.5   30.5   5.0   .14
73   Elton Brand      47.4   40.5   6.9   .15
59   Tracy McGrady    54.7   46.6   8.1   .15
124  Detlef Schrempf  33.5   28.5   5.0   .15
123  Chris Mullin     33.6   28.1   5.6   .17
117  Mookie Blaylock  34.8   29.0   5.8   .17
75   Chris Webber     45.7   37.3   8.4   .18
18   Kevin Garnett   121.3   99.0  22.3   .18
63   Larry Nance      53.6   43.7   9.9   .18
69   Vince Carter     49.3   39.8   9.6   .19
36   Gary Payton      71.2   57.2  14.1   .20

After Johnston, McAdoo would be the next oldest, being '70s-80s.
The rest are into the '90s or later. Relatively under-represented by their postseasons.

Posted: **Thu Dec 17, 2015 3:27 pm**

Mike G wrote:And the other side of the coin: From the top 100 RS HR, the lowest % of HR from playoffs:After Johnston, McAdoo would be the next oldest, being '70s-80s.
The rest are into the '90s or later. Relatively under-represented by their postseasons.

You have a valid concern, and I have made some adjustments:

1. If there was no earlier round in the playoffs (no first round), then that "length" isn't included on the playoff run (I had been doing so).
2. All series that were played, however, will be scaled up to have gone 7 games, no matter how long they actually were (I don't want to penalize sweeps, or 5 game series!)
3. Instead of weighting all playoff series x2 vs. regular season, I have adopted the following multipliers: 4x for Finals, 3x for conference finals, 2x for conf semi finals, and 1x for first round series. This makes sense to me, keeps an overall average of x2, and at the same time doesn't penalize older players for not playing the first round or two very much.

Here are the updated hall ratings, incorporating these changes.

Note: at some point, I will do era adjustments, and that will hurt the ABA players the most I suspect.

https://docs.google.com/spreadsheets/d/ ... sp=sharing

Posted: **Thu Dec 17, 2015 5:40 pm**

That seems to fix the most glaring disparities.
Summing all HallRating, RS and PO, by decade of players' debut, it's still higher PO/HR ratio
in the old days.

Code: Select all

year 1    po%    HR     RS    PO
. -1959   .38   1215    749   466
1960-69   .37   1578    995   583
1970-79   .34   1493    986   508
1980-89   .35   2493   1619   874
1990-99   .30   2235   1562   673
2000-12   .30   1507   1049   458

po% = PO/HR
Magic is one whose PO now account for more than half of his total HR.
Robert Horry's PO are .72 of his total. His rank moved up from #83 to #68.
Since PO minutes are only about .07 of RS minutes every year, these PO-HR seem to be magnified 4-5 times.

EDIT: Sorry, that table represents just the top 200 in Hall Rating.

APBRmetrics

Kobe Bryant and All-Time Ranking of Players

Re: Kobe Bryant and All-Time Ranking of Players

Re: Kobe Bryant and All-Time Ranking of Players

Re: Kobe Bryant and All-Time Ranking of Players

Re: Kobe Bryant and All-Time Ranking of Players

Re: Kobe Bryant and All-Time Ranking of Players

Re: Kobe Bryant and All-Time Ranking of Players

Re: Kobe Bryant and All-Time Ranking of Players

Re: Kobe Bryant and All-Time Ranking of Players

Re: Kobe Bryant and All-Time Ranking of Players

Re: Kobe Bryant and All-Time Ranking of Players

Re: Kobe Bryant and All-Time Ranking of Players

Re: Kobe Bryant and All-Time Ranking of Players

Re: Kobe Bryant and All-Time Ranking of Players

Re: Kobe Bryant and All-Time Ranking of Players

Re: Kobe Bryant and All-Time Ranking of Players