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

Re: [APBR_analysis] Assigning Credit to Players

Expand Messages
  • Michael K. Tamada
    Here are my gut feel answers: ... Passer 2/3, dunker 1/3. ... A not so gut feel answer: The identity of the shooter is what has me thinking; ideally we d want
    Message 1 of 10 , Mar 14, 2002
    • 0 Attachment
      Here are my gut feel answers:

      On Thu, 14 Mar 2002, HoopStudies wrote:

      >
      > I posed the following to a bunch of coaches. This group is a bit
      > tainted by my opinion, but I'll ask anyway...
      >
      > I am interested in obtaining some opinions on distributing
      > credit to players for different situations in basketball.
      >
      > For example
      >
      > 1. The point guard passes through the defense to a big man
      > underneath the basket for a dunk. What percentage of the credit for
      > the TEAM score should go to the passer vs. the scorer? Obviously,
      > one player gets an assist and one gets the field goal, but,
      > subjectively, should they split credit 50-50?

      Passer 2/3, dunker 1/3.

      >
      > 2. A point guard passes to an unguarded mediocre shooter at the
      > perimeter who makes a jump shot. How should credit be split here?

      A not so gut feel answer: The identity of the shooter is what has me
      thinking; ideally we'd want the point guard to pass to unguarded good
      shooters rather than unguarded mediocre shooters. It's a clearly superior
      situation, and is likely to be harder for the passer to achieve. (If it's
      Shawn Bradley 30 feet from the basket, getting the ball to him is no great
      achievement).

      All that suggests that the passer should get more credit when passing to
      the good shooter than the mediocre shooter.

      BUT this has a potentially nasty implication: given that the basket is
      worth only 2 points (gross; net value is even less than that because the
      team loses possession), if the passer gets more credit, then the shooter
      is getting LESS credit. So if Reggie Miller and Jeff Foster both hit open
      20 footers, should Reggie get less credit than Foster?

      Miller WILL make more of those baskets, so maybe it's okay to give him
      less credit -- but it better not be TOO much less. Otherwise, he might go
      5-10 one night on open 20 footers, and Foster would go 4-10 ... and Foster
      might be rated as having a better performance!

      So although taking the identity of the shooter into account seems like an
      attractive idea, it has some pitfalls.


      Back to the original question: ignoring the identity of the shooter, in
      such situations give 1/3 credit to the passer, 2/3 to the shooter.


      >
      > 3. On a pick and roll, the guard gets double-teamed and finds the
      > roll-guy near the basket for an open 5 ft shot.

      Guard gets 2/3 credit.

      >
      > 4. On a pick and roll, the guard gets double-teamed and finds the
      > roll-guy popping out for a 15 ft jump shot.

      Guard gets half credit.

      >
      > 5. On a pick and roll, the guard's man is picked off with no switch,
      > freeing him for an open 18 ft jump shot.

      Guard gets 3/4 credit, pick-setter gets 1/4. Hmm, but how much of this
      was not due to EITHER of them, but instead should be counted as a negative
      against the defenders who failed to do their job? Should a made basket
      always represent only positive things, to be apportioned among the
      offensive players, or should we instead deduct some credits from the
      defenders?

      In accounting terms, do we do double-entry bookkeeping, where a 2 point
      basket give 2 points (or so) of credit to be apportioned amongst the
      offensive players, and 2 points of blame to be apportioned amongst the
      defenders? Or would a more complex single-entry system be better, where
      maybe the offense was responsible for +1.2, and the defense was
      responsible for -.8.

      > 6. In situation 3, how do you split the credit/blame to the two
      > defenders who didn't shut down the pick and roll? Should the rest of
      > the team also receive blame? How much?

      See above, but in a double-entry system (or equivalenty, just focusing on
      the defenders) 2/3 blame to the double-teamer. Depends of course on how
      the play unfolded -- maybe the defensive guard needed help and
      should've switched, maybe the defensive double-teamer should not have
      gone to the double team -- and also on what the defense's tactics were
      supposed to be in such situations... who's responsibility was it?

      >
      > 7. Same question for splitting defensive credit for #4.

      50-50 defensive blame.

      >
      > 8. Same question for splitting defensive credit for #5.

      3/4 blame on the defender who failed to switch. Or maybe less, maybe the
      defensive guard should've fought through the pick, depends on the
      situation.

      >
      > 9. A guard misses a jump shot, which is rebounded and stuck back in
      > by a teammate. How much credit goes to the two players?

      Negative credits for the misser, he hurt the team (not by a lot, it's not
      nearly as bad as a turnover, but still, a negative event). The rebounder
      did two positive things: unassisted basket (2 points gross), plus an
      offensive rebound (worth maybe 2/3 of a point).

      >
      > 10. A guard misses a jump shot, which is rebounded by a teammate.
      > He passes back to the guard to restart the offense, which, after
      > several passes, scores. How much credit should go to the offensive
      > rebounder?

      About 2/3 of a point, but again in this example although the offense
      scored two points, some players should be charged with negative
      contribution (the shot misser) while the scorer and offensive rebounder
      get positive credits.


      --MKT


      > > I'm not looking for formal studies, but rather gut feel from people
      > on these, maybe with some explanation for the answers, maybe
      > without. I recognize that coaches often don't think about
      > distributing credit this way and that it, in some ways, goes against
      > the spirit of the roles players should have. This is an attempt to
      > identify whether people have thought about decisions about credit and
      > to collect results into a report for coaches to view.
      >
      > Dean Oliver
      > Journal of Basketball Studies
      > www.rawbw.com/~deano/index.html
      >
      >
      >
      >
      > To unsubscribe from this group, send an email to:
      > APBR_analysis-unsubscribe@yahoogroups.com
      >
      >
      >
      > Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
      >
      >
    • mikel_ind
      This raises the spectre of French judges, and tends to drive me to the side of those who like to say stats don t matter; only winning matters . But I will
      Message 2 of 10 , Mar 14, 2002
      • 0 Attachment
        This raises the spectre of French judges, and tends to drive me to
        the side of those who like to say "stats don't matter; only winning
        matters".

        But I will endeavor to make constructive comments.

        --- In APBR_analysis@y..., "HoopStudies" <deano@r...> wrote:
        >

        > 1. The point guard passes through the defense to a big man
        > underneath the basket for a dunk. What percentage of the credit for
        > the TEAM score should go to the passer vs. the scorer? Obviously,
        > one player gets an assist and one gets the field goal, but,
        > subjectively, should they split credit 50-50?

        The guard lobs a pass up high, the big man has to reach behind
        himself while leaping out of the gym: 120% to the big man, -20% to
        the passer.
        (Obviously, every case is different).

        > 2. A point guard passes to an unguarded mediocre shooter at the
        > perimeter who makes a jump shot. How should credit be split here?
        >
        Depends on the shot clock and a zillion other things. Who decides
        that guy is mediocre? Maybe that is his spot. Maybe he was knocking
        them down in practice. Maybe the guy needs a vote of confidence.
        Ignoring your wide-open shooter is poison to team chemistry, besides
        surrendering to the opposition at that matchup.

        > 3. On a pick and roll, the guard gets double-teamed and finds the
        > roll-guy near the basket for an open 5 ft shot.
        >
        > 4. On a pick and roll, the guard gets double-teamed and finds the
        > roll-guy popping out for a 15 ft jump shot.
        >
        > 5. On a pick and roll, the guard's man is picked off with no
        switch,
        > freeing him for an open 18 ft jump shot.

        My guts tell me success is its own reward, players get back to the
        guy that set them up, teams gel or don't gel, playing time is the
        result of good play.

        > 6. In situation 3, how do you split the credit/blame to the two
        > defenders who didn't shut down the pick and roll? Should the rest
        of
        > the team also receive blame? How much?

        How many ways could one split a hair? Further, why should one?

        ....
        > 9. A guard misses a jump shot, which is rebounded and stuck back in
        > by a teammate. How much credit goes to the two players?
        >
        > 10. A guard misses a jump shot, which is rebounded by a teammate.
        > He passes back to the guard to restart the offense, which, after
        > several passes, scores. How much credit should go to the offensive
        > rebounder?

        Now we are looking at the spectre of several "French judgments" on
        the same play.

        Sorry, I just don't get it.


        Mike Goodman
      • mikel_ind
        ... Any number from .1 to .9 might be appropriate. ... superior ... (If it s ... no great ... Passing to Bradley 30 feet away will yield no assists, so this
        Message 3 of 10 , Mar 14, 2002
        • 0 Attachment
          --- In APBR_analysis@y..., "Michael K. Tamada" <tamada@o...> wrote:
          >
          > Here are my gut feel answers:
          > >
          > > 1. The point guard passes through the defense to a big man
          > > underneath ...., should they split credit 50-50?
          >
          > Passer 2/3, dunker 1/3.

          Any number from .1 to .9 might be appropriate.


          >...ideally we'd want the point guard to pass to unguarded good
          > shooters rather than unguarded mediocre shooters. It's a clearly
          superior
          > situation, and is likely to be harder for the passer to achieve.
          (If it's
          > Shawn Bradley 30 feet from the basket, getting the ball to him is
          no great
          > achievement).

          Passing to Bradley 30 feet away will yield no assists, so this credit
          is built-in.
          (Unless the clock is about to expire, and Bradley is the only guy
          open. In this case, not-passing-to-Bradley should get you yanked
          from the game.)

          > All that suggests that the passer should get more credit when
          passing to
          > the good shooter than the mediocre shooter.
          >.... So if Reggie Miller and Jeff Foster both hit open
          > 20 footers, should Reggie get less credit than Foster?
          >
          > Miller WILL make more of those baskets, so maybe it's okay to give
          him
          > less credit -- but it better not be TOO much less. Otherwise, he
          might go
          > 5-10 one night on open 20 footers, and Foster would go 4-10 ... and
          Foster
          > might be rated as having a better performance!

          Indeed, the longer you think about it, the more unimaginable the idea
          becomes. (Shouldn't Foster really be banging the boards, rather than
          hanging out at 20 feet?)

          > > 3. On a pick and roll, the guard gets double-teamed and finds the
          > > roll-guy near the basket for an open 5 ft shot.
          >
          > Guard gets 2/3 credit.
          >
          > >
          > > 4. On a pick and roll, the guard gets double-teamed and finds the
          > > roll-guy popping out for a 15 ft jump shot.
          >
          > Guard gets half credit.

          We could have concentric arcs that assign % credit in this way.
          While we're at it, install sensors under the floor, in the ball, and
          in the shoes of the players, so it is all tabulated electronically.

          This is reminiscent of some suggested extra arcs that give a team
          only 1 point inside 5 feet (or 10, or whatever), and 4 points beyond
          40 feet, etc.


          > > 5. On a pick and roll, the guard's man is picked off with no
          switch,
          > > freeing him for an open 18 ft jump shot.
          >
          > Guard gets 3/4 credit, pick-setter gets 1/4. Hmm, but how much of
          this
          > was not due to EITHER of them,

          Indeed, there are players setting picks for the pickers, acting as
          decoys, combining these actions with positioning for rebounding... My
          guess is, there are infinite possibilities you could consider.

          > but instead should be counted as a negative
          > against the defenders who failed to do their job? Should a made
          basket
          > always represent only positive things, to be apportioned among the
          > offensive players, or should we instead deduct some credits from the
          > defenders?

          Another odious can of worms: Negative statistics. Perhaps
          competition at any level should have some accounting system that adds
          up to zero. Every success is the result of some mistake.

          At a game, one is sometimes surrounded by fans who only see the
          negative aspects of the game, and a "positive" comment is "About time
          he hit that shot", or some such.

          I see the game as 90-99% beauty and excellence. Sometimes I can't
          remember the last time someone misplayed anything.

          If it were an orchestra, of course more than 90% of the notes have to
          be essentially perfect. But there is no D in music.

          What the heck were we talking about here?
        • alleyoop2
          Fascinating topic Dean. I put in my $0.02 below each item. ... Normally, I think the shooter should get most of the credit, since he has to both get open and
          Message 4 of 10 , Mar 14, 2002
          • 0 Attachment
            Fascinating topic Dean. I put in my $0.02 below each item.

            --- In APBR_analysis@y..., "HoopStudies" <deano@r...> wrote:
            >
            > I posed the following to a bunch of coaches. This group is a bit
            > tainted by my opinion, but I'll ask anyway...
            >
            > I am interested in obtaining some opinions on distributing
            > credit to players for different situations in basketball.
            >
            > For example
            >
            > 1. The point guard passes through the defense to a big man
            > underneath the basket for a dunk. What percentage of the credit for
            > the TEAM score should go to the passer vs. the scorer? Obviously,
            > one player gets an assist and one gets the field goal, but,
            > subjectively, should they split credit 50-50?

            Normally, I think the shooter should get most of the credit, since he
            has to both get open and make the shot. In this case, since the pass
            led directly to a dunk, I think it should be split more evenly. 50-
            50, or 60-40 for the shooter at worst.

            >
            > 2. A point guard passes to an unguarded mediocre shooter at the
            > perimeter who makes a jump shot. How should credit be split here?
            >

            Passing to a mediocre shooter is its own reward. He'll make a lower
            percentage so the guard will get fewer assist "credits." I would
            consider this a fairly normal assist and give 2/3 to the shooter, 1/3
            to the passer.

            > 3. On a pick and roll, the guard gets double-teamed and finds the
            > roll-guy near the basket for an open 5 ft shot.
            >
            > 4. On a pick and roll, the guard gets double-teamed and finds the
            > roll-guy popping out for a 15 ft jump shot.
            >

            Very little difference between these two IMHO, since a lot depends on
            where the roll guy decides to go. I'd go 2/3 for the shooter on each.

            > 5. On a pick and roll, the guard's man is picked off with no
            switch,
            > freeing him for an open 18 ft jump shot.
            >

            Interesting - a pick leading directly to a basket. Theoretically that
            should be worth as much as an assist. I would give as much as 1/3 of
            the credit to the screener.


            > 6. In situation 3, how do you split the credit/blame to the two
            > defenders who didn't shut down the pick and roll? Should the rest
            of
            > the team also receive blame? How much?
            >
            > 7. Same question for splitting defensive credit for #4.
            >
            > 8. Same question for splitting defensive credit for #5.
            >

            I think, if you're doing this, you have to assign -2 points to your
            defense somehow on this play. On the pick and roll it should probably
            be split among the two guys, unless one made a particularly poor play.

            > 9. A guard misses a jump shot, which is rebounded and stuck back in
            > by a teammate. How much credit goes to the two players?
            >

            Negative credit for the guard, unless you think he was passing it off
            the rim. Two credits for the rebounder - 1) regaining possession, and
            2) making the basket.

            > 10. A guard misses a jump shot, which is rebounded by a teammate.
            > He passes back to the guard to restart the offense, which, after
            > several passes, scores. How much credit should go to the offensive
            > rebounder?

            This is why I separate the regaining possession from making the
            basket above. Same negative credit to the guard.

            >
            > I'm not looking for formal studies, but rather gut feel from people
            > on these, maybe with some explanation for the answers, maybe
            > without. I recognize that coaches often don't think about
            > distributing credit this way and that it, in some ways, goes
            against
            > the spirit of the roles players should have. This is an attempt to
            > identify whether people have thought about decisions about credit
            and
            > to collect results into a report for coaches to view.
            >
            > Dean Oliver
            > Journal of Basketball Studies
            > www.rawbw.com/~deano/index.html
          • HoopStudies
            I won t say much yet on this, but I do want to introduce one aspect with regard to ... in ... off ... and ... One of the things I remember hearing about
            Message 5 of 10 , Mar 14, 2002
            • 0 Attachment
              I won't say much yet on this, but I do want to introduce one aspect
              with regard to

              > > 9. A guard misses a jump shot, which is rebounded and stuck back
              in
              > > by a teammate. How much credit goes to the two players?
              > >
              >
              > Negative credit for the guard, unless you think he was passing it
              off
              > the rim. Two credits for the rebounder - 1) regaining possession,
              and
              > 2) making the basket.

              One of the things I remember hearing about Dominique Wilkins and have
              heard about Allen Iverson is that their shots are more often
              rebounded by their own team because their teammates know that they
              are going to shoot the ball. We already have put forth evidence (I
              think, I know I did the work) showing that Iverson's presence does
              help the Philly offense even though the boy sometimes can't hit the
              side of a barn with his shot.

              Just something to consider.

              DeanO
            • McKibbin, Stuart
              I m not looking for formal studies, but rather gut feel from people on these, maybe with some explanation for the answers, maybe without. 1. The point guard
              Message 6 of 10 , Mar 14, 2002
              • 0 Attachment
                I'm not looking for formal studies, but rather gut feel from people
                on these, maybe with some explanation for the answers, maybe
                without.

                1. The point guard passes through the defense to a big man
                underneath the basket for a dunk. What percentage of the credit for
                the TEAM score should go to the passer vs. the scorer? Obviously,
                one player gets an assist and one gets the field goal, but,
                subjectively, should they split credit 50-50?

                If the pass is truly through the defense, that is, if the point guard had to see the opening and thread the ball at high velocity past defenders to the big man, (who then has to be able to handle the pass and not drop it) who dunks the ball without the defense being able to react then I'd give ALL of the credit to the PG. The PG made the play.

                2. A point guard passes to an unguarded mediocre shooter at the
                perimeter who makes a jump shot. How should credit be split here?

                NO credit should ever be given for making a basic pass. However, if the point guard penetrates and draws Mr. Mediocre's defender to him, and subsequently passes the ball to Mr. Mediocre who hits his wide open shot then PG should get ALL of the credit. Again, the PG made the play.


                9. A guard misses a jump shot, which is rebounded and stuck back in
                by a teammate. How much credit goes to the two players?

                I agree with others---what the hell did the guard do to earn credit? It sounds like his offensive rebounding teammate just bailed him out.

                10. A guard misses a jump shot, which is rebounded by a teammate.
                He passes back to the guard to restart the offense, which, after
                several passes, scores. How much credit should go to the offensive
                rebounder?

                No more than should be given a guy for grabbing a defensive rebound, or stealing the ball, or taking the ball out of the net. Or for that matter, to the opponent for kicking the ball out of bounds. In other words if the offensive rebound doesn't lead DIRECTLY to the basket, who cares how you got the ball?
              • dlirag
                ... back ... have ... A shot can be seen as a pass that results in one of the following things happening: 2 or 3 points are immediately scored (45% chance),
                Message 7 of 10 , Mar 29, 2002
                • 0 Attachment
                  --- In APBR_analysis@y..., "HoopStudies" <deano@r...> wrote:
                  > > > 9. A guard misses a jump shot, which is rebounded and stuck
                  back
                  > in
                  > > > by a teammate. How much credit goes to the two players?
                  > > >
                  > >
                  > > Negative credit for the guard, unless you think he was passing it
                  > off
                  > > the rim. Two credits for the rebounder - 1) regaining possession,
                  > and
                  > > 2) making the basket.
                  >
                  > One of the things I remember hearing about Dominique Wilkins and
                  have
                  > heard about Allen Iverson is that their shots are more often
                  > rebounded by their own team because their teammates know that they
                  > are going to shoot the ball. We already have put forth evidence (I
                  > think, I know I did the work) showing that Iverson's presence does
                  > help the Philly offense even though the boy sometimes can't hit the
                  > side of a barn with his shot.

                  A shot can be seen as a pass that results in one of the following
                  things happening: 2 or 3 points are immediately scored (45% chance),
                  the ball ends up in a teammate's hands (16% chance), or the ball goes
                  to the opposing team (39% chance).

                  Seen this way, a missed jump shot that's put back in by a teammate
                  could be recast as a bad pass to a teammate who manages to score off
                  this pass. Maybe one should allocate credit with this revised view of
                  missed shots in mind, but I'm not yet sure.
                • HoopStudies
                  ... they ... (I ... does ... the ... chance), ... goes ... This gets at what Gary Skoog did in one of James abstracts, where they assessed the expected runs
                  Message 8 of 10 , Mar 29, 2002
                  • 0 Attachment
                    --- In APBR_analysis@y..., "dlirag" <dlirag@h...> wrote:
                    > >
                    > > One of the things I remember hearing about Dominique Wilkins and
                    > have
                    > > heard about Allen Iverson is that their shots are more often
                    > > rebounded by their own team because their teammates know that
                    they
                    > > are going to shoot the ball. We already have put forth evidence
                    (I
                    > > think, I know I did the work) showing that Iverson's presence
                    does
                    > > help the Philly offense even though the boy sometimes can't hit
                    the
                    > > side of a barn with his shot.
                    >
                    > A shot can be seen as a pass that results in one of the following
                    > things happening: 2 or 3 points are immediately scored (45%
                    chance),
                    > the ball ends up in a teammate's hands (16% chance), or the ball
                    goes
                    > to the opposing team (39% chance).

                    This gets at what Gary Skoog did in one of James' abstracts, where
                    they assessed the expected runs for every given state and assigned
                    credit/blame for transitions between those states.

                    At the point before a shot, every possession is worth about 1 pt.
                    After the shot, it could go to 2 pts with a made shot and end of
                    possession, 3 pts with a made shot and end of possession, rare cases
                    of >3 pts with or without continued possession, 0 pts with the
                    probability of an offensive rebound (expected value of 1 pt after
                    OR), 0 pts with probability of opponent getting the ball (expected
                    value of 0 after DR). In the 0 pts with offensive rebound, you
                    diminish the value of the missed shot by about 2/3 (assuming 1/3
                    chance of OR), so, using relative to the mean credit (as Skoog did),
                    the credit is -2/3 to the shooter who missed (bring expected points
                    from 1 to 1/3), 2/3 to the offensive rebounder (who brought it back
                    to 1), then 1 to the scorer (for going from 1 to 2). If this
                    methodology is carried out in detail, you could modify the shooter-
                    specific OR%, so that rather than diminishing Iverson's misses by
                    2/3, you diminish them by 1/2 if you think that half of his misses
                    are rebounded by the offense.

                    Though in this case the rebounder and scorer are the same, note that
                    there is more relative credit to the scorer (+1) than to the
                    rebounder (+2/3). This may be in conflict with what the offensive
                    rebound studies we did earlier here said, but I'd need to check.

                    Dean Oliver
                  • Michael K. Tamada
                    ... [...] ... [...] This is how my model of basketball works. It can be used to generate means and standard deviations, but can even more directly be used to
                    Message 9 of 10 , Mar 29, 2002
                    • 0 Attachment
                      On Fri, 29 Mar 2002, HoopStudies wrote:

                      > --- In APBR_analysis@y..., "dlirag" <dlirag@h...> wrote:

                      [...]

                      > > A shot can be seen as a pass that results in one of the following
                      > > things happening: 2 or 3 points are immediately scored (45%
                      > chance),
                      > > the ball ends up in a teammate's hands (16% chance), or the ball
                      > goes
                      > > to the opposing team (39% chance).
                      >
                      > This gets at what Gary Skoog did in one of James' abstracts, where
                      > they assessed the expected runs for every given state and assigned
                      > credit/blame for transitions between those states.
                      >
                      > At the point before a shot, every possession is worth about 1 pt.
                      > After the shot, it could go to 2 pts with a made shot and end of
                      > possession, 3 pts with a made shot and end of possession, rare cases

                      [...]

                      This is how my model of basketball works. It can be used to generate
                      means and standard deviations, but can even more directly be used to
                      estimate win probabilities. The trouble is that the models make it easy
                      and tempting to assign all the credit for an action to a player, e.g. the
                      offensive rebounder gets the +2/3 that you describe, when some of the
                      credit perhaps should go to teammates who blocked out (unlikely however in
                      an offensive rebounding situation) and maybe some of the blame should go
                      to the defensive rebounder who failed to block the guy out. So this model
                      makes assignment of credit to players easy to do -- but not necessarily
                      truly accurate.

                      > Though in this case the rebounder and scorer are the same, note that
                      > there is more relative credit to the scorer (+1) than to the
                      > rebounder (+2/3). This may be in conflict with what the offensive
                      > rebound studies we did earlier here said, but I'd need to check.

                      No, scoring a field goal is clearly more valuable than grabbing an
                      offensive rebound, even when the loss of possession is taken into account.
                      Missing a field goal is the act that is approximately equal to an
                      offensive rebound.


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