## Re: [APBR_analysis] Kidd v. KJ and comparing eras

Expand Messages
• harlan - using the online historical basketball stats database at the APBR website, you can compare KJ vs. Kidd in a number of categories, including per game
Message 1 of 11 , Mar 8, 2001
harlan -

using the online historical basketball stats database at the APBR website,
you can compare KJ vs. Kidd in a number of categories, including per game and
per minute stats....

bob chaikin
bchaikin@...
• ... thought ... he ... statistical ... must ... I have been looking for an occasion to present an example of my analysis technique. Kidd vs. KJ will do just
Message 2 of 11 , Mar 8, 2001
--- In APBR_analysis@y..., harlanzo@y... wrote:
>
> My friend swears that Jason Kidd is a better player than Kevin
> Johnson ever was. Now my intial reaction was, based solely upon my
> memory of the two players in action, was that I prefer KJ. I
thought
> he could shoot and finish and while not as good a passer as Kidd,
he
> definitely was more than adequate in that department.
>
> The question I have is how to quantify this difference in
statistical
> terms. (I realize that Kidd is still young so that any argument
must
> be restricted to peak value [ie the top 3 years of each player]).

I have been looking for an occasion to present an example of my
analysis technique. Kidd vs. KJ will do just fine.
I find Kevin Johnson's peak year to be 1991, when he averaged 22.2
pts, 10.1 assists. In that year, he played about 36 minutes per game
(mpg) for a high-scoring Suns team, 114.0 ppg and 107.5 ppg allowed.
I scale all per-game averages to an arbitrary 36 mpg. In KJ's case,
he receives a scaling factor of 1. The minutes factor is then re-
scaled to give a bonus or penalty for playing more or less than 32
mpg, respectively. This involves a 4th root, and KJ's scaling factor
becomes 1.03, meaning all his per-game averages are multiplied by
this. 22.2 * 1.03 = 22.9 ppg
Another factor in scoring is the shooting pct. I use a combined pct
which considers points scored from all sources, FTs, 3-pointers, and
2s. Total shot attempts are FGA + FTA/2. Pct. = equivalent baskets
per shooting attempt. KJ shot .588 in 1991.
The historic NBA average combined pct. is about .527. Take .588/.527
as another factor in the equation, and KJ looks like a 25.5 scorer.
However, as Harlan notes, the Suns were a high-scoring team. I don't
penalize (or reward) a scorer for what his team scores, but for what
they allow their opposition to score. Against a historical average
of 100 ppg, I factor in the Suns 107.5 ppga. 25.5 * 100 / 107.5 =
23.8 ppg.

This is what KJ might average in 36 min. for a historically average
team. In other words, part of the reason for KJ's very high shooting
pct was that he was surrounded by other excellent scorers. On an
average team, he might shoot more, but at a lower pct, etc.

KJ's 3.5 rebound average is similarly scaled by his minutes factor,
1.03, and then by the total team + opponent rebound number, 45.5 +
42.6 = 88.1, almost equal to the historical avg of 88 total rpg.
3.5 * 1.03 * 88 / 88.1 = 3.6 rpg.

Assists are partly a result of game pace, which I measure by
totalling the points by both teams, in this case 114 + 107.5 =
221.5. So KJ gets 10.1 apg * 1.03 * 200 / 221.5 = 9.4 apg, in
standardized terms.
In other words, KJ's assist numbers were somewhat inflated that year.
Steals, turnovers, blocks are all scaled by the 1.03 factor.

Johnson turns out to have the following ratings for 1991:
23.8 sco, 3.6 reb, 9.4 ast, 2.2 stl, 3.6 TO, .1 blk
By the same method, Jason Kidd in 1999 turns up these numbers:
16.6 sco, 6.7 reb, 10.6 ast, 2.1 stl, 2.8 TO, .4 blk

When I add all these up, I scale assists by 1 1/3, to account for the
extra turnovers; and I scale blocks by 1 1/2, to give credit for
altered shots, etc.
Sco + Reb + Ast*1.333 + Stl - TO + Blk*1.5 = Total Rating
For their best seasons, KJ rates 38.8, Kidd 37.3
For their careers, KJ was 33.9, Kidd is presently 30.6, present
season not included.
Kidd is having a 34.8 year so far.

My rough rule of thumb is, 40 is a superstar, 35 is allstar, 30 is a
star, 25 is a starter, 20 is a contributor, 15 is a bench guy, 10 is
a scrub. 45 might be a megastar. Only a handful of players have had
a 50 season: Wilt, Kareem, Jordan, Shaq last year. Others have been
close, including Bird, the Admiral, Hakeem, and Malone.

A
> rudimentary tendex analysis of their tendex peak (kidd 97-00 and kj
> 88-91) gives KJ a huge edge 720 to 569. Of course tendex is hardly
> perfect in that it does not quantify value and does not take into
> account the difference in scoring between the NBA of the late 80s
and
> the late 90s.
>
> The second of these questions is highlighted by an examination of
the
> suns teams in different eras. KJ's higher tendex, it should be
noted
> came on a team that scored 115.8 per game while Kidd played on a
team
> that scored 98 per game. If we assume this difference in scoring
is
> not the result of one team being better but the basketball
> environment in which each played, we can assume that tendex (or
> anyother formula) is flawed. This could be adjusted by pro rating
> points and assits of each player (compared with each other and the
> league average that they competed within). However, some could
argue
> that the difference in scoring could be balanced off in the idea
that
> theoretically there could be more rebounds in that late 90s than
the
> late 80s. Indeed, the 80s/kj suns shot around 49% whereas the
> 90s/kidd suns shot closer to 44-46%.
>
> With all these variables in mind, how do we most efficiently
compare
> these players from two different eras with statistical formulas?
> (Secondarily, I would like to know who is better kidd or kj).
> Finally, I realize there is something called floor percentage but I
> don't know its formula and the problems and solutions it might
> present over other formulas. So, please give me some feedback on
> these issues. Thanks for your time.
• ... I ve personally had to work on other basketball tasks as well as real work. ... thought ... he ... Opinion first -- KJ was generally a better offensive
Message 3 of 11 , Mar 8, 2001
--- In APBR_analysis@y..., harlanzo@y... wrote:
> I see things on the site have slowed down of late so I thought I

work.

>
> My friend swears that Jason Kidd is a better player than Kevin
> Johnson ever was. Now my intial reaction was, based solely upon my
> memory of the two players in action, was that I prefer KJ. I
thought
> he could shoot and finish and while not as good a passer as Kidd,
he
> definitely was more than adequate in that department.
>

Opinion first -- KJ was generally a better offensive weapon because
he could score. He wasn't as good a passer. Kidd has always been a
good defender, though, one thing that he's never gotten the credit
for. KJ generally did seem to have a greater impact on better teams.
What Kidd does to pass a ball is legendary. Overall, I don't think
he has been as effective as KJ, but his passing flare is impressive.
He helped make Michael Finley a star that he hasn't matched
in Dallas (though KJ did, too). He made Antonio McDyess a star that
he hasn't matched in Denver. He is making Shawn Marion a star. He
even made Tom Gugliotta look good for a while. But it just bugs me
that he has weaknesses and they do get in the way of the team winning
big.

> With all these variables in mind, how do we most efficiently
compare
> these players from two different eras with statistical formulas?
> (Secondarily, I would like to know who is better kidd or kj).
> Finally, I realize there is something called floor percentage but I
> don't know its formula and the problems and solutions it might
> present over other formulas. So, please give me some feedback on
> these issues. Thanks for your time.

Floor percentage is just an offensive evaluation, representing the
fraction of a player's possessions that he contributes a score to.
The formula is complex (and has simpler versions), but I'll get it
out there eventually. I also calculate offensive ratings (points
produced per 100 possessions) and individual win-loss records.

What I have (going back only to '92 because I need to consolidate
some stats)

KJ
Scor. Poss. Floor RTG Points Def. Net Net
Poss. Pct. Prod. Rtg. Win% W-L
1992 900 1581 0.569 113.5 1794 107.2 0.721 9.6-3.7
1993 450 794 0.567 114.1 906 108.3 0.703 5.0-2.1
1994 754 1322 0.571 116.5 1540 108.3 0.768 8.7-2.6
1995 416 720 0.578 118.3 851 111.2 0.736 4.8-1.7
1996 587 993 0.591 121.0 1202 109.1 0.848 7.8-1.4
1997 738 1286 0.573 122.7 1578 108.6 0.883
11.1-1.5
1998 266 503 0.528 108.6 546 103.5 0.689 3.2-1.5
2000 21 36 0.569 121.8 44 100.0 0.963 0.3-0.0

Kidd
1995 547 1129 0.484 102.3 1155 106.8 0.329 4.0-8.2
1996 759 1628 0.467 104.1 1694 107.0 0.388 5.8-9.1
1997 369 749 0.494 108.7 814 104.8 0.647 5.3-2.9
1998 576 1193 0.483 104.5 1247 98.7 0.720 9.7-3.8
1999 489 922 0.530 112.4 1037 99.4 0.883 8.6-1.1
2000 557 1150 0.484 104.1 1197 95.7 0.801 9.6-2.4

Totals
Scor. Poss. Floor RTG Points Def. Net
Poss. Pct. Prod. Rtg. Win% W-L
KJ 4131 7235 0.571 117.0 8462 107.0 0.777 51-15
Kidd 3298 6771 0.487 105.5 7144 102.8 0.610 43-27

At their peaks, they are proving to be about equally effective. KJ
really didn't have much of a ramp-up period (I don't show it),
though, which makes sense since he actually stayed in school 4 years.
Kidd may have co-won rookie of the year, but KJ actually had better
numbers and Kidd won a lot of hype because Dallas didn't win many
games.

Between the two, you probably take KJ if you need a scorer and Kidd
if you need a distributor. Both help a lot.

Dean Oliver
• ... teams. ... Brain spasm here. Kidd never played with Finley. They were traded for each other (something I even wrote about when it happened). I had been
Message 4 of 11 , Mar 8, 2001
--- In APBR_analysis@y..., "Dean Oliver" <deano@t...> wrote:
> >
>
> Opinion first -- KJ was generally a better offensive weapon because
> he could score. He wasn't as good a passer. Kidd has always been a
> good defender, though, one thing that he's never gotten the credit
> for. KJ generally did seem to have a greater impact on better
teams.
> What Kidd does to pass a ball is legendary. Overall, I don't think
> he has been as effective as KJ, but his passing flare is impressive.
> He helped make Michael Finley a star that he hasn't matched
> in Dallas (though KJ did, too). He made Antonio McDyess a star that

Brain spasm here. Kidd never played with Finley. They were traded
for each other (something I even wrote about when it happened). I had
been making this list in my head for a while of the forwards that
seemed to play best with Kidd. And Finley accidentally creeped onto
it. I do have to wonder how Marion would do with another team. His
numbers this year are pretty gaudy, but how much of that is Jason
Kidd. Of course, neither Mashburn nor J. Jackson played very well
with Kidd.

Dean Oliver
• ... My point is that McDyess was MUCH better with the Suns than he has ever been with the Nuggets -- this All-Star year fully withstanding. Gugliotta before
Message 5 of 11 , Mar 8, 2001
>
> Just some observation here Dean-
>
> While I think having Kidd to distribute the ball helps somewhat, most
> of Marion's points come from short jumpers around the basket and
> offensive rebounds. (his outside shot still isn't that reliable) Kidd
> certainly is helpful when it comes to the transition game, he's
> probably one of the best ever at getting the ball up the court.
>

My point is that McDyess was MUCH better with the Suns than he has ever
been with the Nuggets -- this All-Star year fully withstanding. Gugliotta
before his hurts was also much better with Kidd giving him the ball.
Rodney Rogers and Clifford Robinson seems to have improved a little with
Kidd distributing the ball (not as much). There is a Kidd effect, I
think. I do wonder how big it is on Marion, who was not deemed to be as
good as he apparently is.

> As for the Kidd-KJ comparison, first I'll have to note that I'm a big
> KJ fan, so keep that in mind. Kidd's strengths of course are passing,

Frankly, I am a big KJ fan, too. Ever since Cleveland, in fact.

My thought is that, ignoring defense, KJ is clearly the better player.
Defense is a much harder thing to get your hands around, but Kidd looks
better to me statistically and observationally. That is more important in
this defense-oriented era than it was in KJ's earlier days.

Two ways of looking at the comparison then:

1. Would the Suns of right now win more with KJ than they are with Kidd?
2. Would most teams win more with KJ than they would with Kidd? (This
gets at my overall evaluation of players.)

I can do a quick analysis of the first one (but not right now during my 5
minute break from the torture known as Charnock -- my environmental
project). Bob Chaikin can definitely do this and I'd be curious to see
that. Use KJ at the same age as Kidd.

The second question is a lot more extensive. But Bob could take a stab
with his software, too.

> Until then I'll pick Johnson because hew was virtually unguardable and
> his teams were much more successful that Kidd's have been so far.

This gets at Q#2 and I would tend to agree (not strongly). Looking at the
current Suns team is a tougher call for me. I think Kidd is a very
important part of a good defense and KJ would hurt that. (Completely
ignoring the injury factor, too.)

Dean Oliver
www.tsoft.com/~deano/index.html
deano@...
• ... Personally, I think Johnson would be a better fit on the current team considering their weaknesses (which are scoring and perimeter shooting - two areas he
Message 6 of 11 , Mar 8, 2001
> This gets at Q#2 and I would tend to agree (not strongly). Looking
> at the
> current Suns team is a tougher call for me. I think Kidd is a very
> important part of a good defense and KJ would hurt that. (Completely
> ignoring the injury factor, too.)

Personally, I think Johnson would be a better fit on the current team
considering their weaknesses (which are scoring and perimeter shooting
- two areas he has a big advantage over Kidd). Kidd would probably fit
in much better on some teams such as Philadelphia or the Lakers where
he would have a big scorer to compliment him. And I agree Kidd's a
better defender, but smaller, quicker guards do give him a problem at
times.

Now who's better - Paul Westphal or Dennis Johnson? Talk about
complete opposites....

__________________________________________________
Do You Yahoo!?
Get email at your own domain with Yahoo! Mail.
http://personal.mail.yahoo.com/
• ... Very opposite. I really will have to work to dig up their numbers. I m still accumulating old data through a source with the league. The Group database
Message 7 of 11 , Mar 8, 2001

> Now who's better - Paul Westphal or Dennis Johnson? Talk about
> complete opposites....
>

Very opposite. I really will have to work to dig up their numbers.
I'm still accumulating old data through a source with the league. The
Group database that we started with APBR has very incomplete
individual turnover data and that's what my source is getting for me.
For just these two players, I can hand enter the info, but it's a
total of about 3200 man-seasons of individual turnovers I'm missing.

I always respected Johnson's game more than Westphal's. But, frankly,
I was pretty young when Westphal was in his heyday. And my judgment
is somewhat clouded by my general disdain for Westphal's coaching
style. I'll run the numbers soon.

Dean Oliver
• ... For their careers: Westphal: sco 20.8, reb 2.5, ast 5.6, stl 1.7, TO 3.2, blk .4 DJohnson: 14.4 4.3 5.2 1.5 2.6 .7 regular
Message 8 of 11 , Mar 8, 2001
> Now who's better - Paul Westphal or Dennis Johnson? Talk about
> complete opposites....
>
For their careers:
Westphal: sco 20.8, reb 2.5, ast 5.6, stl 1.7, TO 3.2, blk .4
DJohnson: 14.4 4.3 5.2 1.5 2.6 .7

regular season totals: Westphal 29.9, DJ 25.4
playoff totals: Westphal 25.8, DJ 25.7

Westphal was the more skilled player in general.
Dennis Johnson played better under pressure and against greater
competition.

I rate DJ #66 alltime, Westphal #120
>
>
> __________________________________________________
> Do You Yahoo!?
> Get email at your own domain with Yahoo! Mail.
> http://personal.mail.yahoo.com/
• ... Johnson of course enjoyed a much longer period of productivity and was far superior defensively, Westphal was as good offensively as Maravich an Gervin in
Message 9 of 11 , Mar 8, 2001
> For their careers:
> Westphal: sco 20.8, reb 2.5, ast 5.6, stl 1.7, TO 3.2, blk .4
> DJohnson: 14.4 4.3 5.2 1.5 2.6 .7
>
> regular season totals: Westphal 29.9, DJ 25.4
> playoff totals: Westphal 25.8, DJ 25.7
>
> Westphal was the more skilled player in general.
> Dennis Johnson played better under pressure and against greater
> competition.
>
> I rate DJ #66 alltime, Westphal #120

Johnson of course enjoyed a much longer period of productivity and was
far superior defensively, Westphal was as good offensively as Maravich
an Gervin in the late 70s when he was playing in Phoenix before his
foot problems began.

An interesting trade - Westphal, the great offensive player, for
Johnson the great defensive player. Obviously Phoenix got the better
of the deal with Westphal's injury in Seattle, but up to that point I'd
consider Westy superior.

Did your alltime rating take into account longevity or team W-L record?
(just curious)

__________________________________________________
Do You Yahoo!?
Get email at your own domain with Yahoo! Mail.
http://personal.mail.yahoo.com/
• To all who wonder: Mike Goodman s alltime rankings, based on statistics compiled since 1952, are based on standardized per-game averages, career totals,
Message 10 of 11 , Mar 8, 2001
To all who wonder: Mike Goodman's alltime rankings, based on
statistics compiled since 1952, are based on standardized per-game
averages, career totals, per-minute averages, and playoff performance.
>
> Johnson of course enjoyed a much longer period of productivity and
was
> far superior defensively, Westphal was as good offensively as
Maravich
> an Gervin in the late 70s when he was playing in Phoenix before his
> foot problems began.
>
> An interesting trade - Westphal, the great offensive player, for
> Johnson the great defensive player. Obviously Phoenix got the
better
> of the deal with Westphal's injury in Seattle, but up to that point
I'd
> consider Westy superior.
>
> Did your alltime rating take into account longevity or team W-L
record?
> (just curious)
>
>
> __________________________________________________
> Do You Yahoo!?
> Get email at your own domain with Yahoo! Mail.
> http://personal.mail.yahoo.com/
Your message has been successfully submitted and would be delivered to recipients shortly.