since his defensive variation from 0 is basically 6x
his offensive one, can we conclude that he's out there
for his defense? or is that stretching it?


--- Roland Beech <roland@...> wrote:
> there are pages for each player with the on/off
> influence, so for Patterson see
> http://www.82games.com/03POR12D.HTM
>
> on a points per 100 possessions basis the Blazers
>
> scored +0.2 more per poss. with Patterson out there
> allowed -1.3 fewer points on defense with him
>
> ...so they were +1.5 points per 100 poss. net better
> with him (which contrasts with his straight +/-
> on/off effect that does not adjust for the higher
> defensive possessions faced)
>
> as for the suggestion for the study I'll make a
> note, and yes it would be easy enough to do, and yes
> I imagine most of the guys with imbalanced
> possessions are better in the facet for which they
> are on court more (eg take out the offensive guy
> when the team is facing a def. poss coming up)
>
>   ----- Original Message -----
>   From: Gabe Farkas
>   To: APBR_analysis@yahoogroups.com
>   Sent: Tuesday, July 27, 2004 11:39 AM
>   Subject: Re: [APBR_analysis] Re: the Ruben
> Patterson effect
>
>
>   well that in itself might be an interesting study:
>   does Patterson's presence as a "defensive stopper"
>   influence the other team's scoring ability at all?
>
>   it might be interesting to look at guys who are on
> for
>   way more offensive than defensive possessions, or
>   vice-versa, and see if they really do have an
> impact.
>
>   is that possible?
>




__________________________________
Do you Yahoo!?
Yahoo! Mail Address AutoComplete - You start. We finish.
http://promotions.yahoo.com/new_mail

--- In APBR_analysis@yahoogroups.com, "Roland Beech" <roland@t...> wrote:
> it's a bit of a play on Kevin's references to the "Hedo Turkoglu
effect" but in the Patterson effect it refers to the fact that players
may be on court for more defensive possessions than offensive ones, or
vice versa, which biases the raw plus/minus type numbers.
>
> Patterson was the last time I checked the most extreme case of this,
facing something like 50-60 more defensive possessions, which with the
value of a possession being roughly 1 point, means he had like a -60
'handicap' to his raw plus/minus just on account of the way he was
substituted.
>
> By doing the on/off numbers based on points per possesion instead of
pure points you can weed out this problem.

As a defender, I've never quite seen how Patterson is considered a
great defender.  He's a decent defender, but even his splits on the
court don't make a great case.  In 2004, the D was 1 pt/100 poss
better (not big), but in 2003, it was about the same (estimating from
Roland's data).  Most of my techniques suggest he is a point or two
better than average at his position.  Having seen the "Kobe-stopper"
get trounced many times also influences me.

DeanO

Dean Oliver
Author, Basketball on Paper
http://www.basketballonpaper.com
"Basketball On Paper aims to lay out new measures for examining team
and player performance and must be considered the finest effort in
this regard yet seen for basketball." 82games.com's Roland Beech

While Bowen does do the heavy lifting, shouldn't Bowen actually be
responsible for his pathetic PER (whether or not his man shoots 7/20
or not hes still being outscored, outrebounded, outpassed even on
most of his good days and oppositions bad days).

I can understand some uncertainty, but to say Hedo or any other
Spur's PER is solid only cause Bowen looked the best guy up -- well
thats like saying Bowens 7pts 2reb 2ast is akin to getting 10-11ppg
5reb 4apg or something like that.

Sure its nice Bowen does well on his man, but wouldn't you rather
have 18pts on 7-17 shooting 6reb 5 asts in 33 min over Bowens 6pts
2reb 2asts on 2-4 shooting in 34 minutes?

At some point Hedo/DBrown/Ginobili have to be helping the team out
somehow no?

I don't know if opposing PER means much as far as individual defense,
but it could mean that the opposing player as at least SOLID on D if
he outperforms his opposing positional player (whether he guards them
or not the entire game) -- or at least more efficient than one other
perimeter player on the floor.

How can you really grasp individual defense unless you chart each
shot attempt that was contested by an opposing defender?

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Wed, 28 Jul 2004 NikoTMP@... wrote:

> How can you really grasp individual defense unless you chart each
> shot attempt that was contested by an opposing defender?

A good question, but it doesn't go far enough. You should be charting how
many shot attempts are contested by each defender. A guy like Bowen or
Artest will be contesting many more shots than somebody like Turkoglu or
Stojakovic, and that should count.

And also: You should chart how many shot attempts are uncontested while a
defender is on the floor, particularly for perimiter defenders like Bowen
and Artest. A good perimiter defender will rush to cover a man who's
gotten open somehow (a pick or whatever).
- --
GnuPG public key available from
http://homepage.mac.com/danflemming/gnupg_key.html
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: Made with pgp4pine 1.76

iD8DBQFBB7ZRER493M5r4PIRAsa3AKCHeEF7AnegBVhyaJxC89DnRraATQCgjwYn
W1VseIqC0zXaxH3IliTYcek=
=PuXK
-----END PGP SIGNATURE-----

Patterson doesn't have to be a superior defender to help Portland as
situational defensive sub.  He only needs to be better than the guy
he's replacing in the lineup.

--- In APBR_analysis@yahoogroups.com, "Dean Oliver" <deano@r...> wrote:
> --- In APBR_analysis@yahoogroups.com, "Roland Beech" <roland@t...>
wrote:
> > it's a bit of a play on Kevin's references to the "Hedo Turkoglu
> effect" but in the Patterson effect it refers to the fact that players
> may be on court for more defensive possessions than offensive ones, or
> vice versa, which biases the raw plus/minus type numbers.
> >
> > Patterson was the last time I checked the most extreme case of this,
> facing something like 50-60 more defensive possessions, which with the
> value of a possession being roughly 1 point, means he had like a -60
> 'handicap' to his raw plus/minus just on account of the way he was
> substituted.
> >
> > By doing the on/off numbers based on points per possesion instead of
> pure points you can weed out this problem.
>
> As a defender, I've never quite seen how Patterson is considered a
> great defender.  He's a decent defender, but even his splits on the
> court don't make a great case.  In 2004, the D was 1 pt/100 poss
> better (not big), but in 2003, it was about the same (estimating from
> Roland's data).  Most of my techniques suggest he is a point or two
> better than average at his position.  Having seen the "Kobe-stopper"
> get trounced many times also influences me.
>
> DeanO
>
> Dean Oliver
> Author, Basketball on Paper
> http://www.basketballonpaper.com
> "Basketball On Paper aims to lay out new measures for examining team
> and player performance and must be considered the finest effort in
> this regard yet seen for basketball." 82games.com's Roland Beech

> While Bowen does do the heavy lifting, shouldn't Bowen actually be
> responsible for his pathetic PER (whether or not his man shoots
> 7/20 or not hes still being outscored, outrebounded, outpassed even
> on most of his good days and oppositions bad days).

That isn't really relevant to my argument. What I'm saying is that
while both Turkoglu and Bowen end up with opponent PERs around 11, in
reality it's probably more like Bowen's is 8 or 9 and Turkoglu's 13
or 14, because they are crossmatched nearly half the time, with Bowen
taking the better offensive player.

This is a key reason I'm careful not to take the opponent PER too
literally. It doesn't necessarily imply anything about Bowen's
relative value to Turkoglu, although I'll point out that just saying
Turkoglu is more productive overall doesn't necessarily make him more
valuable to the Spurs, who can afford to have Bowen not producing
much on offense because of the guys they have around him, especially
now with a very offensive-biased player in Brent Barry replacing
Turkoglu.

> You should chart how many shot attempts are uncontested while a
> defender is on the floor, particularly for perimiter defenders like
> Bowen and Artest. A good perimiter defender will rush to cover a
> man who's gotten open somehow (a pick or whatever).

I'm going to disagree here ... I think a Bowen's/Artest's
responsibility to go to an open shooter is much lower because they're
covering the best offensive player on the court and that is their
primary responsibility. Players get left open for a reason.

A player like Iverson who plays a defensive back-style roaming
defense would be much more responsible for open shooters in my
opinion.

I constructed a pretty simple win over replacement level rating
based on the PER difference.  I pretty casually set the replacement
level at -5 (I had in mind a player with 10 per, opponent with the
average 15).  I also conveniently, but I think not too inaccurately
decided that a team of replacement level players would be lucky to
win any games.  After all, last year the best player from the
previous year by PER was on a league worst 21 win team.  I genuinely
think a team of replacement level players would have a much rougher
time in the NBA than in MLB.  But this allowed me to divvy up wins
based on how number of minutes played * (Perdiff  + 5).  Those with
per difference less than -5 were simply given zero wins.

I also looked at salaries a bit.  The league average for marginal
dollar spent/marginal win is approximately 1.3 million per win.  [I
simply took the league average payroll - 12*366,000.  For an
individual I just examine (salary - 366,000)/wins. ]

I give here the top 40 players.  The Permin is simply a perminute
rating based on my win rating.  Based on my calculations, among top
40 players, only Shaq and Michael Finley were overpaid (paid more
than 1.3 mill/win) last year.

This was really just done for fun, but I'd like to hear any thoughts
people might have.

ben

Name Team Minutes Salary Wins  Permin mill/win
Garnett min 81% 28.00 27.8 34.3 0.995
Duncan san 63% 12.00 20.3 32.2 0.573
O'Neal lal 62% 26.50 18.8 30.3 1.390
Kirilen uta 73% 1.00 18.5 25.3 0.034
Ming hou 68% 3.50 16.5 24.2 0.190
Bryant lal 62% 13.50 16.3 26.2 0.807
O'Neal ind 70% 12.00 16.0 22.9 0.725
Artest ind 68% 5.20 13.2 19.4 0.366
Cassell min 72% 5.30 12.8 17.7 0.386
Redd mil 76% 2.90 12.5 16.4 0.203
Davis nor 68% 10.10 12.4 18.3 0.783
Gasol mem 62% 3.40 12.4 20.0 0.245
Marbury tot 81% 13.50 12.3 15.1 1.068
McGrady orl 67% 13.30 12.2 18.2 1.060
Jeffers njn 79% 1.60 12.1 15.4 0.102
Randolp por 77% 1.20 12.1 15.8 0.069
Pierce bos 78% 12.00 12.1 15.5 0.965
Nowitzk dal 74% 12.00 12.0 16.2 0.968
Carter tor 70% 12.00 12.0 17.1 0.973
Wallace det 77% 5.50 11.6 15.0 0.443
Brand lac 67% 11.00 11.5 17.1 0.926
Allen sea 54% 13.50 11.1 20.5 1.184
Jones mia 76% 12.30 10.6 13.9 1.127
Odom mia 76% 8.10 10.6 13.9 0.731
Marion pho 81% 9.70 10.5 13.0 0.886
Dampier gsw 60% 7.50 10.5 17.5 0.678
Ilgausk cle 64% 13.50 10.3 16.1 1.278
Kidd njn 62% 12.20 10.2 16.5 1.159
Billups det 70% 5.00 10.0 14.3 0.464
Finley dal 70% 13.30 9.9 14.1 1.310
Ginobil san 56% 1.50 9.7 17.3 0.117
Jamison dal 60% 12.00 9.7 16.1 1.203
Boozer cle 65% 0.60 9.5 14.7 0.025
James cle 79% 4.00 9.4 11.9 0.387
Miller ind 57% 4.50 9.1 16.0 0.452
Magloir nor 70% 1.70 9.1 13.1 0.146
Posey mem 61% 4.90 8.9 14.6 0.510
Anthony den 75% 3.20 8.8 11.8 0.321
Martin njn 57% 5.10 8.8 15.4 0.540
Maggett lac 66% 5.90 8.7 13.2 0.636

Kevin Pelton wrote:

>What I'm saying is that
>while both Turkoglu and Bowen end up with opponent PERs around 11, in
>reality it's probably more like Bowen's is 8 or 9 and Turkoglu's 13
>or 14, because they are crossmatched nearly half the time, with Bowen
>taking the better offensive player.
>
I think Calbert Cheaney's season with the Warriors may be another
example of this.  He had a lousy per but a decent +/-, and he was
frequently matched up against either the other teams stud (ex, Marion),
or sometimes even the other teams PF (ex. Camby).

I forgot to add Abdur Rahim's Atlanta and Portland totals together,
they put him at 9.5 wins, good for 33rd place.  But this also adds
him to the overpaid category since his marginal dollar/win is 1.376,
higher than the league average which is about 1.3 million.

ben

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Wed, 28 Jul 2004 kpelton08@... wrote:

> > You should chart how many shot attempts are uncontested while a
> > defender is on the floor, particularly for perimiter defenders like
> > Bowen and Artest. A good perimiter defender will rush to cover a
> > man who's gotten open somehow (a pick or whatever).
>
> I'm going to disagree here ... I think a Bowen's/Artest's
> responsibility to go to an open shooter is much lower because they're
> covering the best offensive player on the court and that is their
> primary responsibility. Players get left open for a reason.

Hmmm, there you have a point.

Perhaps you could get around it by charting the FG% of opposing shooters who
took open shots while the player was on the floor.

Still, though: Guys like Bowen and Artest often cover guys who would
normally be double-teamed. Because they're covering the guy one-on-one, that
should cut down on the number of open shots taken.
- --
GnuPG public key available from
http://homepage.mac.com/danflemming/gnupg_key.html
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: Made with pgp4pine 1.76

iD8DBQFBCTdyER493M5r4PIRArTMAJwJFc1xA/tShC4GzxT/XyDkOyzImwCgg+e+
qMlxNXDX5SFQVaUtyeUDWMM=
=v2W/
-----END PGP SIGNATURE-----

> That isn't really relevant to my argument. What I'm saying is that
> while both Turkoglu and Bowen end up with opponent PERs around 11,
in
> reality it's probably more like Bowen's is 8 or 9 and Turkoglu's 13
> or 14, because they are crossmatched nearly half the time, with
Bowen
> taking the better offensive player.

Bowens opponent PER around 8 or 9?  Is that what you are saying?
Just unclear thats all.

Im not sure I know about that.  There are several games where
scorers  score 22pts and do other things on the floor while shooting
something like 9-22 against Bowen.  The problem is if Bowen is not
hitting his three pointer he is a very limited player on the floor.
Thats nice if his man shoots 40% or less, but his man is also gonna
get to the line 7-8 times, grabs 4-5 rebounds, gets 3-4 assists while
Bowen gets barely anything else but a 3-6 shooting statline, thus hes
still going to be outproduced greatly.  Thing is Bowen has Duncan to
offset that as you alluded too.  But does that mean Hedo and Ginobili
are merely outproducing their opponents because of Bowen?  If Bowen
isn't hitting his shot hes really no better than some of the worst
starting perimeter players in the league on average.  Of course if
his shot is on he is EXTREMELY effective.

I think the Spurs will become better the more offense they get on the
perimeter.  I would take the gamble and play Bowen a few min less,
than take a simple defensive approach in the regular season, then
possibly get burned in the playoffs anyway (as Bowen did -- he was a
BIG reason they lost to LA).

I am just trying to understand where you are coming from in your
analysis and opinion of the Spurs perimeter players.

Are Hedo and Ginobili really slightly below mediocre and carried by
Bowens D?  Would the Spurs even need Ginobili or Hedo over Bowen to
be effective like they were last season? It seems they would be
better served by simple shooters (spot up even, not neccesarily
balanced players), Bowen, Parker and Duncan.

I just find it hard to beleive that Bowen holds his man to average of
8 PER and that without him the Spurs D would collapse on the
perimeter.  It would indeed get worse, but I think the offense might
make up for it eventually as their perimeter becomes more acclimated
(Barry, Manu, Devin and Parker).

What do you think?  Is the PER differential meaningless?  Which
ranking system do you value or use out of curiosity?

I think PER (both a player's own and his opps) are too team dependant
for role players. Well coached teams try to maximize their scoring
chances and minimize the the opps. It doesn't matter to the Spurs if
Bowen doesn't score much bcs they have other options that are much
better. As long as Bowen is playing on the floor with Duncan / Parker
and either Manu or Turkgolu, it's better for the Spurs for the
majority of the offensive production to go through them (Bowen's
lousy PER has no real effect on the team's PER). Similiarly, Bowen
covers the toughest match-ups to limit their chances. Kobe / T-Mac /
Peja / 'Melo / etc. are not going to be putting up 8-9 PER's but
Bowen's defense should be making them underperform their average (if,
as generally accepted, he is a good defensive player).

Until PER and PER diference finds a good way to recognize these
factets, I think you will continue to have trouble applying PER Dif
to second tier individual players. With the best players it works
very well since thier role generally don't change. No matter who Shaq
plays with, the offense will run through him and Marbs pretty much
will always have the ball in his hands. Role players, however, will
have more variance. Consider for example if Bowen was playing next to
Kobe and Kobe was constantly matching up against the T-Macs while
Bowen was covering Stevensons. Would this have a dramatic effect on
Bowen PER dif, you bet, although Kobe - who already gets stuck
covering T-Mac - would be unchanged.

It will be interesting to see Harping performance with a more
talented Utah team and how Brian Cardinal fits into Mem.

How can one account for these "Bowen" effects?

Most people here never give an objective analysis to how they would
rank players combined with stats such as per diff etc......

But most will quickly mention the effect Bowen has despite putting up
bad stats across the board. Isn't that somewhat objective as well to
say that Bowen scoring nothing is fine as long as he has Duncan and a
few other mediocre scorers?

Does it mean that any non all star who has a very good per diff that
its pretty much meaningless?  But stats like DanVal or others who
favor guys like Brent Barry are more meaningful?

I am having a hard time seeing how any of these forumlas really
seperate from another.  I tend to like the PER diff best of most of
them.  I think it accounts for more than simple points scored etc...

I also like ppfga and comparing it with Team PSA and seeing how it
changes for a player from year to year.  It helps see whos efficient
and whose not.

For example Q and Joe Johnson are thought of as clear cut top 10 sgs
by many objectively, but their stats show they are terribly
inefficient at scoring (well below average even for non all stars).
Not to say they dont have other skills, but most would not know this
without seeing stats such as PSA, or even per diff to see how well an
opposing SG does against them etc.....

I guess I am asking, what statistical forumas are better than PER
diff in your opinions and why?  I know there is no holy grail stat,
but why are total efficiency stats meaningful or any measure for that
matter if you don't account for the ACTUAL matchup for a player?
Many people are quick to mention Bowens value (which the stats don't
show AT ALL).

So where are stats more meaningful than objectivity?  Where do they
meet?

The Wallace from Detroit is Ben Wallace, right?

--- In APBR_analysis@yahoogroups.com, "wimpds" <wimpds@y...> wrote:
>
> Name Team Minutes Salary Wins  Permin mill/win

> Wallace det 77% 5.50 11.6 15.0 0.443
> Brand lac 67% 11.00 11.5 17.1 0.926
> Allen sea 54% 13.50 11.1 20.5 1.184
> Jones mia 76% 12.30 10.6 13.9 1.127
> Odom mia 76% 8.10 10.6 13.9 0.731
> Marion pho 81% 9.70 10.5 13.0 0.886
> Dampier gsw 60% 7.50 10.5 17.5 0.678
> Ilgausk cle 64% 13.50 10.3 16.1 1.278
> Kidd njn 62% 12.20 10.2 16.5 1.159
> Billups det 70% 5.00 10.0 14.3 0.464
> Finley dal 70% 13.30 9.9 14.1 1.310
> Ginobil san 56% 1.50 9.7 17.3 0.117
> Jamison dal 60% 12.00 9.7 16.1 1.203
> Boozer cle 65% 0.60 9.5 14.7 0.025
> James cle 79% 4.00 9.4 11.9 0.387
> Miller ind 57% 4.50 9.1 16.0 0.452
> Magloir nor 70% 1.70 9.1 13.1 0.146
> Posey mem 61% 4.90 8.9 14.6 0.510
> Anthony den 75% 3.20 8.8 11.8 0.321
> Martin njn 57% 5.10 8.8 15.4 0.540
> Maggett lac 66% 5.90 8.7 13.2 0.636

> Does it mean that any non all star who has a very good per diff
> that its pretty much meaningless?  But stats like DanVal or others
> who favor guys like Brent Barry are more meaningful?

No, it's not meaningless. But it has to be put into context. If
Brent Barry rated well by one system, and incredibly poorly by
another, that would make it difficult to tout him. Instead, Barry
rates very well by most all of our methods, including PER
Difference. That's why I'm comfortable saying he's better than he's
conventionally regarded.

I'm not comfortable saying Bruce Bowen and Hedo Turkoglu are elite
NBA players because that's how WinVAL rated them this season,
because of the discrepancy with the other available evidence.

It can be a crutch to say you have to look at many different
numbers, but it's absolutely true. Rating NBA players remains just
as much an art as a science, and, no matter what number you use, if
you strictly rank players by it, you're going to have blind spots.

I would also say the quality of the system you're using is not as
important as the understanding of its limitations. The possession-
based numbers I use are probably even more context-dependent than
Opponent PER; a decently large component of a player's defensive
rating is based on his team's performance, whether he's leading the
charge or riding a teammate's coattails. I understand that and I
account for it. It does a better job with defense than linear
weights do, but it still can't accurately capture the value of a
Bruce Bowen. I understand that and I account for it.

You need to create explanations that make logical sense. When I say
that Bowen's Opponent PER is probably lower because of the way he
was deployed defensively, I can make an argument to back that up.
(The 8 or 9 PER I threw out there is probably ridiculous, but I
don't see why he couldn't match Tayshaun Prince, who was around a 10
Opponent PER.)

Am I confident that when I, say, rank free agents, that I've got
them in the right order? Not tremendously. I know there's a great
deal of uncertainty and importance of role and situation and all the
things DeanO points out when he talks about the lack of a Holy Grail
rating system. In the meantime, I try to do the best I can with
what's available.

There's a line of thinking out there that the way you evaluate
players is to consider both

a.) How well they fill the role they're asked to fill, and
b.) How important that role is

To some extent, it's unfair to compare Allen Iverson and Eric Snow
on defense, because their roles are so tremendously different. As
Dean points out in Basketball on Paper, Joe Dumars may be a great
perimeter defender, but he's useless against Shaq.

But there are tradeoffs being made between signing a perimeter
defender and an interior defender. A team like the Sonics, when
looking for answers on defense, has to decide whether to build from
the outside in or the inside out. There are ways, like looking at
how money is spent and draft picks used, to determine how teams
implicitly value different roles. And there are ways to
theoretically determine where they're overrating and underrating
specific roles.

In Bowen's case, I think the majority of us would agree that his
defensive stopper role probably tends to be overrated by teams.
However, he does a very good job of doing what the Spurs ask him to
do, which is why they keep re-signing him. If his statistics aren't
great, oh well.

Yes, that's correct.  Though Rasheed's rating at Detroit was quite
high, his complete season wasn't enought to crack the top 40.

Here's Rasheed's line:

Name    Team Minutes Salary Wins  Perminute   mill/win
WallaceR   tot 60% 18.00 7.3 13.7     2.423

ben


--- In APBR_analysis@yahoogroups.com, dlirag@h... wrote:
> The Wallace from Detroit is Ben Wallace, right?
>
> --- In APBR_analysis@yahoogroups.com, "wimpds" <wimpds@y...> wrote:
> >
> > Name Team Minutes Salary Wins  Permin mill/win
>
> > Wallace det 77% 5.50 11.6 15.0 0.443
> > Brand lac 67% 11.00 11.5 17.1 0.926
> > Allen sea 54% 13.50 11.1 20.5 1.184
> > Jones mia 76% 12.30 10.6 13.9 1.127
> > Odom mia 76% 8.10 10.6 13.9 0.731
> > Marion pho 81% 9.70 10.5 13.0 0.886
> > Dampier gsw 60% 7.50 10.5 17.5 0.678
> > Ilgausk cle 64% 13.50 10.3 16.1 1.278
> > Kidd njn 62% 12.20 10.2 16.5 1.159
> > Billups det 70% 5.00 10.0 14.3 0.464
> > Finley dal 70% 13.30 9.9 14.1 1.310
> > Ginobil san 56% 1.50 9.7 17.3 0.117
> > Jamison dal 60% 12.00 9.7 16.1 1.203
> > Boozer cle 65% 0.60 9.5 14.7 0.025
> > James cle 79% 4.00 9.4 11.9 0.387
> > Miller ind 57% 4.50 9.1 16.0 0.452
> > Magloir nor 70% 1.70 9.1 13.1 0.146
> > Posey mem 61% 4.90 8.9 14.6 0.510
> > Anthony den 75% 3.20 8.8 11.8 0.321
> > Martin njn 57% 5.10 8.8 15.4 0.540
> > Maggett lac 66% 5.90 8.7 13.2 0.636

Bob,

Very eloquently put why the sum total of a combination of individual
PER do not equal a team's PER, even though the break-down of a Team
PER must equal the individual PERs.

BTW, I am guessing your stats don't indicate the move fromHoward
to 'Melo accounted for the 26 game improvement . . .just kidding, but
I would be curious to see the comps between the 2003 and 2004 Nugs
for who really did account, statistically, for the improvement (game
shares).

PS Phoenix wouldn't have been better with AI at guard, yet Philly
built two seasons around having nothing but him? Is it fair to
conclude that the supporting defensive players (and maybe people like
Bowen) on those teams were vastly underrated as opposed to overrated?


--- In APBR_analysis@yahoogroups.com, bchaikin@a... wrote:
> How can one account for these "Bowen" effects?
>
> Most people here never give an objective analysis to how they would
> rank players combined with stats such as per diff etc......
>
> this is where computer simulation of nba games can help...
>
> But most will quickly mention the effect Bowen has despite putting
up bad
> stats across the board. Isn't that somewhat objective as well to
say that Bowen
> scoring nothing is fine as long as he has Duncan and a few other
mediocre
> scorers?
>
> Does it mean that any non all star who has a very good per diff
that its
> pretty much meaningless?  But stats like DanVal or others who favor
guys like
> Brent Barry are more meaningful? 
>
> I am having a hard time seeing how any of these forumlas really
seperate from
> another.  I tend to like the PER diff best of most of them.  I
think it
> accounts for more than simple points scored etc...
>
> I also like ppfga and comparing it with Team PSA and seeing how it
changes
> for a player from year to year.  It helps see whos efficient and
whose not.
>
> For example Q and Joe Johnson are thought of as clear cut top 10
SGs by many
> objectively, but their stats show they are terribly inefficient at
scoring
> (well below average even for non all stars). 
>
> they are - at best - average SGs....
>
> Not to say they dont have other skills, but most would not know
this without
> seeing stats such as PSA, or even per diff to see how well an
opposing SG does
> against them etc.....
>
> I guess I am asking, what statistical forumas are better than PER
diff in
> your opinions and why?  I know there is no holy grail stat, but why
are total
> efficiency stats meaningful or any measure for that matter if you
don't account
> for the ACTUAL matchup for a player?  Many people are quick to
mention Bowens
> value (which the stats don't show AT ALL).
>
> So where are stats more meaningful than objectivity?  Where do they
meet?
>
> In Bowen's case, I think the majority of us would agree that his
defensive
> stopper role probably tends to be overrated by teams. However, he
does a very
> good job of doing what the Spurs ask him to do, which is why they
keep
> re-signing him. If his statistics aren't
> great, oh well.
>
> last season joe johnson was an excellent example to use for
comparing how a
> team would do with or without him using simulation, i.e. how much
better or
> worse would phoenix have been with a different SG in his place.
why? because he
> played just about the maximum time a player can play in a single
season - every
> game and 40 min/g...
>
> last season the suns went 29-53, yet had the point differential of
a team
> that should have gone 31-51. not a big difference, and simulation
shows the team
> at, on average, 30-52. now let's look objectively at joe johnson...
>
> he had a scoring FG% of just .486, over two percentage points worse
than the
> league average (.508). not awful but since he accounted for almost
1/5 (18.5%)
> of the suns scoring opportunities (FGA + FTA/2), that's not good.
he led the
> team in overall turnovers but in reality turned the ball over on
just 5% of
> his touches, which is very good. he rarely committed a foul (which
is good) but
> rarely got fouled (which is bad because he shot 75% from the line).
he was an
> average rebounder for an SG at best, but more importantly got few
steals and
> blocks. so statistically speaking he wasn't anything special, and i
have him
> rated as a slightly worse than average defender...
>
> when i run a simulation of quentin richardson on the 03-04 suns at
SG for 40
> min/g replacing joe johnson the suns improve to, on average, just
32.5-49.5, a
> 2.5 game improvement over 82 games. his scoring FG% of .479 in 03-
04 was
> actually worse than johnson's, and he turned the ball over slightly
more (on 6% of
> his touches), but he got fouled more often and rebounded better. i
have him
> rated as an even worse defender than johnson...
>
> so how do these two compare to some of the best and worst SGs
(statistically
> speaking) from around the league? run calbert cheaney for example
on the suns
> for 40 min/g replacing joe johnson and phoenix averages a 30-52
record per 82
> games. same with jamal crawford - just 30-52. using davis wesley
the suns go
> 29.5-52.5, deshawn stevenson just 29-53, with deshawn stevenson
also 29-53,
> raja bell is even worse at 28.5-53.5. with jason richardson the
suns go 34-48,
> with derek anderson 31-51, allen iverson just 31-51 (35-47 using
his 02-03
> stats), larry hughes 32-50. so it looks like what joe johnson did
for the suns in
> 03-04 is just a notch above what some of the league's worst SGs
(statistically
> speaking) would have done...
>
> having said that what about a really good SG? kobe bryant on the 03-
04 suns
> for 40 min/g replacing joe johnson has the suns go on average 41-
41, 11 full
> games better (12 games better using kobe's 02-03 stats). ray allen
makes them 8
> games better (38-44), as does tracy mcgrady (but a whopping 15
games better,
> up to 45-37, using mcgrady's 02-03 stats)....
>
> surprisingly manu ginobili makes them 7 games better (37-45). using
a tendex
> type production rating on a per minute basis i have ginobili rated
about the
> same as jason richardson and allen iverson (but nowhere near that
of bryant,
> mcgrady, or r.allen). yet simulation shows the suns 3-6 games
better with
> ginobili than with either of those two. part of the reason is i
have ginobili rated
> as a very good defender (possibly due to being on the spurs) and
iverson and
> j.richardson closer to an average defender, but also he has a
scoring FG% of
> .526, commits turnovers on just 5% of his touches (iverson and
j.richardson are
> at 6%), he's a good rebounder for an SG, plus he got the most
steals on a per
> minute basis of all SGs in the league. even if i had him rated as a
very poor
> defender, his stats are so good that he'd still make the suns 2-3
games better
> over 82 games than joe johnson did...
>
> back to bruce bowen. truly a one dimensional player, he accumlates
literally
> no statistics yet plays significant minutes. most would agree that
he is a
> very good stand-up defender (i have him rated excellent) but a
player who gets
> few rebounds, steals, or blocks, and is no factor on offense. he
gets just 0.5
> touches/min, incredibly low for a player not a PF or C. he is all
D, no O. but
> what about the opposite type of player? most would agree glenn
robinson is all
> O and no D (i have him rated a very poor defender). if you swap
robinson onto
> the spurs for bowen for 40 min/g at SF versus what bowen would do
for 40
> min/g, the spurs are on average 3-4 games worse per 82 game season.
how can that
> be if robinson gets you 16-17 pts/g?...
>
> the reason is robinson's poor defense, and his touches/min are
almost twice
> that of bowen's (0.9 to 0.5). playing 40 min/g on the spurs because
of his high
> touches/min his presence reduces tim duncan's overall touches on
offense by
> 10%-15%, and drops duncan's scoring by 3-4 pts/g. one key reason
why bowen
> plays is he doesn't need the ball on offense, yet can contribute in
a way needed
> by the spurs...
>
> so the question might be why not employ 2 or 3 bowen type players
on the
> spurs to increase duncan's touches even more? because there is a
point (i'm not
> quite sure exactly where that point is yet) where this logic breaks
down, and if
> duncan's teammates cannot be a factor on offense he will get double
and
> triple teamed consistently. but teams have employed this kind of
tactic in the past
> - in 00-01 the 76ers got to the finals with 3 of their top 5
players in
> minutes played having touches/min of 0.6, 0.6, and 0.5 touches/min
(george lynch,
> tyrone hill, and theo ratliff), players whose main contributions
were on
> defense and rebounding. consequently iverson scored over 31 pts/g.
same scenario for
> the 99-00 76ers too...
>
> simulation also shows that the league's very best players can -
individually
> - make their teams as much as 18-20 games better per 82 game season
than the
> worst players (starters) at their same position. play tim duncan 40
min/g on
> the spurs and they finish 18-20 games better per 82 game season
than if duncan
> was replaced by the likes of antoine walker, walter mccarty, or
clifford
> robinson. kevin garnett replacing duncan for 40 min/g would make
the spurs just a
> single game worse, andrei kirilenko just 3 games worse, per 82 game
season...
>
> bob chaikin
> bchaikin@b...

Kevin,

I know you are Sonic fan and that you read Bill Simmons . .so I
wanted to send him a recommendation of Portland-Seattle regular
season game from the early 90's for NBA TV. Not sure if you remember
it since Portland beat the Seattle - in triple/doule OT but it was
the basically the Blazers last hurrah in Pacific NW basketball. I
think Porter hit a 35 footer to send it into the second OT. I think
Seattle beat Portland three or four times the next season as
Kemp/Payton began to come of age. Don't suppose you recall that game?

tajallie@... wrote:
> Kevin,
>
> I know you are Sonic fan and that you read Bill Simmons . .so I
> wanted to send him a recommendation of Portland-Seattle regular
> season game from the early 90's for NBA TV. Not sure if you remember
> it since Portland beat the Seattle - in triple/doule OT but it was
> the basically the Blazers last hurrah in Pacific NW basketball. I
> think Porter hit a 35 footer to send it into the second OT. I think
> Seattle beat Portland three or four times the next season as
> Kemp/Payton began to come of age. Don't suppose you recall that game?
>

I found a contemporary description of that game, which took place on Dec 1
1990, by John Gossman on usenet: http://shorl.com/gugoprodrafipro

  Saturday night the Sonics and Blazers played in Seattle one of the most
exciting, wildest basketball games I have ever seen.   In the end Portland
won 130-124 to go 13-1, while the Sonics fell to a heart breaking 4-8 as
they start out on a five game eastern road trip.  They'll be tired when they
show up in the Garden Monday night.

  Sonics led almost the entire game, usually by more than 6 but less than 12
points.  Xavier McDaniel was the main man on offense:  he eventually
finished with 41 points, 6 assists, and a handfull of rebounds and steals.
Defense was the key though.  The normally hot shooting Blazers were held to
42% for the game, even though they shot well in the 4th period and parts of
the overtimes.  Going into the final quarter the Sonics were leading by 14
and seemed in good position.

  In the fourth quarter the Sonics had one of their common offensive
breakdowns, characterized by poor shooting, turnovers, and an inability to
run a half-court offense.  Portland charged back to lead by 2, then the
teams traded the lead until the final seconds.  The Sonics led 96-95 when
Kersey drove to the basket, several players went to the floor (including
McKey but not Kersey) and Kersey put up a layin but missed and time ran out.
However, McKey was called for his 6th foul and Kersey went to the line to
hit one and miss one and tie the game.  The replay showed that McKey was
nowhere near Kersey, and the only possible call was against Michael Cage
(who had 2 fouls) and that was questionable.  Anyway,  considering the
comeback, Portland deserved another shot.

  To my surprise the Sonics came out in the OT and led from buzzer to buzzer,
by as much as six at one point.  Portland came back again.  Porter and
McDaniel did most of the scoring again, and with 12 seconds left trailed by
2.  There followed a wild sequence that started with Porter driving inside,
two blocked shots and a missed put back then a tip to the corner where it
was saved by Sedale Threatt,  back to Cage and Williams in the middle,
tipped to courtside where it was saved by Kemp, to Kersey and McDaniel who
fought over it and stopped the clock with a foul on X.  Kersey then shoved X
(the refs missed X's little bump on Jerome) and was called for a T, which
Sedale sunk on the other end. Portland trailed by 3 with 2 seconds left.
The inbounds went to Porter, who was clearly fouled by Gary Payton.  The two
scrambled for the ball just outside the 3 point line, and Porter got off a
miracle shot, lifting the ball from the floor and up into the air as the
buzzer sounded.  It went in.  The refs counted it. The replay however,
clearly showed that the ball was still only waist high when the clock hit 0.
No matter, 2nd OT.  By the way, the Sonics top bench player, Shawn Kemp,
fouled out in the OT.

  In the second OT the Sonics again pulled away.  Gary Payton and Sedale
Threatt provided most of the scoring, while Cage was scooping up rebounds to
the final tune of 15.  Ahead by four with 8 seconds left, Payton fouled
Ainge on Portland's inbounds.  Ainge hit the FTs and Gary was gone.  Then
Portland stole the Sonics inbound and scored an easy layup to tie again.
Cage missed a short, off-balance jumper to win the game with .9 secs.

  Third OT.  Sonics are playing three guards -- McMillan, Threatt, and
Barros, along with X and Cage and Polynice and Dailey coming in for instant
D and instant O.  With the Sonics shot blockers (Kemp and McKey) and best
defender (Payton) out, Porter and Drexler took it to the hoop, while the
Sonics couldn't quite hit their shots.  The Sonics were able to eek out a
four point lead, then fell apart down the stretch to lose.

  Porter had 38 for Portland.  The game was so late no box scores are out, so
I don't know much more.  Kersey had at least 19, Williams and Duckworth were
not factors.  Robinson played well, especially defensively. For the Sonics,
Gary Payton had 8 assists and at least 12 points, including some important
drives inside at crunch time.  Threatt had a least 15, and Shawn Kemp had
only 6 points, but 9 rebounds and a number of key blocks. Biggest difference
was the usual (for the Sonics).  Portland was 35-40, while the Sonics were
only 19-21 from the FT line.  Sonics also severely lacked from 3-point range
(last two years they led the NBA in 3 pointers.  Come back Dale all is
forgiven :-)).

  A very, very exciting game.  The Sonics could have really used this win to
push them to 5-7 and give them some momentum going back east.  Instead they
are tired and 4-8.  Dale Ellis was in street clothes, but stayed on the
bench and was in the huddle supporting his team--a very important sign,
considering his fight with X last week.  He should be back in a couple more
games.

--
ed

Hey Bob, out of curiosity, based on your stats, objectivity, and what
you watch of the NBA -- what would your Top 15 SG's be in order?

Also do you have a spreadsheet or a link to download your simulations
and overall tendex per minute/touch statistics?  Im interested in
testing and viewing them if its no trouble.

Also, what do you think this says of Shawn Marion and Amare, if the
Suns can't win respectably?  Is Marion overatted slightly?  He might
be an elite SF, but is he really as good as his stats might seem on
the surface?

How much will Nash and Q help the Suns? What if you ultize Q's 6th
man of the year calibur stats in 01-02?  How do you know which stats
to use?  Could you try is 0102 stats and Nash's previous season stats
to see what the Suns might be this season record wise?

Thanks!

> I know you are Sonic fan and that you read Bill Simmons . .so I
> wanted to send him a recommendation of Portland-Seattle regular
> season game from the early 90's for NBA TV. Not sure if you
> remember it since Portland beat the Seattle - in triple/doule OT
> but it was the basically the Blazers last hurrah in Pacific NW
> basketball. I think Porter hit a 35 footer to send it into the
> second OT. I think Seattle beat Portland three or four times the
> next season as Kemp/Payton began to come of age. Don't suppose you
> recall that game?

I'm glad Ed could step in here, because that one's a little far back
for me. I was eight then, after all.

I wouldn't quite call it their "last hurrah", by the way; the two
teams met in the playoffs that last year, with the Blazers pulling
out an underrated series 3-2 thanks to the home-court advantage.

It's more of an APBR thing, I know, but anyone else interested in
posting their favorite Simmons picks (for those who don't read him:
late 80s or early 90s; not on ESPN Classic regularly)?

I would probably go with a game that is more personally meaningful
to me than it is relevant in the grand scheme of things. It was the
Sonics against the Suns on April 10, 1994. That happened to be my
12th birthday, and I was at the game, which was nationally televised
on NBC.

With a win, the Sonics would clinch the Pacific Division, which was
also important as a symbolic transfer of power in the Pacific and
the West after the Suns beat the Sonics in the Western Conference
Finals the year before.

It was a truly wild game. The Sonics went up 40-25 after one
quarter, but the Suns dominated periods two and three and led 89-75
going to the fourth. Over the final 12 minutes, the Sonics outscored
Phoenix 36-19, withstanding a late run of three-pointers, to pull
out a 111-108 win.

This game also featured one of the greatest dunks of Shawn Kemp's
career; I recall his hands getting up to about the top of the sqare
on the backboard on an alley-oop lob from Payton. Good times, good
times. That was basically the yin to that year's playoffs' yang.

Good stuff, one of the many benefits of the PER-diff is that you can
quantify the value of a superstar, as you have done here.

-- In APBR_analysis@yahoogroups.com, "wimpds" <wimpds@y...> wrote:
> I constructed a pretty simple win over replacement level rating
> based on the PER difference.  I pretty casually set the replacement
> level at -5 (I had in mind a player with 10 per, opponent with the
> average 15).  I also conveniently, but I think not too inaccurately
> decided that a team of replacement level players would be lucky to
> win any games.  After all, last year the best player from the
> previous year by PER was on a league worst 21 win team.  I
genuinely
> think a team of replacement level players would have a much rougher
> time in the NBA than in MLB.  But this allowed me to divvy up wins
> based on how number of minutes played * (Perdiff  + 5).  Those with
> per difference less than -5 were simply given zero wins.
>
> I also looked at salaries a bit.  The league average for marginal
> dollar spent/marginal win is approximately 1.3 million per win.  [I
> simply took the league average payroll - 12*366,000.  For an
> individual I just examine (salary - 366,000)/wins. ]
>
> I give here the top 40 players.  The Permin is simply a perminute
> rating based on my win rating.  Based on my calculations, among top
> 40 players, only Shaq and Michael Finley were overpaid (paid more
> than 1.3 mill/win) last year.
>
> This was really just done for fun, but I'd like to hear any
thoughts
> people might have.
>
> ben
>
> Name Team Minutes Salary Wins  Permin mill/win
> Garnett min 81% 28.00 27.8 34.3 0.995
> Duncan san 63% 12.00 20.3 32.2 0.573
> O'Neal lal 62% 26.50 18.8 30.3 1.390
> Kirilen uta 73% 1.00 18.5 25.3 0.034
> Ming hou 68% 3.50 16.5 24.2 0.190
> Bryant lal 62% 13.50 16.3 26.2 0.807
> O'Neal ind 70% 12.00 16.0 22.9 0.725
> Artest ind 68% 5.20 13.2 19.4 0.366
> Cassell min 72% 5.30 12.8 17.7 0.386
> Redd mil 76% 2.90 12.5 16.4 0.203
> Davis nor 68% 10.10 12.4 18.3 0.783
> Gasol mem 62% 3.40 12.4 20.0 0.245
> Marbury tot 81% 13.50 12.3 15.1 1.068
> McGrady orl 67% 13.30 12.2 18.2 1.060
> Jeffers njn 79% 1.60 12.1 15.4 0.102
> Randolp por 77% 1.20 12.1 15.8 0.069
> Pierce bos 78% 12.00 12.1 15.5 0.965
> Nowitzk dal 74% 12.00 12.0 16.2 0.968
> Carter tor 70% 12.00 12.0 17.1 0.973
> Wallace det 77% 5.50 11.6 15.0 0.443
> Brand lac 67% 11.00 11.5 17.1 0.926
> Allen sea 54% 13.50 11.1 20.5 1.184
> Jones mia 76% 12.30 10.6 13.9 1.127
> Odom mia 76% 8.10 10.6 13.9 0.731
> Marion pho 81% 9.70 10.5 13.0 0.886
> Dampier gsw 60% 7.50 10.5 17.5 0.678
> Ilgausk cle 64% 13.50 10.3 16.1 1.278
> Kidd njn 62% 12.20 10.2 16.5 1.159
> Billups det 70% 5.00 10.0 14.3 0.464
> Finley dal 70% 13.30 9.9 14.1 1.310
> Ginobil san 56% 1.50 9.7 17.3 0.117
> Jamison dal 60% 12.00 9.7 16.1 1.203
> Boozer cle 65% 0.60 9.5 14.7 0.025
> James cle 79% 4.00 9.4 11.9 0.387
> Miller ind 57% 4.50 9.1 16.0 0.452
> Magloir nor 70% 1.70 9.1 13.1 0.146
> Posey mem 61% 4.90 8.9 14.6 0.510
> Anthony den 75% 3.20 8.8 11.8 0.321
> Martin njn 57% 5.10 8.8 15.4 0.540
> Maggett lac 66% 5.90 8.7 13.2 0.636

Here's what I sent to Bill Simmons, verbatim:

1) the playoff game where Hakeem just clobbered D-Rob
right after Robinson won the MVP award that year. If
it's that noteworthy on the vengeance scale, then it
should be shown.

2) Game 6 or Game 7 of the Knicks/Rockets finals (Game
6 where Hakeem blocked Starks' last-second shot, and
Game 7 where Starks couldn't find the bucket).

3) the Bulls/Suns Finals game where Barkley went off
for a triple-double. Also from that series, the game
where Barkley tries to steal the ball from Pippen,
which leads to the Suns' defense being out of position
and Paxon hitting the nail-in-the-coffin 3-pointer.

4) Game 5 of the Sonics/Nuggets first-round series,
where Denver upset the Sonics and Mutumbo was on the
ground crying tears of joy after the game.

5) '97 All-Star Game where Glen Rice got the MVP of
the game.

6) Can you show all-star contests? If so, Larry's
record-setting 3-point adventure, Brent Barry's slam
dunk contest, Isaiah Rider's sideways-between-the-legs
dunk contest would be awesome.


--- Kevin Pelton <kpelton08@...> wrote:

> > I know you are Sonic fan and that you read Bill
> Simmons . .so I
> > wanted to send him a recommendation of
> Portland-Seattle regular
> > season game from the early 90's for NBA TV. Not
> sure if you
> > remember it since Portland beat the Seattle - in
> triple/doule OT
> > but it was the basically the Blazers last hurrah
> in Pacific NW
> > basketball. I think Porter hit a 35 footer to send
> it into the
> > second OT. I think Seattle beat Portland three or
> four times the
> > next season as Kemp/Payton began to come of age.
> Don't suppose you
> > recall that game?
>
> I'm glad Ed could step in here, because that one's a
> little far back
> for me. I was eight then, after all.
>
> I wouldn't quite call it their "last hurrah", by the
> way; the two
> teams met in the playoffs that last year, with the
> Blazers pulling
> out an underrated series 3-2 thanks to the
> home-court advantage.
>
> It's more of an APBR thing, I know, but anyone else
> interested in
> posting their favorite Simmons picks (for those who
> don't read him:
> late 80s or early 90s; not on ESPN Classic
> regularly)?
>
> I would probably go with a game that is more
> personally meaningful
> to me than it is relevant in the grand scheme of
> things. It was the
> Sonics against the Suns on April 10, 1994. That
> happened to be my
> 12th birthday, and I was at the game, which was
> nationally televised
> on NBC.
>
> With a win, the Sonics would clinch the Pacific
> Division, which was
> also important as a symbolic transfer of power in
> the Pacific and
> the West after the Suns beat the Sonics in the
> Western Conference
> Finals the year before.
>
> It was a truly wild game. The Sonics went up 40-25
> after one
> quarter, but the Suns dominated periods two and
> three and led 89-75
> going to the fourth. Over the final 12 minutes, the
> Sonics outscored
> Phoenix 36-19, withstanding a late run of
> three-pointers, to pull
> out a 111-108 win.
>
> This game also featured one of the greatest dunks of
> Shawn Kemp's
> career; I recall his hands getting up to about the
> top of the sqare
> on the backboard on an alley-oop lob from Payton.
> Good times, good
> times. That was basically the yin to that year's
> playoffs' yang.
>
>




__________________________________
Do you Yahoo!?
New and Improved Yahoo! Mail - Send 10MB messages!
http://promotions.yahoo.com/new_mail

Based on PER-diff, I'm expecting HUGE things from the Jazz this
season. How does this lineup do?
Let's assume they all play 36 minutes a night
Arroyo
Giricek
Kirilenko
Boozer
Okur

Cool, thank you. I think 60+ is a good possiblity because I expect
each starter to improve over last season.

If you care to share any more breakdown (strengths/weaknesses) on
that lineup I'd be interested.




In APBR_analysis@yahoogroups.com, bchaikin@a... wrote:
> Based on PER-diff, I'm expecting HUGE things from the Jazz this
season. How
> does this lineup do? Let's assume they all play 36 minutes a night
>
> Arroyo
> Giricek
> Kirilenko
> Boozer
> Okur
>
> playing the above lineup against an 03-04 schedule, and based on
who plays
> their bench minutes, simulation shows this team wins games in the
mid 50s (54-56
> wins) because of its defense, with each of the starters averaging
between
> 13-16 pts/g...
>
> bob chaikin
> bchaikin@b...

--- In APBR_analysis@yahoogroups.com, bchaikin@a... wrote:
> with derek anderson 31-51, allen iverson just 31-51 (35-47 using his
02-03

Bob's stuff typically results in similar projections as mine (I also
get about a 10 win difference putting Kobe in for Johnson, for
example).  I go about it through net points or individual win-loss
records, but the ultimate result Bob puts out of team win-loss record
is nice.

I highlight the passage above from his email to highlight the
difficulty with any of these context-sensitive approaches:  Players
change.  AI's 2004 season was a bit worse than his 2003 season, which
was worse than his 2001 season.  Why?  Was that predictable?  Is it
age?  Is it coaching?  Is it injuries (the guy doesn't seem to show
much pain)?

Bob's sim shows sensitivity to those kinds of changes.  My stuff
definitely shows sensitivity, but I'd like to think I understand those
changes better.  I knew AI would do worse once Larry Brown left, but
that doesn't mean I understand it better.

This is relevant for Q, too.  His first 2 seasons in the league were
much better than his last two.  I've been thinking it's
injury-related, but some have suggested that it's related to playing
with Corey Maggette.  There are reasons to doubt the Maggette
rationale, but if it is true, he should be better in Phoenix.  If the
injuries have dogged him the last two years, which is true, and, more
speculatively, he is injury-free in 2005, then perhaps he returns to
form.  I need to look into it more, but my sense is that Q will get a
bit closer to where he was in 2002, which is a player who is about 4
wins better than he was last year.

Finally, Bob mentioned that McGrady's 2003 season was about as good as
some of MJs best.  I can see why he says that, but one big thing would
keep me from agreeing -- consistency.  In MJ's top seasons, he was
dominant every night.  My "game-by-game win-loss record" for MJ was
typically 60+ wins and about 20 losses, implying that he had a good
night about 3/4ths of the time.  McGrady, on the other hand, had some
monster nights that made his season total look good, but also had some
mediocre nights, for a similar total of 47-28 that season.

>
> so the question might be why not employ 2 or 3 bowen type players on
the
> spurs to increase duncan's touches even more? because there is a
point (i'm not
> quite sure exactly where that point is yet) where this logic breaks
down, and if
> duncan's teammates cannot be a factor on offense he will get double and
> triple teamed consistently.

This relates to the skill curves I put in Chapter 19 of the book.
Duncan can maintain his efficiency while using a lot of possessions
(upwards of 30%), but if you have 3 Bowens using 11% of the team's
possessions, that's 3*9%=27% of the team's possessions that have to be
used by the other 2 guys.  Duncan can't handle more than 14% or so of
those before getting inefficient.  And the Spurs don't have anyone
else who can even pick up about 9% of the rest.

>
> simulation also shows that the league's very best players can -
individually
> - make their teams as much as 18-20 games better per 82 game season
than the
> worst players (starters) at their same position. play tim duncan 40
min/g on
> the spurs and they finish 18-20 games better per 82 game season than
if duncan
> was replaced by the likes of antoine walker, walter mccarty, or
clifford
> robinson.

Interestingly, my range is about 14-17, not 18-20, for these players.
  I'm not sure if you can do the comparison with reality to say which
is "right".  Regardless, that's a lot of wins between some of the best
and worst starters.

DeanO

Dean Oliver
Author, Basketball on Paper
http://www.basketballonpaper.com
When basketball teams start playing Moneyball, this is the book
they'll use!