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

Effect of Losing Man on Defense

Expand Messages
  • Dean Oliver
    As mentioned a couple messages ago, I have started to look at the effect of losing certain players on their team defense. Knowing that defense is the toughest
    Message 1 of 6 , Mar 25 12:21 PM
    • 0 Attachment
      As mentioned a couple messages ago, I have started to look at the effect
      of losing certain players on their team defense. Knowing that defense is
      the toughest thing to quantify, I look for opportunities to do so when
      players miss games or get traded. Specifically, I looked at some of Doug
      Steele's top defensive players

      http://shell.rmi.net/~doug/00-01DefensiveTendex.html

      and tried to see whether their absence has been reflected in their teams'
      defensive rating.

      Here is what I found (including offensive numbers)

      With Player Without Player
      Off. Rtg. Def. Rtg. Off. Rtg. Def. Rtg.
      Webber 104.0 97.4 105.2 100.0
      Mutombo 95.6 98.7 99.7 109.3
      Iverson 102.2 96.1 99.4 99.8
      Marion 98.1 94.8 91.5 95.6
      Camby 100.5 96.8 96.6 94.8
      Olajuwo 103.5 100.8 107.2 106.0
      EJones 99.3 95.7 100.0 100.5

      (Iverson is not listed as one of the top D players, but I personally have
      him fairly high. You generally have to be interested in his influence.)

      Of these players, the ones whose teams defense improved significantly
      (at the 5% level) when they were in the game are only Mutombo, Olajuwon
      and Eddie Jones. The Knicks actually got a little better without Camby,
      which raises the always important issue in doing this kind of study -- it
      reflects not only the players of interest, but the players who fill in for
      them. In the case of Camby, the players getting the most additional time
      when he's out are Othella Harrington, Luc Longley, and Kurt Thomas --
      pretty good defensive players. In the case of Eddie Jones, the players
      getting more time are Cedric Ceballos and Eddie House -- two pretty bad
      defensive players (someone once called Ceballos the worst defensive player
      in the league).

      Ultimately, I'd really like to predict these effects better. Offensively,
      I can do OK. The Iverson situation can be tough because he is such a
      large part of the offense and gets double teamed an incredible amount
      (makes his teammates better a disproportionate amount), but I actually do
      OK there, too. But defensively, it is hard. The most basic things I
      would like to do -- replace Mutombo's stops by other players' -- show
      little effect of him missing a game on defense. I definitely have Mutombo
      with the best Def Rtg of anyone on his team -- 5 to 10 pts better -- but
      that isn't enough to explain an 11 point drop off. There is some weird
      math necessary to quantify defense -- I'm getting more convinced.

      More study necessary.

      Dean Oliver
      Journal of Basketball Studies
      www.tsoft.com/~deano/index.html
      deano@...
    • aussievamp2
      ... effect ... defense is ... when ... of Doug ... This link doesn t seem to work, is it reproduced anywhere? Thanks, Richard
      Message 2 of 6 , Jul 4, 2002
      • 0 Attachment
        --- In APBR_analysis@y..., Dean Oliver <deano@t...> wrote:
        >
        > As mentioned a couple messages ago, I have started to look at the
        effect
        > of losing certain players on their team defense. Knowing that
        defense is
        > the toughest thing to quantify, I look for opportunities to do so
        when
        > players miss games or get traded. Specifically, I looked at some
        of Doug
        > Steele's top defensive players
        >
        > http://shell.rmi.net/~doug/00-01DefensiveTendex.html


        This link doesn't seem to work, is it reproduced anywhere?

        Thanks,

        Richard
      • HoopStudies
        ... http://home.rmi.net/~doug/00-01DefensiveTendex.html DeanO
        Message 3 of 6 , Jul 8, 2002
        • 0 Attachment
          --- In APBR_analysis@y..., "aussievamp2" <rnmscott@b...> wrote:
          > --- In APBR_analysis@y..., Dean Oliver <deano@t...> wrote:
          > >
          > > As mentioned a couple messages ago, I have started to look at the
          > effect
          > > of losing certain players on their team defense. Knowing that
          > defense is
          > > the toughest thing to quantify, I look for opportunities to do so
          > when
          > > players miss games or get traded. Specifically, I looked at some
          > of Doug
          > > Steele's top defensive players
          > >
          > > http://shell.rmi.net/~doug/00-01DefensiveTendex.html
          >
          >
          > This link doesn't seem to work, is it reproduced anywhere?

          http://home.rmi.net/~doug/00-01DefensiveTendex.html

          DeanO
        • Richard Scott
          Thanks Dean. ... From: HoopStudies [mailto:deano@rawbw.com] Sent: July 08, 2002 6:52 PM To: APBR_analysis@yahoogroups.com Subject: [APBR_analysis] Re: Effect
          Message 4 of 6 , Jul 8, 2002
          • 0 Attachment
            Thanks Dean.

            -----Original Message-----
            From: HoopStudies [mailto:deano@...]
            Sent: July 08, 2002 6:52 PM
            To: APBR_analysis@yahoogroups.com
            Subject: [APBR_analysis] Re: Effect of Losing Man on Defense
            http://home.rmi.net/~doug/00-01DefensiveTendex.html

            DeanO
          • airiverson01
            ... the formula didn t seem to show defense much better than tndex or many other systems do, am i missing something? i just wonder why it is called defensive
            Message 5 of 6 , Jul 9, 2002
            • 0 Attachment
              --- In APBR_analysis@y..., "Richard Scott" <rnmscott@b...> wrote:
              > Thanks Dean.
              >
              > -----Original Message-----
              > From: HoopStudies [mailto:deano@r...]
              > Sent: July 08, 2002 6:52 PM
              > To: APBR_analysis@y...
              > Subject: [APBR_analysis] Re: Effect of Losing Man on Defense
              > http://home.rmi.net/~doug/00-01DefensiveTendex.html
              >
              > DeanO

              the formula didn't seem to show defense much better than tndex or
              many other systems do, am i missing something? i just wonder why it
              is called defensive rating or whatever you call it when it shows
              offense a lot too.

              thanks.
            • HoopStudies
              ... it ... Just to clarify -- Doug calculates his tendex rating on all players but he also matches up players offensively and defensively. What he then does
              Message 6 of 6 , Jul 9, 2002
              • 0 Attachment
                --- In APBR_analysis@y..., "airiverson01" <airiverson01@y...> wrote:
                > the formula didn't seem to show defense much better than tndex or
                > many other systems do, am i missing something? i just wonder why
                it
                > is called defensive rating or whatever you call it when it shows
                > offense a lot too.

                Just to clarify -- Doug calculates his tendex rating on all players
                but he also matches up players offensively and defensively. What he
                then does is evaluate how well defenders lower their opponent's
                tendex rating relative to their averages. This is to show that
                players who are good defenders, like Bruce Bowen, who also get
                matched up against the tough shooters, get credit for bringing down
                good offensive players to only decent offensive players. Doug told
                me that the tendex rating that he calculates for this actually does
                not include steals, blocks, or defensive rebounds after I suggested
                that reduction in those stats really isn't the doing of a defender.

                DeanO
              Your message has been successfully submitted and would be delivered to recipients shortly.