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

Re: NBA Stats site

Expand Messages
  • Dean Oliver
    ... In http://www.rawbw.com/~deano/articles/iwldef.html I post 3 versions of individual win-loss records. The version that Roland posts -- how many games that
    Message 1 of 5 , Oct 10, 2003
    • 0 Attachment
      --- In APBR_analysis@yahoogroups.com, igor eduardo küpfer
      >
      > Perhaps the Pythag would be informative here. Using Tim Duncan as an
      > example:
      >
      > OffPts DefPts Min Win% W-L
      > OnCourt 97.0 88.9 3151 75.6% 50-16
      > OffCourt 91.7 97.9 775 29.9% 5-11
      >
      > 50-16 is a number people can easily understand, while points scored and
      > allowed appeals only to us statheads :-)
      >
      > BTW DeanO: how do your individual win numbers match up to these?
      >

      In

      http://www.rawbw.com/~deano/articles/iwldef.html

      I post 3 versions of individual win-loss records. The version that
      Roland posts -- how many games that person's team outscores the
      opponent when the player is on the court -- is most like the first
      listed there, I believe because players on good teams generally look
      pretty good. Roland's is better than what my first one does. Method
      2 shown on my page is one of my favorite tools because it indicates
      how well players do their roles. For some players, I think it is
      similar to Roland's, especially the better players (Duncan 58-23 by
      Roland and 62-19 by me). Method 3 of mine is applying Pythagorean
      methods to individual offensive and defensive ratings, which is more
      answering the question you raise. It looks, however, to estimate how
      many games a player contributes to, so how many wins can they add.
      Duncan comes out at 15.9-1.0 (.941), so it's not comparable in
      magnitude.

      I think it's key to note that all these interpretations of an
      individual's win-loss record are fair. The one Ed coins is
      interesting. The one that Roland has should be similar to Ed's
      (theoretically, they should be very comparable), but I haven't checked.

      >
      > I have a few other suggestions. I'll send them off list.

      That data set could lead to a flood of ideas on what to do with it.
      How many years of pbp's do you have?

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