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

Dick Cramer's baseball work -- Evaluating history vs today

Expand Messages
  • Dean Oliver
    I m finally re-reading Hidden Game of Baseball after all these years and finding Chapter 7 particularly relevant to the conversation we had and put on hold
    Message 1 of 8 , Jun 23, 2003
    • 0 Attachment
      I'm finally re-reading Hidden Game of Baseball after all these years
      and finding Chapter 7 particularly relevant to the conversation we
      had and put on hold some time ago -- about whether the league has
      evolved and whether the old guys would do better or worse now.

      In the Hidden Game, they point to work by Dick Cramer in 1980 where
      he looked at how existing players saw their batting averages change
      through time. Through analysis that isn't presented in the Hidden
      Game (but in Cramer's 1980 article that I need to get), Cramer shows
      very clearly that old guys would do worse in modern baseball. Ty
      Cobb's lifetime batting average would be well below .367
      (around .330). Rod Carew would be about the greatest average
      hitter. This caught my eye for another reason, which is that I did
      quick work over in APBR looking at how FG%'s for long time players
      trended through the '80's and '90's. That showed that most players
      did have their FG%'s decline in the '90's, even before they "got
      old." I'd be curious to look more at players from the '70's to
      the '80's, where league FG%'s were going up -- were longtime guys
      seeing their FG% also go up?

      Cramer did his work more rigorously than I did in my 10 minute study,
      normalizing relative to league averages, etc. Does anyone have that
      work from the Baseball Research Journal? It may be useful here.

      DeanO
    • bricks299
      Dean I haven t looked at the numbers; but I think percentages decreased as the 3-point FGA became more common. Have you compared their 2-point FG%? ... years
      Message 2 of 8 , Jun 23, 2003
      • 0 Attachment
        Dean

        I haven't looked at the numbers; but I think percentages decreased as
        the 3-point FGA became more common. Have you compared their 2-point
        FG%?


        --- In APBR_analysis@yahoogroups.com, "Dean Oliver" <deano@r...>
        wrote:
        >
        > I'm finally re-reading Hidden Game of Baseball after all these
        years
        > and finding Chapter 7 particularly relevant to the conversation we
        > had and put on hold some time ago -- about whether the league has
        > evolved and whether the old guys would do better or worse now.
        >
        > In the Hidden Game, they point to work by Dick Cramer in 1980 where
        > he looked at how existing players saw their batting averages change
        > through time. Through analysis that isn't presented in the Hidden
        > Game (but in Cramer's 1980 article that I need to get), Cramer
        shows
        > very clearly that old guys would do worse in modern baseball. Ty
        > Cobb's lifetime batting average would be well below .367
        > (around .330). Rod Carew would be about the greatest average
        > hitter. This caught my eye for another reason, which is that I did
        > quick work over in APBR looking at how FG%'s for long time players
        > trended through the '80's and '90's. That showed that most players
        > did have their FG%'s decline in the '90's, even before they "got
        > old." I'd be curious to look more at players from the '70's to
        > the '80's, where league FG%'s were going up -- were longtime guys
        > seeing their FG% also go up?
        >
        > Cramer did his work more rigorously than I did in my 10 minute
        study,
        > normalizing relative to league averages, etc. Does anyone have
        that
        > work from the Baseball Research Journal? It may be useful here.
        >
        > DeanO
      • harlanzo
        ... years ... where ... change ... shows ... did ... players ... Dean, can you explain this? This assumes, I am guessing, that guady numbers that Cobb put up
        Message 3 of 8 , Jun 23, 2003
        • 0 Attachment
          --- In APBR_analysis@yahoogroups.com, "Dean Oliver" <deano@r...>
          wrote:
          >
          > I'm finally re-reading Hidden Game of Baseball after all these
          years
          > and finding Chapter 7 particularly relevant to the conversation we
          > had and put on hold some time ago -- about whether the league has
          > evolved and whether the old guys would do better or worse now.
          >
          > In the Hidden Game, they point to work by Dick Cramer in 1980
          where
          > he looked at how existing players saw their batting averages
          change
          > through time. Through analysis that isn't presented in the Hidden
          > Game (but in Cramer's 1980 article that I need to get), Cramer
          shows
          > very clearly that old guys would do worse in modern baseball. Ty
          > Cobb's lifetime batting average would be well below .367
          > (around .330). Rod Carew would be about the greatest average
          > hitter. This caught my eye for another reason, which is that I
          did
          > quick work over in APBR looking at how FG%'s for long time players
          > trended through the '80's and '90's. That showed that most
          players
          > did have their FG%'s decline in the '90's, even before they "got
          > old." I'd be curious to look more at players from the '70's to
          > the '80's, where league FG%'s were going up -- were longtime guys
          > seeing their FG% also go up?

          Dean, can you explain this? This assumes, I am guessing, that guady
          numbers that Cobb put up were at least partly a result of inferior
          competition. I don't necessarily disagree but remember Cramer wrote
          this in 1980 when batting averages were way down from the Cobb era.
          Stats went up later. How do we separate cyclical stat trends (such
          as shooting %) with honest to goodness improvement. I assume WIlt
          couldn't socre 50 ppg today but how do we adjust AFTER game pace is
          accounted for?
        • Mike G
          ... This I would love to see. I already know it isn t going to convince me. Peripheral factors notwithstanding (i.e., rigors of travel, equipment,
          Message 4 of 8 , Jun 24, 2003
          • 0 Attachment
            --- In APBR_analysis@yahoogroups.com, "Dean Oliver" <deano@r...>
            wrote:
            > ... Cramer shows
            > very clearly that old guys would do worse in modern baseball. Ty
            > Cobb's lifetime batting average would be well below .367
            > (around .330). Rod Carew would be about the greatest average
            > hitter.

            This I would love to see. I already know it isn't going to convince
            me. Peripheral factors notwithstanding (i.e., rigors of travel,
            equipment, medicine,...), it just isn't possible to prove such a
            thing.

            > ... I did
            > quick work over in APBR looking at how FG%'s for long time players
            > trended through the '80's and '90's. That showed that most players
            > did have their FG%'s decline in the '90's, even before they "got
            > old." I'd be curious to look more at players from the '70's to
            > the '80's, where league FG%'s were going up -- were longtime guys
            > seeing their FG% also go up?

            (Somehow, I didn't see anything of this sort at the other site.)

            It's hard to imagine league-wide FG% going up without individuals'
            FG% going up (likewise, down).
          • Dean Oliver
            ... convince ... Pete and John make the analogy with the Olympics, saying that the athletes are simply much better now. They do admit that conditions were not
            Message 5 of 8 , Jun 24, 2003
            • 0 Attachment
              --- In APBR_analysis@yahoogroups.com, "Mike G" <msg_53@h...> wrote:
              > --- In APBR_analysis@yahoogroups.com, "Dean Oliver" <deano@r...>
              > wrote:
              > > ... Cramer shows
              > > very clearly that old guys would do worse in modern baseball. Ty
              > > Cobb's lifetime batting average would be well below .367
              > > (around .330). Rod Carew would be about the greatest average
              > > hitter.
              >
              > This I would love to see. I already know it isn't going to
              convince
              > me. Peripheral factors notwithstanding (i.e., rigors of travel,
              > equipment, medicine,...), it just isn't possible to prove such a
              > thing.

              Pete and John make the analogy with the Olympics, saying that the
              athletes are simply much better now. They do admit that conditions
              were not as good then, but don't imply that Dick accounted for that.

              I do tend to agree that it's not possible to "prove" such a thing. I
              also haven't seen a lot of value in doing such a comparison except in
              the Who is Best debate, which hasn't been my focus (though is fun).
              But the inference from a number of things can be pretty strong.

              As I say, I don't know the details without getting the article. But
              HGOB describes it as

              "Dick Cramer employed an ingenious method of evaluating average
              batting skill across time. He compared the batting performance of
              the same player in different seasons, avoiding the pitfall of
              environmental factors such as ball resilience, rule changes, racial
              mix, playing conditions, etc. by subtracting league averages before
              making any comparison. "Of course," Cramer wrote, "direct comparison
              cannot be made for seasons more than 20 years apart; few played much
              in both periods, say, 1950 and 1970. But these seasons can be
              compared indirectly by comparing 1950 to 1955 to 1960, etc., and
              adding the results."

              Cramer states the major trends as

              1. The average level of batting skill has improved steadily and
              substantially since 1876. The .120-point difference [between then
              and 1980] implies that a batter with 1979-average skills would in
              1876 have had the value of an otherwise 1876-average batter who hit
              enough extra singles for a.385 average.
              2. The AL and NL were closeley matched, blah not related to NBA.
              3. The recent and also the earliest expansions had only slight and
              short-lived effects on batting competitiveness. However, the blip
              around 1900 shows the substantial effect on competition that changing
              the number of teams from 12 to 8 to 16 can have!

              I highly recommend reading HGOB and/or the Cramer article for those
              interested in doing this kind of work for hoops.

              >
              > > ... I did
              > > quick work over in APBR looking at how FG%'s for long time
              players
              > > trended through the '80's and '90's. That showed that most
              players
              > > did have their FG%'s decline in the '90's, even before they "got
              > > old." I'd be curious to look more at players from the '70's to
              > > the '80's, where league FG%'s were going up -- were longtime guys
              > > seeing their FG% also go up?
              >
              > (Somehow, I didn't see anything of this sort at the other site.)

              APBR Message 11790. I looked only at 2pt FG% to avoid biases. It's
              reproduced below:

              --- In APBR@yahoogroups.com, Keith Ellis <jolk85@y...> wrote:
              > --- Bob Kuska <bobkuska@h...> wrote:
              > ---------------------------------
              > [.....] when the world's greatest players clank shot
              > attempt after shot attempt - even if the defensive
              > strategy of the opposing team is brilliant - viewers
              > feel like the series is a fraud. >>
              > >>
              >
              > In attempting to avoid isolated examples, we miss out
              > on some interesting...isolated examples. John
              > Stockton's FG% from '94 thru 2002 remained above 50%
              > altho the league's "defense" improved dramatically as
              > per league averages.
              >
              > Karl Malone, Hakeem, Ewing, & MJ finally suffered
              > lower FG%s due to age or rust, not better defensing.
              > Shaq, Grant Hill, Van Exel, Kenny Anderson, Clifford
              > Robinson, even Reggie Miller all maintained their good
              > or bad shooting percentages as the 90s wore on.
              >

              This is an interesting comment. I went back to find players who
              played at least 10 years across the '80's and '90's to see if they
              showed declines in 2pt FG% (to avoid biases of players shooting more
              3s). Specifically, I looked at the correlation between season and
              2pt FG% for all these guys. Only 17% showed non-negative correlation
              going into the 1990s. Only 2.5% showed positive correlation above
              0.3. Roughly 68% showed negative correlation below -0.3. This
              implies that long-term players do generally show a decline through
              time.

              Some of this could definitely be age. If I screen out the last 3
              years of a player's career, the effects are not as strong as above,
              becoming 38%, 18%, and 46%. This still says long-term guys were
              feeling some declining effects in 2pt shooting percentage through
              time.

              Keith does point out some valid counterexamples, though MJ and Ewing
              actually did show significant decline in FG% from their early days.
              Kevin Johnson is an exception not pointed out.

              Big guys tended to show the biggest decline through time, which is
              consistent with the Riley Effect I pointed out in the previous
              referenced article. Too much brutality in the middle. Gotta call it
              tighter, not allow hard fouls, call more intentional fouls. I think
              the league is going this way.


              > Why is it only the low FG%s of newstyle players like
              > Iverson & Kidd, maybe Spree are held up as evidence of
              > great D? We surely agree that fewer Centers today
              > defend as well as those of 10, 20, & 30 years ago.
              >

              Not sure about that last thing. I'd say that a lot of centers now
              are only in because of their defense. It sure ain't their offense.

              DeanO


              >
              > It's hard to imagine league-wide FG% going up without individuals'
              > FG% going up (likewise, down).

              Take a look at the book, Mike. It doesn't explain it all, but it
              discusses a lot of concerns people may have.

              DeanO
            • Michael Tamada
              [...] ... Interesting, it s similar to what I ve been proposing (look at long career players such as Hayes and Havlicek and how their stats changed over time)
              Message 6 of 8 , Jun 24, 2003
              • 0 Attachment
                [...]

                >As I say, I don't know the details without getting the article. But
                >HGOB describes it as
                >
                >"Dick Cramer employed an ingenious method of evaluating average
                >batting skill across time. He compared the batting performance of
                >the same player in different seasons, avoiding the pitfall of
                >environmental factors such as ball resilience, rule changes, racial
                >mix, playing conditions, etc. by subtracting league averages before
                >making any comparison. "Of course," Cramer wrote, "direct comparison
                >cannot be made for seasons more than 20 years apart; few played much
                >in both periods, say, 1950 and 1970. But these seasons can be
                >compared indirectly by comparing 1950 to 1955 to 1960, etc., and
                >adding the results."
                >
                >I highly recommend reading HGOB and/or the Cramer article for those
                >interested in doing this kind of work for hoops.

                Interesting, it's similar to what I've been proposing (look at long
                career players such as Hayes and Havlicek and how their stats
                changed over time) and sounds like it's also similar to the work that
                MikeG did with minutes played statistics and how they vary from
                season to season.

                One disadvantage that basketball analysts will face is that basketball
                careers are typically shorter than baseball careers, so the long
                chains of year-to-year data will be scarcer.

                I can't tell from the description what corrections if any Cramer made
                for a player's age/experience. But those corrections will probably
                be even more important for NBA players; so we need good models of
                predicting the trajectory of a player's stats (most players have
                low-productivitity rookie seasons, rise, and as Ed noted here yesterday,
                peak around their 5th year, at least as measured by his tendex-type
                stat). I don't know how FG% trajectories look, but they probably
                peak later, and probably have a less drastic fall-off at the end.

                I think this trajectory analysis will be important for correcting the
                year-to-year data. MikeG found consistent downward trends in players'
                minutes from year-to-year, and DeanO found consistent downward trends in
                players' FG%. Could be due to strength of competition, could be due to
                league strategies, but I'm guessing that some of it is due to players
                having a few years of increase followed by many more years of gradual
                decline in many of their stats.


                --MKT
              • Gary Collard
                ... If you don t find it, let me know, I m sure I can get it mailed to me/you. -- Gary Collard SABR-L Moderator collardg@earthlink.net Working at the CIA has
                Message 7 of 8 , Jun 24, 2003
                • 0 Attachment
                  Dean Oliver wrote:
                  >
                  > Cramer did his work more rigorously than I did in my 10 minute study,
                  > normalizing relative to league averages, etc. Does anyone have that
                  > work from the Baseball Research Journal? It may be useful here.

                  If you don't find it, let me know, I'm sure I can get it mailed to me/you.

                  --
                  Gary Collard
                  SABR-L Moderator
                  collardg@...

                  "Working at the CIA has taught me that any American boy or girl really
                  can grow up to be president. You just might not have control over
                  which Middle Eastern country you wind up being president of."
                  -- The Covert Comic
                • Mike G
                  ... much ... that ... What I found was that in every non-expansion NBA season, minutes are reduced for the average player (from the previous season). These
                  Message 8 of 8 , Jun 25, 2003
                  • 0 Attachment
                    --- In APBR_analysis@yahoogroups.com, "Michael Tamada" <tamada@o...>
                    wrote:
                    > > [DeanO:]
                    > > Cramer wrote, "direct comparison
                    > >cannot be made for seasons more than 20 years apart; few played
                    much
                    > >in both periods, say, 1950 and 1970. But these seasons can be
                    > >compared indirectly by comparing 1950 to 1955 to 1960, etc., and
                    > >adding the results."

                    >
                    > Interesting, it's similar to what I've been proposing (look at long
                    > career players such as Hayes and Havlicek and how their stats
                    > changed over time) and sounds like it's also similar to the work
                    that
                    > MikeG did with minutes played statistics and how they vary from
                    > season to season.

                    What I found was that in every non-expansion NBA season, minutes are
                    reduced for the average player (from the previous season). These
                    minutes are presumably taken up by rookies, who play more minutes
                    than guys in their final seasons.

                    Whether you look at consecutive seasons or the cumulative effects of
                    5 seasons, there is no conclusion available about whether declining
                    minutes are more the result of (1) aging, or (2) increasing
                    competition.

                    What I did find was that there was significant re-apportionment
                    (inflation) of minutes in expansion years. This to me implies a
                    certain reduction in competition for those years. Presumably, the
                    competition remains (partly) diluted until the curve returns to the
                    cumulative norm.

                    It's in the definition of this 'norm' that we all get our prejudices
                    out on the table. If we 'know' the competition is stiffer in the
                    present era, then the reducing-minutes phenomenon is primarily the
                    result of increasing competition.

                    My gut feeling is that there is so much 'noise' in the eral
                    differences of travel, training, etc, that we may as well resign
                    ourselves to an arbitrary definition, such as 1967 = 1977, draw a
                    straight line thru those, and extrapolate forwards and backwards.

                    So, if the cumulative annual ups-and-downs in average minutes yields
                    an annual average over that interval of, say, .9785, then that can
                    be the annual standard that accounts for 'aging'; and any departures
                    can be attributed to changes in competitive level.

                    Using such a short baseline (11 years) could yield startling
                    suggestions for 1953 and 2003, of course.


                    > MikeG found consistent downward trends in players'
                    > minutes from year-to-year, and DeanO found consistent downward
                    trends in
                    > players' FG%. Could be due to strength of competition, could be
                    due to
                    > league strategies, but I'm guessing that some of it is due to
                    players
                    > having a few years of increase followed by many more years of
                    gradual
                    > decline in many of their stats.

                    Actually, it shouldn't depend on the shape of the trajectory, but
                    only on the fact that a player whose rookie-year minutes were 2000
                    and whose 10th-and-final-year minutes were 1000 shows an average
                    loss of 100 minutes per year.

                    Whether he peaks at 2000 or 3000 minutes, the average for his career
                    comes out to the same thing. No difference whether the trajectory
                    is smooth or all ups-and-downs.

                    I spent a couple of hours tracking year-to-year changes in (1) FG%
                    and (2) scoring % (points divided by attempts).

                    I've got it tracked by NBA season (1952-2002) and by players' years-
                    in-the-league. There are a couple of complications, though.

                    A 15-year player has a much different trajectory over his first 5
                    years than does a 5-year player. The data is skewed very much to
                    the short-career player.

                    Rather than attempting the monumental task of categorizing 2nd-year
                    players from 1995, 3rd-year players from 1995, etc (some 15*60
                    subcategories); I just grouped them into 2-5th-year, 6-10 year, 11-
                    15 year. (Beyond 15 years is insubstantial sample.)

                    I will post this In Some Form, if anyone is interested. When I get
                    home.
                  Your message has been successfully submitted and would be delivered to recipients shortly.