Loading ...
Sorry, an error occurred while loading the content.

Defensive ratings

Expand Messages
  • jimmy_purnell
    I know everyone with a rating system is always working on figuring out how to judge a player s defense, so I was wondering how things have been working out.
    Message 1 of 16 , Aug 24, 2004
    • 0 Attachment
      I know everyone with a rating system is always working on figuring out
      how to judge a player's defense, so I was wondering how things have
      been working out. Who does everyone have ranked as the top defenders
      in the league last season? I'm interested not only in PER or +/-, but
      in all the different rating systems people have.
    • Kevin Pelton
      ... 84.6 C Ben Wallace 85.0 SF Andrei Kirilenko 85.7 C Chris Andersen 86.0 C Theo Ratliff* 86.1 PF Tim Duncan ... 86.4 PF Kevin Garnett 86.6 C
      Message 2 of 16 , Aug 24, 2004
      • 0 Attachment
        > I know everyone with a rating system is always working on figuring
        > out how to judge a player's defense, so I was wondering how things
        > have been working out. Who does everyone have ranked as the top
        > defenders in the league last season? I'm interested not only in
        > PER or +/-, but in all the different rating systems people have.

        84.6 C Ben Wallace
        85.0 SF Andrei Kirilenko
        85.7 C Chris Andersen
        86.0 C Theo Ratliff*
        86.1 PF Tim Duncan
        ---------------------------------
        86.4 PF Kevin Garnett
        86.6 C Marcus Camby
        86.8 C Shawn Bradley
        87.0 C Dan Gadzuric
        87.0 PF Jermaine O'Neal
        ---------------------------------
        87.1 C Stromile Swift
        87.1 C Rasho Nesterovic
        87.1 C Samuel Dalembert
        87.3 C Dikembe Mutombo
        87.6 PF Rasheed Wallace*
        ----------------------------------
        87.6 SF Ron Artest
        87.6 PF Donyell Marshall*
        87.7 SF Shawn Marion
        87.9 PF Kelvin Cato
        87.9 SG Manu Ginobili
        ----------------------------------
        87.9 C Calvin Booth
        87.9 PF Kenyon Martin
        87.9 C Zydrunas Ilgauskas
        87.9 C Steven Hunter
        87.9 SF Linton Johnson

        Kerry Kittles (87.9) is the second-best shooting guard. The top two
        point guards are Jason Kidd (88.0) and Lindsey Hunter (88.0).
      • igor eduardo küpfer
        ... I make no claims for my system, basically a points per 100 plays faced measure. I don t incorporate rebounds, because I can t really figure out how much
        Message 3 of 16 , Aug 24, 2004
        • 0 Attachment
          jimmy_purnell wrote:
          > I know everyone with a rating system is always working on figuring out
          > how to judge a player's defense, so I was wondering how things have
          > been working out. Who does everyone have ranked as the top defenders
          > in the league last season? I'm interested not only in PER or +/-, but
          > in all the different rating systems people have.
          >

          I make no claims for my system, basically a points per 100 plays faced
          measure. I don't incorporate rebounds, because I can't really figure out
          how much they're worth -- and because I think everyone else is overvaluing
          them. Here's the top 10, by position.


          1 Mike James 78
          2 Chauncey Billups 78
          3 Jamaal Tinsley 79
          4 Earl Watson 79
          5 Milt Palacio 81
          6 Eric Snow 83
          7 Raul Lopez 83
          8 Jason Kidd 83
          9 Derek Fisher 83
          10 Kirk Hinrich 84


          1 Fred Jones 80
          2 Emanuel Ginobili 81
          3 Reggie Miller 81
          4 Bonzi Wells 83
          5 Baron Davis 86
          6 Fred Hoiberg 86
          7 Dwayne Wade 87
          8 Cuttino Mobley 87
          9 Anthony Peeler 87
          10 Kerry Kittles 87


          1 Tayshaun Prince 76
          2 Aaron McKie 83
          3 Hedo Turkoglu 83
          4 Bruce Bowen 84
          5 Morris Peterson 85
          6 Ron Artest 86
          7 Latrell Sprewell 86
          8 Shane Battier 86
          9 Rodney White 86
          10 Jim Jackson 86


          1 Tim Duncan 79
          2 Kelvin Cato 83
          3 Corliss Williamson 84
          4 Kevin Garnett 84
          5 Bo Outlaw 86
          6 Robert Horry 86
          7 Mehmet Okur 87
          8 Kenyon Martin 87
          9 Zach Randolph 88
          10 Clifford Robinson 89


          1 Brendan Haywood 81
          2 Michael Olowokandi 82
          3 Mark Madsen 84
          4 Rasho Nesterovic 84
          5 Jeff Foster 85
          6 Jason Collins 85
          7 Theo Ratliff 86
          8 Elden Campbell 86
          9 Chris Bosh 86
          10 Shaquille O'Neal 86

          --
          ed
        • Mike G
          ... figuring out ... Hey, I m not. And glancing over these handful of short lists, I m not encouraged that there s much hope for it, at present. I see a
          Message 4 of 16 , Aug 25, 2004
          • 0 Attachment
            --- In APBR_analysis@yahoogroups.com, igor eduardo küpfer
            <edkupfer@r...> wrote:
            > jimmy_purnell wrote:
            > > I know everyone with a rating system is always working on
            figuring out
            > > how to judge a player's defense, ..

            Hey, I'm not.

            And glancing over these handful of short lists, I'm not encouraged
            that there's much hope for it, at present. I see a number of players
            at the top of some lists, who are at the bottom of others'.

            Since I only know defensive reputations by 2nd-sources and what
            meager viewing I manage, I'm not literally Laughing ... but let's
            face it boys, we've got a long way to go.
          • Dean Oliver
            ... I don t know if we have a LONG way to go, but we ve all admitted that defense is the most poorly evaluated aspect of the game. And because it s half the
            Message 5 of 16 , Aug 25, 2004
            • 0 Attachment
              --- In APBR_analysis@yahoogroups.com, "Mike G" <msg_53@h...> wrote:
              > --- In APBR_analysis@yahoogroups.com, igor eduardo küpfer
              > <edkupfer@r...> wrote:
              > > jimmy_purnell wrote:
              > > > I know everyone with a rating system is always working on
              > figuring out
              > > > how to judge a player's defense, ..
              > Since I only know defensive reputations by 2nd-sources and what
              > meager viewing I manage, I'm not literally Laughing ... but let's
              > face it boys, we've got a long way to go.

              I don't know if we have a LONG way to go, but we've all admitted that
              defense is the most poorly evaluated aspect of the game. And because
              it's half the game, all overall statistical ratings of players are
              somewhat questionable (Dan's method notwithstanding since he backs the
              defense out from a different set of stats).

              In terms of discrepancy, it looks like Jason Kidd being at the bottom
              of Dan's list is the biggest one I see. And Ben Wallace not even
              making Ed's list. There are others (Calvin Booth, Casey Jacobsen,
              Walter McCarty, etc.), but Kidd is a pretty big discrepancy. Note
              that Doug Steele has defensive evaluations at

              http://www.dougstats.com/03-04DefensiveTendex.html

              I use these in combination with mine because these are done in a very
              clever way that combines a lot of very good ideas.

              Reposting mine by position and trying to avoid the space limit...

              Rank Player Team DRtg
              1 wallace,ben det 87
              2 nesterovic,rado san 91
              3 cato,kelvin hou 95
              4 camby,marcus den 96
              5 ming,yao hou 97
              6 dalembert,samue phi 97
              7 collins,jason njn 98
              8 o'neal,shaquill lal 99
              9 dampier,erick gsw 101
              10 magloire,jamaal nor 101
              11 wright,lorenzen mem 101
              12 ostertag,greg uta 102
              13 blount,mark bos 102
              14 ratliff,theo atl 103
              15 miller,brad sac 103
              16 ilgauskas,zydru cle 103
              17 divac,vlade sac 106
              18 voskuhl,jake pho 107
              19 kaman,chris lac 109

              Rank Player Team DRtg
              1 duncan,tim san 88
              2 garnett,kevin min 91
              3 o'neal,jermaine ind 93
              4 martin,kenyon njn 93
              5 okur,mehmet det 94
              6 foster,jeff ind 95
              7 harrington,al ind 97
              8 williamson,corl det 98
              9 swift,stromile mem 99
              10 bosh,chris tor 99
              11 taylor,maurice hou 100
              12 thomas,kenny phi 101
              13 outlaw,bo mem 101
              14 thomas,kurt nyk 102
              15 boozer,carlos cle 102
              16 brown,p.j. nor 102
              17 grant,brian mia 102
              18 hilario,nene den 102
              19 haslem,udonis mia 102
              20 gasol,pau mem 102
              21 vanhorn,keith nyk 103
              22 thomas,etan was 103
              23 davis,antonio chi 104
              24 lynch,george nor 104
              25 stoudemire,amar pho 104
              26 skinner,brian mil 104
              27 randolph,zach por 105
              28 brown,kwame was 105
              29 curry,eddy chi 105
              30 davis,dale por 105
              31 wallace,rasheed por 106
              32 nowitzki,dirk dal 106
              33 collins,jarron uta 107
              34 abdur-rahim,sha atl 107
              35 jeffries,jared was 107
              36 brand,elton lac 107
              37 walker,antoine dal 108
              38 bibby,mike sac 109
              39 jamison,antawn dal 109
              40 gooden,drew orl 110

              Rank Player Team DRtg
              1 ginobili,emanue san 92
              2 turkoglu,hidaye san 93
              3 artest,ron ind 95
              4 bowen,bruce san 96
              5 marshall,donyel tor 96
              6 prince,tayshaun det 97
              7 hamilton,richar det 98
              8 kirilenko,andre uta 98
              9 odom,lamar mia 99
              10 jefferson,richa njn 99
              11 george,devean lal 101
              12 posey,james mem 101
              13 battier,shane mem 102
              14 butler,caron mia 102
              15 marion,shawn pho 103
              16 sprewell,latrel min 103
              17 pierce,paul bos 103
              18 cardinal,brian gsw 103
              19 dunleavy,mike gsw 105
              20 kukoc,toni mil 105
              21 anthony,carmelo den 105
              22 mccarty,walter bos 105
              23 james,lebron cle 105
              24 miller,mike mem 105
              25 patterson,ruben por 106
              26 anderson,shando nyk 106
              27 robinson,cliff gsw 106
              28 howard,josh dal 107
              29 cheaney,calbert gsw 107
              30 hayes,jarvis was 108
              31 lewis,rashard sea 109
              32 radmanovic,vlad sea 109
              33 jaric,marko lac 110
              34 howard,juwan orl 114

              Rank Player Team DRtg
              1 kittles,kerry njn 98
              2 jones,fred ind 99
              3 hoiberg,fred min 100
              4 Jackson,Jim hou 100
              5 miller,reggie ind 100
              6 mobley,cuttino hou 100
              7 peterson,morris tor 101
              8 carter,vince tor 102
              9 harris,lucious njn 102
              10 bryant,kobe lal 102
              11 alston,rafer mia 103
              12 mckie,aaron phi 103
              13 hassell,trenton min 103
              14 williams,alvin tor 103
              15 salmons,john phi 103
              16 jones,eddie mia 104
              17 richardson,jaso gsw 105
              18 welsch,jiri bos 105
              19 davis,ricky bos 105
              20 christie,doug sac 105
              21 crawford,jamal chi 106
              22 hughes,larry was 106
              23 bell,raja uta 106
              24 stojakovic,pred sac 107
              25 stevenson,desha uta 108
              26 lenard,voshon den 108
              27 jackson,stephen atl 108
              28 diaw,boris atl 108
              29 anderson,derek por 108
              30 mason,desmond mil 109
              31 redd,michael mil 109
              32 houston,allan nyk 109
              33 barry,brent sea 110
              34 johnson,joe pho 110
              35 jacobsen,casey pho 110
              36 jones,damon mil 111
              37 murray,ronald sea 111
              38 finley,michael dal 111
              39 allen,ray sea 112
              40 maggette,corey lac 112
              41 richardson,quen lac 112
              42 mcgrady,tracy orl 112
              43 bogans,keith orl 114

              Rank Player Team DRtg
              1 kidd,jason njn 96
              2 parker,tony san 98
              3 billups,chaunce det 98
              4 francis,steve hou 99
              5 johnson,anthony ind 99
              6 cassell,sam min 102
              7 fisher,derek lal 102
              8 davis,baron nor 102
              9 armstrong,darre nor 103
              10 iverson,allen phi 103
              11 wade,dwyane mia 103
              12 watson,earl mem 104
              13 claxton,speedy gsw 104
              14 payton,gary lal 104
              15 rose,jalen tor 104
              16 smith,joe mil 104
              17 miller,andre den 104
              18 snow,eric phi 105
              19 lopez,raul uta 106
              20 hinrich,kirk chi 106
              21 james,mike bos 106
              22 wesley,david nor 106
              23 arenas,gilbert was 106
              24 williams,jason mem 106
              25 arroyo,carlos uta 107
              26 marbury,stephon nyk 107
              27 boykins,earl den 108
              28 terry,jason atl 109
              29 stoudamire,damo por 109
              30 daniels,antonio sea 113
              31 nash,steve dal 113
              32 lue,tyronn orl 116



              Dean Oliver
              Author, Basketball on Paper
              http://www.basketballonpaper.com
              When basketball teams start playing Moneyball, this is the book
              they'll use!
            • dan_t_rosenbaum
              Measuring defense has never been something I have put a lot of effort into. Measuring the overall effectiveness of player is far more useful and at least with
              Message 6 of 16 , Aug 25, 2004
              • 0 Attachment
                Measuring defense has never been something I have put a lot of
                effort into. Measuring the overall effectiveness of player is far
                more useful and at least with my stuff a lot easier to measure. For
                me, measuring defense requires a good measure of overall
                effectiveness and a good measure of offensive productivity. I don't
                have a good measure of offensive productivity, so my defensive
                ratings are going to be much, much noisier than my overall ratings.

                And what by the way do we count as defense? What about if teams
                alter their lineups based upon matchups? If a certain player forces
                teams to often play worse offensive players in order to match up,
                should we count that player as a good defender, even if his
                individual defense is nothing to write home about? What about the
                player who runs his opponent ragged so that by the fourth quarter he
                no longer can hit a jump shot? It is really the total effectiveness
                of a player that we really care about. Defensive ratings might be
                useful sometimes as a check, but in a lot of cases they probably
                serve more as a distraction than anything else.

                --- In APBR_analysis@yahoogroups.com, "Mike G" <msg_53@h...> wrote:
                > --- In APBR_analysis@yahoogroups.com, igor eduardo küpfer
                > <edkupfer@r...> wrote:
                > > jimmy_purnell wrote:
                > > > I know everyone with a rating system is always working on
                > figuring out
                > > > how to judge a player's defense, ..
                >
                > Hey, I'm not.
                >
                > And glancing over these handful of short lists, I'm not encouraged
                > that there's much hope for it, at present. I see a number of
                players
                > at the top of some lists, who are at the bottom of others'.
                >
                > Since I only know defensive reputations by 2nd-sources and what
                > meager viewing I manage, I'm not literally Laughing ... but let's
                > face it boys, we've got a long way to go.
              • Dean Oliver
                ... These cross-correlations between offense and defense are pretty small factors. 90+% of defense is not affected by what you re talking about. There are
                Message 7 of 16 , Aug 25, 2004
                • 0 Attachment
                  --- In APBR_analysis@yahoogroups.com, "dan_t_rosenbaum"
                  <rosenbaum@u...> wrote:
                  > And what by the way do we count as defense? What about if teams
                  > alter their lineups based upon matchups? If a certain player forces
                  > teams to often play worse offensive players in order to match up,
                  > should we count that player as a good defender, even if his
                  > individual defense is nothing to write home about? What about the
                  > player who runs his opponent ragged so that by the fourth quarter he
                  > no longer can hit a jump shot? It is really the total effectiveness
                  > of a player that we really care about.

                  These cross-correlations between offense and defense are pretty small
                  factors. 90+% of defense is not affected by what you're talking
                  about. There are cross-matchups where Bruce Bowen matches up with
                  Kobe, but Kobe doesn't match up with Bruce. The fatigue affect is
                  pretty much averaged out over all games and what would be left is
                  toward the end of games, a small part of defense.

                  > Defensive ratings might be
                  > useful sometimes as a check, but in a lot of cases they probably
                  > serve more as a distraction than anything else.
                  >

                  Saying that overall ratings are more useful than defensive ratings is
                  like saying that the only thing that matters in evaluating the space
                  shuttle is whether it blows up or not. Yeah, survival of the people
                  and craft is most important, but if that's all you measure, you don't
                  really have a good sense for indicators of whether it's going to blow
                  up. You gotta go in and check the o-rings, the heat shield, the
                  steering, the flaps, all the stuff I don't know about. Nor is a
                  country defined purely by its GDP, though that is a pretty important
                  summary. If you want to understand how a country got to have such a
                  big economy, you better understand its other economics (of which you
                  know more than I).

                  Or if you have to pick between Dan Marino and Michael Vick based on an
                  overall rating being the same, don't you want more information?

                  Or if all that matters is who is highest rated by overall measures,
                  why don't we just throw KG, McGrady, AK, Duncan, and Shaq all on the
                  floor at the same time?

                  So, yeah, overall ratings are nice, but they lack the predictive power
                  of understanding all that goes into them, which includes defense.

                  DeanO

                  Dean Oliver
                  Author, Basketball on Paper
                  http://www.basketballonpaper.com
                  "Excellent writing. There are a lot of math guys who just rush from
                  the numbers to the conclusion. . .they'll tell you that Shaq is a real
                  good player but his team would win a couple more games a year if he
                  could hit a free throw. Dean is more than that; he's really
                  struggling to understand the actual problem, rather than the
                  statistical after-image of it. I learn a lot by reading him." Bill
                  James, author Baseball Abstract


                  > --- In APBR_analysis@yahoogroups.com, "Mike G" <msg_53@h...> wrote:
                  > > --- In APBR_analysis@yahoogroups.com, igor eduardo küpfer
                  > > <edkupfer@r...> wrote:
                  > > > jimmy_purnell wrote:
                  > > > > I know everyone with a rating system is always working on
                  > > figuring out
                  > > > > how to judge a player's defense, ..
                  > >
                  > > Hey, I'm not.
                  > >
                  > > And glancing over these handful of short lists, I'm not encouraged
                  > > that there's much hope for it, at present. I see a number of
                  > players
                  > > at the top of some lists, who are at the bottom of others'.
                  > >
                  > > Since I only know defensive reputations by 2nd-sources and what
                  > > meager viewing I manage, I'm not literally Laughing ... but let's
                  > > face it boys, we've got a long way to go.
                • bchaikin@aol.com
                  Measuring defense has never been something I have put a lot of effort into.  Measuring the overall effectiveness of player is far more useful and at least
                  Message 8 of 16 , Aug 25, 2004
                  • 0 Attachment
                    Measuring defense has never been something I have put a lot of effort into.  Measuring the overall effectiveness of player is far more useful and at least with my stuff a lot easier to measure.  For me, measuring defense requires a good measure of overall effectiveness and a good measure of offensive productivity.  I don't have a good measure of offensive productivity, so my defensive ratings are going to be much, much noisier than my overall ratings.

                    And what by the way do we count as defense?  What about if teams alter their lineups based upon matchups?  If a certain player forces teams to often play worse offensive players in order to match up, should we count that player as a good defender, even if his
                    individual defense is nothing to write home about?  What about the player who runs his opponent ragged so that by the fourth quarter he no longer can hit a jump shot?  It is really the total effectiveness of a player that we really care about.  Defensive ratings might be
                    useful sometimes as a check, but in a lot of cases they probably serve more as a distraction than anything else.

                    i could not disagree more. i think an effort to measure defensive FGM/FGA (trying to assign all FGM/FGA by players on offense to a specific defender) and forced turnovers (any turnover not a steal as being credited to a defender as a forced turnover) would go a long way in starting the process of rating players defensively. you can talk all you want about the intricacies of what a defender does, or of a switching defense, or of playing a zone, but the bottom line is that points are the currency by which games are won and lost and any measurement of preventing points from being scored is a positive move in the right direction of how to rate players defensively...

                    prior to 1973-74 the nba did not keep track of steals and blocked shots or rebounds broken down as being either offensive or defensive, nor turnovers prior to 1977-78. i couldn't imagine trying to rate players now without knowing ST, BS, TO, and rebounds broken down as either off or def, because through analysis we all know how important each of these categories are, and how wide ranging players abilities can be in each of these categories (can you imagine not knowing the impact of a player like andrei kirilenko who is heads and shoulders above most players because of his abilities to block shots and get steals?). i think any efforts to rate players defensively by keeping track of any knew defensive-minded stats is a sure step in the right direction...

                    sure any system of gathering this kind of information (all FGM/FGA assigned to defenders) will have its problems, but the data will be infinitely better than no data at all. you may think that it'd be tough to assign these to defenders, but i believe over the span of a long 82 game season (or 1230 games for 30 teams) most descrepencies would even out. again, look at off FGM/FGA. what does FG% really tell you? two players could both shoot 50%, but one may be from a player who gets that mostly on dunks and layups and short shots, and another on long range 2s. yet we've continued used the FGM/FGA = FG% as a standard for years without thinking twice about it, even though now we have better methods (effective FG%, scoring FG%) of rating player's shooting, not to mention FGM/FGA broken down by distance...

                    gather this kind of defensive information for 2-3 years and we'll wonder how we ever rated players defensively before. any data that can take the subjectivity out of player evaluation is a positive move....

                    It is really the total effectiveness of a player that we really care about.

                    there are multiple methods to do this now, but this is not all we want to know. winval and danval are great because they are a completely different approach to conventional stats that can be shown to produce extremely useful information, in particular information we did not have access to before, but if you want to determine what a team needs, you don't always want, nor can afford, the best or most totally effective player available. to be able to determine those attributes in a player that a team needs is key, and then whether you can go out and find a player with those attributes or get the player available with the closest set of attributes to what your team needs. rating players defensively is a huge step in that direction, and simply measuring def FGM/FGA and forced turnovers would start that process...

                    bob chaikin
                    bchaikin@...







                  • igor eduardo küpfer
                    ... I m pretty sure that you d see almost as much variation if we compared offensive lists. My system, for example, rarely spits out the same lists as yours.
                    Message 9 of 16 , Aug 25, 2004
                    • 0 Attachment
                      Mike G wrote:
                      > --- In APBR_analysis@yahoogroups.com, igor eduardo küpfer
                      > <edkupfer@r...> wrote:
                      >> jimmy_purnell wrote:
                      >>> I know everyone with a rating system is always working on
                      > figuring out
                      >>> how to judge a player's defense, ..
                      >
                      > Hey, I'm not.
                      >
                      > And glancing over these handful of short lists, I'm not encouraged
                      > that there's much hope for it, at present. I see a number of players
                      > at the top of some lists, who are at the bottom of others'.
                      >

                      I'm pretty sure that you'd see almost as much variation if we compared
                      offensive lists. My system, for example, rarely spits out the same lists as
                      yours.

                      > Since I only know defensive reputations by 2nd-sources and what
                      > meager viewing I manage, I'm not literally Laughing ... but let's
                      > face it boys, we've got a long way to go.
                      >

                      Well, sure. But that doesn't mean what we have is useless only because the
                      ratings don't correlate with each other all that well. It just means that
                      each system measures different things, that each system answers a related,
                      but different question.

                      I don't think anyone would doubt the value of measuring different aspects of
                      the same phenomenon, which is essentially what we're doing with our
                      defensive ratings (and offensive ratings as well). For us statheads, there
                      is no single such thing as "defense," only different manifestations of it in
                      the data.

                      --
                      ed
                    • dan_t_rosenbaum
                      I will try to respond to both you and Bob here. I think that I went way too far in what I said, and you two rightly put me back in check. ... small ... The
                      Message 10 of 16 , Aug 25, 2004
                      • 0 Attachment
                        I will try to respond to both you and Bob here. I think that I went
                        way too far in what I said, and you two rightly put me back in check.

                        > These cross-correlations between offense and defense are pretty
                        small
                        > factors. 90+% of defense is not affected by what you're talking
                        > about. There are cross-matchups where Bruce Bowen matches up with
                        > Kobe, but Kobe doesn't match up with Bruce. The fatigue affect is
                        > pretty much averaged out over all games and what would be left is
                        > toward the end of games, a small part of defense.

                        The point I was making here is that suppose Bowen plays less in
                        games where the other team does not have a good wing offensive
                        player. For those teams playing against the Spurs, having a good
                        wing offensive player helps the defense because they get to see more
                        of Bowen (presuming that Bowen is worse on offense than his
                        substitutes). Neither this nor the fatigue factor would really
                        average out over all games, but I agree that in most cases this
                        probably is not that big of a factor.

                        > Saying that overall ratings are more useful than defensive ratings
                        is
                        > like saying that the only thing that matters in evaluating the
                        space
                        > shuttle is whether it blows up or not.

                        What I said was a little too strong, but what you say seems to me an
                        overreaction. There is a huge difference between "are more useful"
                        and "the only thing that matters."

                        > Or if you have to pick between Dan Marino and Michael Vick based
                        on an
                        > overall rating being the same, don't you want more information?

                        Again, I think you are putting words in my mouth here. I don't
                        think I ever said that the overall rating is the only thing that
                        matters. You know as well as anyone that I have put a lot of work
                        in trying to figure out which box score statistics correlate well
                        with the adjusted plus/minus statistics. One of the big reasons for
                        this is to get some sense for why a given player is higher rated
                        than another.

                        It is just that with my stuff it is very tough to distinguish
                        offense from defense and so for my stuff coming up with defensive
                        ratings is always going to be more difficult than it is using PER
                        differences and the like.

                        For me it is easier to measure a different parameter - that part of
                        overall effectiveness not measured well by an efficiency index or by
                        PER or by some other offensively oriented index. I think that
                        parameter can be every bit as interesting as a defense parameter for
                        the very reasons that you cite above.

                        > Or if all that matters is who is highest rated by overall measures,
                        > why don't we just throw KG, McGrady, AK, Duncan, and Shaq all on
                        the
                        > floor at the same time?

                        Again, I am not sure where this point is going. I never made any
                        statement of this sort.

                        > So, yeah, overall ratings are nice, but they lack the predictive
                        power
                        > of understanding all that goes into them, which includes defense.

                        I agree with this point. I believe that overall ratings are the
                        most useful but can be uninformative without a solid understanding
                        of the components that add up to the overall rating.

                        Above I was suggesting a different way of breaking down the
                        components that with my stuff is easier to measure.

                        And on Bob's points that weren't covered above, I agree that
                        measuring FGA/FGM and forced turnovers can be very useful - just
                        like measuring those things on offense can be useful. But there are
                        some problems, as with any defensive measure (or offensive measure,
                        for that matter). For example, in the Finals because of the good
                        defense of Ben and Rasheed Wallace, Detroit felt comfortable single-
                        teaming Shaq. If the Wallaces had been lesser defenders, perhaps
                        Detroit would have double-teamed more, which might have led to a
                        lower field goal percentage and more turnovers for Shaq (but higher
                        shooting percentages and fewer turnovers for his teammates).

                        These things may not average out over a season because some teams
                        may be more quick to double team, thus protecting their poor
                        defenders from accumulating poor defensive statistics. On the other
                        hand, these things would be picked up by a measure of a player's
                        overall effectiveness while on the floor.

                        However, the same kind of thing can affect measures of offensive
                        production as well, so this is just a consideration and does not
                        invalidate the use of FGM/FGA and forced turnovers.
                      • Gabe Farkas
                        ... Bob, I want to juxtapose these two statements that you make (which are both very valid), and offer a hypothetical scenario. What if we all went back in a
                        Message 11 of 16 , Aug 26, 2004
                        • 0 Attachment
                          --- bchaikin@... wrote:

                          > prior to 1973-74 the nba did not keep track of
                          > steals and blocked shots or
                          > rebounds broken down as being either offensive or
                          > defensive, nor turnovers prior
                          > to 1977-78. i couldn't imagine trying to rate
                          > players now without knowing ST,
                          > BS, TO, and rebounds broken down as either off or
                          > def, because through
                          > analysis we all know how important each of these
                          > categories are, and how wide
                          > ranging players abilities can be in each of these
                          > categories (can you imagine not
                          > knowing the impact of a player like andrei kirilenko
                          > who is heads and shoulders
                          > above most players because of his abilities to block
                          > shots and get steals?). i
                          > think any efforts to rate players defensively by
                          > keeping track of any knew
                          > defensive-minded stats is a sure step in the right
                          > direction...


                          > gather this kind of defensive information for 2-3
                          > years and we'll wonder how
                          > we ever rated players defensively before. any data
                          > that can take the
                          > subjectivity out of player evaluation is a positive
                          > move....
                          >
                          > It is really the total effectiveness of a player
                          > that we really care about.
                          >


                          Bob, I want to juxtapose these two statements that you
                          make (which are both very valid), and offer a
                          hypothetical scenario.

                          What if we all went back in a time machine to 1955,
                          before they kept track of steals, blocks, turnovers,
                          and other "defensive" statistics. If you were a
                          statistician back then, and you had to devise a
                          defensive statistical tracking system, what would you
                          come up with? Do you think you would come up with the
                          same statistics that were deemed important? Maybe PPS
                          or eff% would become more important, and steals and
                          blocks would shrink to the background.

                          Once certain statistics were gathered for a few years,
                          first informally, then officially, they would become
                          more and more prominent.

                          Look at the Chinese league, where many individual
                          statistics (both offensive and defensive) are not kept
                          for many years now. It's basically a blank slate
                          where, barring foreign influence, they could start
                          emphasizing and tracking whatever statistics they felt
                          were most important.

                          I guess what I'm asking is this: is the current
                          statistical tracking system (emphasizing points,
                          rebounds and assists, and to a lesser extent blocks,
                          steals and turnovers) the only way that things could
                          have evolved?




                          __________________________________
                          Do you Yahoo!?
                          New and Improved Yahoo! Mail - 100MB free storage!
                          http://promotions.yahoo.com/new_mail
                        • bchaikin@aol.com
                          Bob, I want to juxtapose these two statements that you make (which are both very valid), and offer a hypothetical scenario. What if we all went back in a time
                          Message 12 of 16 , Aug 26, 2004
                          • 0 Attachment
                            Bob, I want to juxtapose these two statements that you make (which are both very valid), and offer a hypothetical scenario.

                            What if we all went back in a time machine to 1955, before they kept track of steals, blocks, turnovers, and other "defensive" statistics. If you were a statistician back then, and you had to devise a defensive statistical tracking system, what would you come up with? Do you think you would come up with the same statistics that were deemed important? Maybe PPS or eff% would become more important, and steals and
                            blocks would shrink to the background.

                            Once certain statistics were gathered for a few years, first informally, then officially, they would become more and more prominent.

                            Look at the Chinese league, where many individual statistics (both offensive and defensive) are not kept for many years now. It's basically a blank slate where, barring foreign influence, they could start emphasizing and tracking whatever statistics they felt
                            were most important.

                            I guess what I'm asking is this: is the current statistical tracking system (emphasizing points, rebounds and assists, and to a lesser extent blocks, steals and turnovers) the only way that things could have evolved?

                            i'd say pretty much yes...

                            no question from my end that the stats they do keep today are critical. anything that helps you evaluate (1) scoring, or (2) a change of possession, are imho critical. these are indeed kept for the most part. i have spoken several times with lee meade who was the original statistician for the old ABA (1967-68). when he got the job at that time the nba only kept track of G, MIN, FGM, FGA, FTM, FTA, REB, AST, and PF. he said for the orginal ABA boxscore/scoresheet he added stuff he used to keep track of when he covered games, such as ST, BS, TO, rebs as OREB and DREB, and team rebounds (read terry pluto's Loose Balls, meade talks about it there) and he said the process of getting everyone on board and doing the stats correctly took a couple of seasons...

                            i think those stats first tracked evolved simply because they were the easiest things to track and most important - anything to do with scoring because points are how you win or loss games, and fouls because after so many a player is disqualified. back in 1946-47 other than those the only other stat tracked was AST. in the early 1940s and late 1930s only FGM, FTM (no attempts, just those made), and PF were tracked.  MIN were not even  added until the 1951-52 season - i.e. at first if it wasn't essential they didn't bother with it...

                            now i think from the statistical analysis side offense is covered very well, so its the defensive side that is lacking in numbers, and those stats i mentioned (def FGM/FGA and forced TOs) would be a good starting point...

                            bob






                          • igor eduardo küpfer
                            Just to lay this thread to rest, I ve adapted DeanO s offensive ratings to the positional counterparts shown at 82games.com. Here s the top 20 defenders
                            Message 13 of 16 , Aug 30, 2004
                            • 0 Attachment
                              Just to lay this thread to rest, I've adapted DeanO's offensive ratings to
                              the positional counterparts shown at 82games.com. Here's the top 20
                              defenders according to this system:

                              PLAYER DRTG
                              1 Tim Duncan 87
                              2 Brendan Haywood 89
                              3 Tayshaun Prince 90
                              4 Jeff Foster 90
                              5 Jason Collins 90
                              6 Rasho Nesterovic 91
                              7 Mark Madsen 91
                              8 Kelvin Cato 92
                              9 Shaquille O'Neal 93
                              10 Reggie Miller 93
                              11 Yao Ming 94
                              12 Chris Bosh 94
                              13 Hedo Turkoglu 94
                              14 Jarron Collins 94
                              15 Vlade Divac 94
                              16 Erick Dampier 95
                              17 Shareef Abdur-Rahim 95
                              18 Samuel Dalembert 95
                              19 Corliss Williamson 95
                              20 Emanuel Ginobili 95
                            • John Hollinger
                              ... Not so fast. Meant to post this before, but FWIW here s my top and bottom 10 from last year. I used a high standard for playing time -- 1300 minutes. BEST
                              Message 14 of 16 , Aug 30, 2004
                              • 0 Attachment
                                --- In APBR_analysis@yahoogroups.com, igor eduardo küpfer
                                <edkupfer@r...> wrote:
                                > Just to lay this thread to rest,

                                Not so fast. Meant to post this before, but FWIW here's my top and
                                bottom 10 from last year. I used a high standard for playing time --
                                1300 minutes.

                                BEST
                                Ben Wallace 6.56
                                Theo Ratliff 6.29
                                Dikembe Mutombo 5.97
                                Kelvin Cato 5.96
                                Tayshaun Prince 5.39
                                Andrei Kirilenko 5.29
                                Manu Ginobili 5.22
                                Rasho Nesterovic 5.20
                                Marcus Camby 5.15
                                Kerry Kittles 5.14

                                WORST
                                Steve Nash -1.25
                                Tracy McGrady -1.07
                                Quentin Richardson -0.84
                                Juwon Howard -0.79
                                Stephon Marbury -0.48
                                Damon Stoudamire -0.34
                                Corey Maggette -0.33
                                Vitaly Potapenko -0.08
                                Antonio Daniels -0.06
                                Ray Allen -0.03
                              • McKibbin, Stuart
                                Laid to rest? Okay. Here s a few theoretical problems I have with Ed s approach, which on first blush sounds reasonable. 1. A portion of the offensive ratings
                                Message 15 of 16 , Aug 30, 2004
                                • 0 Attachment
                                  Laid to rest? Okay. Here's a few theoretical problems I have with Ed's
                                  approach, which on first blush sounds reasonable.

                                  1. A portion of the offensive ratings are earned in transition, say the
                                  first 5 or 6 seconds after a steal or defensive rebound. This favors the big
                                  men because it's the guards and SFs that do most of the damage offensively
                                  in those situations in oddman rushes. Moreover, many times because the big
                                  men are slow to get back defensively the guards can't release to run out to
                                  the shooters. That seems to unfairly penalize the guards for a big man's
                                  lack of speed, hustle or stamina.

                                  My proposal is that GETTING BACK is a team defense category and no
                                  individual blame should be assessed. Everything that happens in the first 5
                                  or 6 seconds after a steal or def reb (provided there isn't an interuption
                                  in play by timeout or something) should be taken out of the equation. With
                                  the 82 games database this could be done.

                                  As it is http://www.82games.com/clock1.htm shows that teams give up anywhere
                                  from 40-44% of offense in the first 10 seconds of a shot clock. This could
                                  be further refined by changing the shot clock segmentation (which would also
                                  eliminate such things as fouling immediately to get the ball back, and
                                  stupid loose ball fouls by attempted offensive rebounds that send the other
                                  team to the line easily) and subtracting out all off reb putbacks and
                                  putting them in a separate category. I would propose by removing those two
                                  things we've isolated half -court offense and defense---which is what we are
                                  really after, right?

                                  2. It doesn't take into account the variable offensive ratings of the
                                  positions. Centers having lower average Ortgs than SGs and PFs, so the
                                  centers as a group benefit again.

                                  3. It doesn't take into account the team's defensive scheme. For example,
                                  I'm told Boston did a lot of trapping of the PGs, so the opposing PGs PER is
                                  lower than other opposing positions. I don't think it's because Boston's PGs
                                  were exceptional defenders. Maybe they were, I just doubt it. Or more to the
                                  point: Reggie Miller shows up as superior to Ron Artest. That a 38-year old
                                  rates better than the alleged Defensive Player of the Year is curious. In
                                  any event, I wouldn't know how to account for it, so I'm just complaining
                                  without a concrete alternative.

                                  4. These rating systems are unfair to good perimeter defenders on poor
                                  defensive teams. For example, I'm sure if Bobby Simmons had half a chance
                                  he'd really contribute to a good defensive team. He makes a big difference
                                  to the Clippers God-awful defense, but because they still give up 108pts per
                                  100 possessions while he's in there he's not picked up by the systems. I
                                  think Dean made a similar point about Dumars in his book, that a single good
                                  perimeter defender wasn't much use on bad defensive teams. It would take
                                  critical mass of perimeter defenders (2, 3, 4?) to overcome their center's
                                  ineptitude.

                                  So the options might be that only centers can be judged across teams, and
                                  all we can do for the others is judge them relative to their own team; or we
                                  give recognition for large relative decreases in pts per 100 defense to
                                  perimeter players.

                                  Just some thoughts.
                                  Best Wishes,
                                  Stuart

                                  -----Original Message-----
                                  From: edkupfer@...
                                  Sent: Monday, August 30, 2004 6:10 PM
                                  To: APBR_analysis@yahoogroups.com; edkupfer@...
                                  Subject: Re: [APBR_analysis] Defensive ratings


                                  Just to lay this thread to rest, I've adapted DeanO's offensive ratings to
                                  the positional counterparts shown at 82games.com. Here's the top 20
                                  defenders according to this system:

                                  PLAYER DRTG
                                  1 Tim Duncan 87
                                  2 Brendan Haywood 89
                                  3 Tayshaun Prince 90
                                  4 Jeff Foster 90
                                  5 Jason Collins 90
                                  6 Rasho Nesterovic 91
                                  7 Mark Madsen 91
                                  8 Kelvin Cato 92
                                  9 Shaquille O'Neal 93
                                  10 Reggie Miller 93
                                  11 Yao Ming 94
                                  12 Chris Bosh 94
                                  13 Hedo Turkoglu 94
                                  14 Jarron Collins 94
                                  15 Vlade Divac 94
                                  16 Erick Dampier 95
                                  17 Shareef Abdur-Rahim 95
                                  18 Samuel Dalembert 95
                                  19 Corliss Williamson 95
                                  20 Emanuel Ginobili 95






                                  Yahoo! Groups Links
                                • igor eduardo küpfer
                                  ... Hell, you didn t even mention the two biggest problems, IMO: I had to estimate both the FGM and OR terms, both in a way which penalizes big men. Anyway...
                                  Message 16 of 16 , Aug 30, 2004
                                  • 0 Attachment
                                    McKibbin, Stuart wrote:
                                    > Laid to rest? Okay. Here's a few theoretical problems I have with Ed's
                                    > approach, which on first blush sounds reasonable.
                                    >

                                    Hell, you didn't even mention the two biggest problems, IMO: I had to
                                    estimate both the FGM and OR terms, both in a way which penalizes big men.
                                    Anyway...

                                    > 1. A portion of the offensive ratings are earned in transition, say
                                    > the first 5 or 6 seconds after a steal or defensive rebound. This
                                    > favors the big men because it's the guards and SFs that do most of
                                    > the damage offensively in those situations in oddman rushes.
                                    ...
                                    > My proposal is that GETTING BACK is a team defense category and no
                                    > individual blame should be assessed. Everything that happens in the
                                    > first 5 or 6 seconds after a steal or def reb (provided there isn't
                                    > an interuption in play by timeout or something) should be taken out
                                    > of the equation. With the 82 games database this could be done.
                                    >
                                    > As it is http://www.82games.com/clock1.htm shows that teams give up
                                    > anywhere from 40-44% of offense in the first 10 seconds of a shot
                                    > clock. This could be further refined by changing the shot clock
                                    > segmentation (which would also eliminate such things as fouling
                                    > immediately to get the ball back, and stupid loose ball fouls by
                                    > attempted offensive rebounds that send the other team to the line
                                    > easily) and subtracting out all off reb putbacks and putting them in
                                    > a separate category. I would propose by removing those two things
                                    > we've isolated half -court offense and defense---which is what we are
                                    > really after, right?
                                    >

                                    As you note above, shot clock time does not equal possession time. About 30%
                                    of all missed shots are offensively rebounded, and therefore not subject to
                                    the "getting back" defense (GBD). Looking through a sample of 143 games that
                                    I have here, only 15% of all possessions end in under 10 seconds.

                                    Another thing is that of the offense recorded with 10 seconds or less on
                                    the shot clock, a lot of it comes on fouls at the end of games where one
                                    team fouls quickly to stop the clock. This too is not subject to GBD. Three
                                    percent of all possessions end in under 3 seconds -- I assume that these end
                                    in a player being fouled or a set play coming off an inbound or putbacks.

                                    That said, your points are well taken. I will see what I can do about
                                    analyzing my data more closely following your advice above.

                                    > 2. It doesn't take into account the variable offensive ratings of the
                                    > positions. Centers having lower average Ortgs than SGs and PFs, so the
                                    > centers as a group benefit again.
                                    >

                                    As far as I know, the difference here isn't all that great. Why don't I just
                                    check that? <sound of ruffling paper and pencil scratching> Hmm. It turns
                                    out that centers actually have *higher* offensive ratings, although the
                                    difference is not significant:

                                    POS ORTG
                                    1 102.0
                                    2 102.0
                                    3 102.2
                                    4 103.1
                                    5 103.5


                                    > 3. It doesn't take into account the team's defensive scheme. For
                                    > example, I'm told Boston did a lot of trapping of the PGs, so the
                                    > opposing PGs PER is lower than other opposing positions. I don't
                                    > think it's because Boston's PGs were exceptional defenders. Maybe
                                    > they were, I just doubt it. Or more to the point: Reggie Miller shows
                                    > up as superior to Ron Artest. That a 38-year old rates better than
                                    > the alleged Defensive Player of the Year is curious. In any event, I
                                    > wouldn't know how to account for it, so I'm just complaining without
                                    > a concrete alternative.
                                    >

                                    A standout example of a lack of strength-of-opponent factor in my system.
                                    Abdul-Rahim shows up on my top-20 list also, presumably for the same reason:
                                    neither of these players are asked to guard tough assignments. DanR's system
                                    is the only one, I think, that begins to address this problem.

                                    > 4. These rating systems are unfair to good perimeter defenders on poor
                                    > defensive teams.
                                    >.... It would take critical mass of perimeter
                                    > defenders (2, 3, 4?) to overcome their center's ineptitude.
                                    >

                                    An interesting idea.

                                    > So the options might be that only centers can be judged across teams,
                                    > and all we can do for the others is judge them relative to their own
                                    > team; or we give recognition for large relative decreases in pts per
                                    > 100 defense to perimeter players.
                                    >

                                    I never look at my defensive stats in isolation. I usually mix them with a
                                    heavy component of team defense, just to provide a safety net.
                                  Your message has been successfully submitted and would be delivered to recipients shortly.