ABA equivalents to NBA Win Shares, PER, etc.
Posted: Wed Aug 10, 2011 5:24 pm
Lots of fans are interested in historic comparisons, and it's never satisfying to see a list of players' "best seasons of the 3-point era", or "best playoffs ever, NBA only". The ABA had many of the best players of the 1970's, and ideally we should be able to compare them to their NBA contemporaries.
http://www.basketball-reference.com/lea ... stats.html
I copied all NBA and ABA players' Advanced Stats from 1965 through 1977 into Excel, sorted alphabetically and by year, and deleted all but those who played in both the NBA and ABA. From these, I retained only consecutive (or closest) player-seasons when one year was in one league and the next year was in the other.
Using the ABA season as reference -- since it was the more rapidly evolving league -- I found average stats for whole groups of [from 9 to 25] players who jumped leagues every year (1968 to 1976). Then from 1976 to '77, about 70 players changed leagues, in the merger year.
From the 1967-68 ABA season -- here called '68 -- some 22 players had at least 100 NBA minutes previously. For 12 of these, one or two years had gone by in the meantime.
Here are their agglomerate stats, from the ABA and from their most recent NBA stints:There are no NBA rebound percents for this era. They could be estimated, though.
Looking at the NBA shooting% and Offensive Win Share rates (OW/48), these had mostly been offensively-challenged players in the NBA, now charged with carrying the offensive load in the ABA. PER and WS/48 well above ABA averages.
The Diff: lines are just A-minus-N values. Most values are positive, in that ABA points and rebounds are easier to come by. Assists not so.
An immediate improvement is shown for the '69 season, only maintained in 1970.
The star quality of the '69 "class" is due to Rick Barry joining the ABA and Connie Hawkins leaving. Barry is compared to his '67 NBA self and Hawkins to 1970.
An assumption here is that the NBA does not change much in the 2-3 years of comparison, and/or that roughly equal player-minutes are being compared to past and future NBA seasons.
This might be called the Early ABA, with huge differences from NBA in what players get in WS/48 and PER.
Next, the Middle ABA era, where the difference is about one-half:
Each year's group of league-jumpers has it's unique character: offense/defense, strong or weak rebounders.
Note that by '74, an ABA rebound is just as hard to get as one in the NBA.
After strong gains just 4 years in, the ABA slipped a bit in '73 and 74. This is probably due to some serious defections, along with less domination of the draft.
In the last 2 ABA years, there was major contraction, concurrent with more domination of top draft picks.According to this study, the 1976 ABA was entirely equal to the NBA, in overall quality. Still a bit easier scoring in the ABA, but harder to get rebounds.
And that's mostly the 1977 NBA; I'll bet the 1976 NBA doesn't look this good.
http://www.basketball-reference.com/lea ... stats.html
I copied all NBA and ABA players' Advanced Stats from 1965 through 1977 into Excel, sorted alphabetically and by year, and deleted all but those who played in both the NBA and ABA. From these, I retained only consecutive (or closest) player-seasons when one year was in one league and the next year was in the other.
Using the ABA season as reference -- since it was the more rapidly evolving league -- I found average stats for whole groups of [from 9 to 25] players who jumped leagues every year (1968 to 1976). Then from 1976 to '77, about 70 players changed leagues, in the merger year.
From the 1967-68 ABA season -- here called '68 -- some 22 players had at least 100 NBA minutes previously. For 12 of these, one or two years had gone by in the meantime.
Here are their agglomerate stats, from the ABA and from their most recent NBA stints:
Code: Select all
68 Min. mpg PER TS% eFG% TRB% AST% OW/48 DW/48 WS/48
A 34848 28.6 17.3 .498 .434 9.6 10.3 .078 .054 .132
N 11195 14.4 10.9 .444 .392 9.4 -.008 .041 .037
Diff: 14.2 6.4 .055 .042 .8 .086 .013 .095
69
A 7260 21.7 21.2 .540 .478 11.4 12.3 .122 .059 .182
N 9889 19.6 17.7 .507 .443 13.7 .076 .044 .122
Diff: 2.2 3.5 .032 .035 -1.4 .045 .015 .060
70
A 23224 25.4 15.0 .498 .455 9.3 13.5 .040 .052 .098
N 9141 11.5 11.6 .448 .405 13.4 -.002 .037 .039
Diff: 13.9 3.4 .050 .051 .2 .042 .015 .059
Looking at the NBA shooting% and Offensive Win Share rates (OW/48), these had mostly been offensively-challenged players in the NBA, now charged with carrying the offensive load in the ABA. PER and WS/48 well above ABA averages.
The Diff: lines are just A-minus-N values. Most values are positive, in that ABA points and rebounds are easier to come by. Assists not so.
An immediate improvement is shown for the '69 season, only maintained in 1970.
The star quality of the '69 "class" is due to Rick Barry joining the ABA and Connie Hawkins leaving. Barry is compared to his '67 NBA self and Hawkins to 1970.
An assumption here is that the NBA does not change much in the 2-3 years of comparison, and/or that roughly equal player-minutes are being compared to past and future NBA seasons.
This might be called the Early ABA, with huge differences from NBA in what players get in WS/48 and PER.
Next, the Middle ABA era, where the difference is about one-half:
Code: Select all
71 Min. mpg PER TS% eFG% TRb% Ast% OW/48 DW/48 WS/48
A 12993 23.8 17.6 .523 .481 13.9 9.8 .059 .057 .116
N 10829 19.9 14.8 .499 .451 9.2 .056 .037 .090
Diff: 3.9 2.8 .024 .029 .6 .004 .020 .026
72
A 29020 23.0 15.7 .510 .463 10.4 10.2 .046 .035 .084
N 23861 17.5 13.8 .471 .425 9.5 13.0 .017 .040 .060
Diff: 5.5 1.9 .038 .038 .9 -2.8 .029 -.005 .024
73
A 11748 18.7 15.9 .512 .468 12.7 14.1 .038 .055 .092
N 13668 17.3 14.2 .475 .429 11.1 13.6 .020 .033 .055
Diff: 1.4 1.8 .037 .040 1.6 .5 .017 .022 .037
74
A 16687 27.0 16.8 .545 .500 9.6 14.0 .081 .038 .120
N 15855 18.4 13.7 .484 .442 9.7 14.3 .027 .050 .076
Diff: 8.6 3.2 .061 .059 -.1 -.3 .054 -.011 .043
Note that by '74, an ABA rebound is just as hard to get as one in the NBA.
After strong gains just 4 years in, the ABA slipped a bit in '73 and 74. This is probably due to some serious defections, along with less domination of the draft.
In the last 2 ABA years, there was major contraction, concurrent with more domination of top draft picks.
Code: Select all
75 Min. mpg PER TS% eFG% TRB% Ast% OW/48 DW/48 WS/48
A 18536 23.4 14.1 .491 .444 8.8 15.8 .032 .038 .073
N 17072 17.5 13.3 .470 .425 10.2 13.9 .013 .045 .062
Diff: 5.9 .8 .021 .020 -1.4 1.9 .019 -.008 .010
76
A 131288 27.1 15.4 .519 .474 10.1 13.4 .0529 .0518 .1047
N 111018 23.8 15.4 .509 .463 10.6 13.3 .0527 .0517 .1042
Diff: 3.3 .0 .009 .011 -.5 .0 .000 .000 .001
And that's mostly the 1977 NBA; I'll bet the 1976 NBA doesn't look this good.