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

Height impacts

Expand Messages
  • Dean Oliver
    First of all, here are average heights (weighted by minutes played) of teams through time: SeasonYear AvgOfMinWtdHt DiffPvsYr 1952 76.2 1953 76.4 0.1 1954 76.6
    Message 1 of 3 , Feb 1, 2003
    View Source
    • 0 Attachment

      First of all, here are average heights (weighted by minutes played) of teams through time:

       

      SeasonYear

      AvgOfMinWtdHt

      DiffPvsYr

      1952

      76.2

       

      1953

      76.4

      0.1

      1954

      76.6

      0.2

      1955

      76.7

      0.2

      1956

      77.1

      0.3

      1957

      77.2

      0.2

      1958

      77.3

      0.0

      1959

      77.2

      0.0

      1960

      77.4

      0.2

      1961

      77.5

      0.1

      1962

      77.5

      0.0

      1963

      77.6

      0.1

      1964

      77.7

      0.1

      1965

      77.8

      0.1

      1966

      77.6

      -0.2

      1967

      77.6

      0.0

      1968

      77.8

      0.2

      1969

      77.9

      0.1

      1970

      77.8

      -0.1

      1971

      77.9

      0.1

      1972

      77.9

      0.0

      1973

      77.9

      0.1

      1974

      77.9

      0.0

      1975

      78.0

      0.1

      1976

      78.2

      0.1

      1977

      78.2

      0.0

      1978

      78.3

      0.1

      1979

      78.3

      0.0

      1980

      78.4

      0.1

      1981

      78.6

      0.2

      1982

      78.7

      0.1

      1983

      78.7

      0.0

      1984

      78.8

      0.1

      1985

      78.9

      0.1

      1986

      79.1

      0.2

      1987

      79.3

      0.2

      1988

      79.3

      0.0

      1989

      79.0

      -0.2

      1990

      78.9

      -0.1

      1991

      79.0

      0.1

      1992

      78.8

      -0.2

      1993

      78.8

      0.0

      1994

      78.9

      0.0

      1995

      79.1

      0.2

      1996

      79.1

      0.0

      1997

      78.9

      -0.2

      1998

      78.9

      -0.1

      1999

      78.9

      0.0

      2000

      78.9

      0.0

      2001

      78.9

      -0.1

       

      (I don’t remember how to post tables so they format right, so I may try this a couple times.)

       

      A long trend of increasing heights (posted in part to compare to MSG’s stuff).  Decreases in average height seem to be the year of or the year after expansion.  Teams are definitely more than an inch taller now than in the ‘60’s.  There does appear to be a correlation between league average height and league average points scored per possession (about 80%).  I ran some experiments within seasons to look at the increment better than the league average points per possessions (offense and defense) as a function of increment above the league average height and I get very statistical significant results (>99.9% significant).  Offensively teams show an improvement of 0.78 pts per 100 possessions for each inch over the league average in height.  Defensively, it’s about 0.57.  However, this by no means explains the full increase in league average points per possessions that has occurred since 1974.  The ABA merger and the 3pt shot both had effects, though the ABA merger perhaps had more of an effect.

       

      I want to do some work to extract what this means on an individual basis – if an inch of height teamwise implies 1.35 pts/100 poss improvement, does it mean the same for individuals?  Maybe, if applied right.  I don’t think it means that a 7 footer is 13.5 pts/100 possessions better than a guy 6’2”.  But a 7’ power forward may be an average of about 4 pts/100 poss better than a 6’9” power forward.  Decisions by basketball people have not been made on pts/100 poss basis.  I’d like to understand whether I can make that correlation to see if they have over- or under-estimated the impact of height.  My gut says that they’ve overestimated that impact.

       

      I also need to think about other ways to identify these team numbers that may be better.

       

      (MikeG – did you notice that your minutes played ratio is over 1 in 1995 before expansion and less than 1 in 1996, the year of expansion?)

       

      Dean Oliver

      deano@...

      Journal of Basketball Studies

      http://www.rawbw.com/~deano/

      6052 Chabot Rd. #9

      Oakland, CA 94618

      (510) 655-4087

       

    • Mike G <msg_53@hotmail.com>
      ... try ... Hey, the table looks great, until you try to reply to it. ... 1995 ... Nah, 1996 is consistent. You might have noticed 1988 looking expansive
      Message 2 of 3 , Feb 3, 2003
      View Source
      • 0 Attachment
        --- In APBR_analysis@yahoogroups.com, "Dean Oliver" <deano@r...>
        wrote:
        > (I don't remember how to post tables so they format right, so I may
        try
        > this a couple times.)

        Hey, the table looks great, until you try to reply to it.


        > (MikeG - did you notice that your minutes played ratio is over 1 in
        1995
        > before expansion and less than 1 in 1996, the year of expansion?)

        Nah, 1996 is consistent. You might have noticed 1988 looking
        expansive while '89 should have but didn't. But '85 also shows
        inflated minutes, for no good reason.

        In any case, I don't entirely 'trust' this table any more than the
        previous one I did. So just for kicks, here are the two side-by-
        side. A 3rd column is merely the average of the other 2.

        I'm only listing the year of measurement (against the previous year);
        then the average ratio in minutes played, for guys playing both
        years; then the average in minutes-per-game; then the average of
        those 2.

        yr min. mpg. avg.
        53 .951 1.023 .987
        54 .905 .946 .925
        55 .891 .918 .905
        56 .922 .935 .929
        57 .998 .944 .971
        58 .995 .955 .975
        59 .975 1.010 .993
        60 .983 .921 .952
        61 .941 .964 .953
        62 1.107 1.050 1.079
        63 .901 .912 .907
        64 .969 .940 .955
        65 .925 .960 .943
        66 .968 .976 .972
        67 1.079 1.061 1.070
        68 1.188 1.201 1.195
        69 1.046 1.014 1.030
        70 .974 .971 .973
        71 1.052 1.050 1.051
        72 .917 .933 .925
        73 .938 .954 .946
        74 1.008 .971 .989
        75 1.019 .963 .991
        76 .951 .923 .937
        77 .942 .870 .906
        78 1.000 .983 .992
        79 .970 .976 .973
        80 .986 .950 .968
        81 1.003 1.003 1.003
        82 .945 .936 .940
        83 .992 .956 .974
        84 .962 .993 .978
        85 1.012 .943 .977
        86 .987 .960 .974
        87 .988 .962 .975
        88 1.010 .987 .998
        89 .990 1.014 1.002
        90 1.025 1.016 1.020
        91 .989 .957 .973
        92 .971 .979 .975
        93 .955 .928 .941
        94 .970 .947 .959
        95 .998 .947 .973
        96 1.012 1.010 1.011
        97 .992 .988 .990
        98 .970 .963 .967
        99 .948 .960 .954
        00 .973 .947 .960
        01 1.015 no data

        Now '88 has sunk beneath the waves, as has '85. Every expansion year
        shows minutes-inflation, while every non-expansion year <1.
      • Mike G <msg_53@hotmail.com>
        I doubt I can explain how I came up with this, tho it s a cumulative index of inflation/contraction of competition for minutes, year by year. Explanation could
        Message 3 of 3 , Feb 3, 2003
        View Source
        • 0 Attachment
          I doubt I can explain how I came up with this, tho it's a cumulative
          index of inflation/contraction of competition for minutes, year by
          year.

          Explanation could be forthcoming, with prompting. But in the name of
          timeliness, here's what I have. A low number is a tight year, where
          minutes are relatively hard to come by. A larger number is a year of
          accumulated expansion, or of dilution of competition.

          yr2 cumX
          53 1.080
          54 1.049
          55 1.007
          56 .980
          57 .975
          58 .971
          59 .976
          60 .961
          61 .947
          62 .991
          63 .953
          64 .940
          65 .921
          66 .917
          67 .956
          68 1.052
          69 1.077
          70 1.072
          71 1.108
          72 1.076
          73 1.056
          74 1.060
          75 1.065
          76 1.041
          77 1.000
          78 1.005
          79 1.000
          80 .993
          81 1.004
          82 .982
          83 .978
          84 .976
          85 .974
          86 .970
          87 .966
          88 .974
          89 .984
          90 1.002
          91 .997
          92 .994
          93 .973
          94 .962
          95 .957
          96 .971
          97 .975
          98 .967
          99 .953
          00 .943
          01 .941


          The software is behaving weirdly (editing produces deletions), so I
          have to be careful. These numbers might be a factor that could be
          divided into other T#nd$x-like number for the designated year.

          For those of us who are into such things.
        Your message has been successfully submitted and would be delivered to recipients shortly.