The latest trade demand from Kobe might cause some to look back at the days when the Lakers were led by players who demanded their team win by producing large quantities of wins (as opposed to Kobe’s method of just demanding the team acquire players who can produce large quantities of wins).
In 1979 the LA Lakers took Magic Johnson with the number one pick in the NBA draft.
Pairing Magic with Kareem Abdul-Jabbar led to the birth of Showtime, and when the 1980 playoffs ended, an NBA championship. Which player, though, mattered more? Was Magic the star of this team? Or was Kareem still leading the way in Magic’s rookie season?
As with my analysis of the 1977 Portland Trail Blazers, the answer (at least the one I am providing) is in Wins Produced. The following table reports the Wins Produced for the 1979-80 LA Lakers.
Table One: The LA Lakers in 1979-80
The Lakers won 60 games that season, an increase of 13 from the previous campaign. The summation of Wins Produced comes to 57, and of these, 23.4 are traced to Kareem. Another 21.1 can be attributed to Magic. When we look at Wins Produced per 48 minutes [WP48] it looks like these two players – who each posted a 0.36 mark — were equally responsible for team success. After Magic and Kareem, the only above average player who played significant minutes was Jamaal Wilkes. The combination of Magic, Kareem, and Silk produced 54.3 of the 57 Wins Produced this team accumulated.
It’s important to remember that Kareem was 32 years old that season while Magic was just a 20–year old kid. As the decade continued, Magic became more and more important while Kareem’s contribution declined. This can be seen when we look at the 1986-87 edition of Showtime.
Table Two: The LA Lakers in 1986-87
The Lakers won 65 games this year, and nearly 29 of these victories can be linked to Magic. Although Magic was truly “magical”, he was not a one-man team. James Worthy, A.C. Green, Byron Scott, Michael Cooper, and Kurt Rambis were all above average players. These five players produced 33.9 victories.
Kareem at this point was 39 years old, but still taking the floor for 31 minutes a night. But he clearly was not a good as he once was. His WP48 was only 0.094 and he only produced 4.8 wins.
Again, the NBA only began tracking the data necessary to calculate Wins Produced, Win Score, PAWS (Position Adjusted Win Score), etc… in 1977-78. If we look at Kareem’s PAWSmin (PAWS per minute), and his corresponding WP48 and Wins Produced (which you can estimate once you know PAWSmin), we see that Kareem was truly an amazing player before he got old.
Table Three: Kareem from 1977-78 to 1988-89
From 1977-78 to 1980-81, Kareem averaged 25.1 wins per season. In 1980-81 he was 33. Over the next four seasons – 1981-82 to 1984-85 — Kareem was in his mid-30s but still very good, producing 57.1 wins. To put that in perspective, Kobe Bryant – the man demanding a trade every other week – has only produced 51.6 wins across the past four seasons.
Still, although Kareem was quite good, he was not the main star of Showtime.
Table Four: Magic Johnson’s Career Productivity
In the same four seasons Kareem produced 57.1 wins, Magic produced 113.3 wins. And in the next four seasons – 1985-86 to 1988-89 – Magic produced another 106 wins. And the two seasons after this, Magic added another 56.8 wins. Yes, Magic produced more in two seasons at the age of 30 and 31 than Kobe has produced the past four campaigns.
It’s important to note that the estimates of wins for Magic and Kareem across their careers are based on the link between PAWSmin and WP48. Because NBA teams played at a faster tempo in the 1980s, these estimates tend to over-state their contribution. But it looks to me that the over-estimation is not very large. Magic’s Wins Produced in 1986-87 (reported in Table Two above), a measure that took into account team pace, indicated that Johnson produced 28.8 wins that season. The estimate of Wins Produced based on PAWSmin says he produced 28.5 wins that season. When we look at Kareem, we see a bigger difference. His WP48 is reported to be 0.094 in Table Two. When we look at PAWSmin, though, we get a WP48 of 0.143.
When I have the time, I prefer to calculate Wins Produced the old-fashioned way (as detailed in Chapters Six and Seven of the book). Still, the picture we paint with PAWSmin is often “close enough”. And in this case, both pictures indicate that Magic initially shared the stage with Kareem in the LA. But by the mid-1980s, it looks like the Lakers were Magic’s show.
History is something I love and I plan on returning to stories from the past again and again in this forum. My next two historical posts will examine the Bad Boys of Detroit and Larry Bird’s Boston Celtics. I also need to start writing about the NBA Draft. Hopefully I can get to all of these topics before the week (or month) is over.
– DJ
Wages of Wins Journal 
Owen
June 19, 2007
Great stuff. These historical posts are great. CAn’t wait for the New York Knicks/Patrick Ewing edition.
Jason
June 19, 2007
I’d suspect that there may be a larger error in the conversion between PAWS and WP48 that even you suspect and may not be “close enough” to evaluate a player’s decline or to evaluate the relative contributions of players across decades.
I did a quick look at the 77-78 season and the 86-87 season and while scoring was slightly up in the 86-87 season (109.9ppg per squad vs 108.5 in 77-78) rebounds were down (average 3863 per club over the season in 77-78 vs 3610 in 86-87). The average WS/min was actually higher in 77-78 than it was in 86-87, as a result.
I think this has something to do with how the corrections are factored. The pace wasn’t *that* much faster in the 80s than in was in the 70s but the number of available rebounds was rather different.
The “scoring” part of WS has, to some degree, a built in correction by discounting points with shots taken. Yes a higher FG% is better (and FG% was higher in the 80s by about 1% a game). This is accounted for in the scoring side of things already, but not in what it means to rebounds. More misses (and there were more misses) means more opportunity to grab rebounds.
Rebounding has no built-in correction inherent except for either the position correction (which need not stay constant nor do the relationships between positions need to stay relatively constant) and the ‘team correction’ for pace and the regression value vs. wins.
If anything, I think that the 70s may show inflated relationships based on converting WS simply because there were more total rebounds. Since the number of wins in a game don’t change, with an increased number of rebounds, the correction for the average should ‘discount’ them to a degree that the conversion didn’t pick up.
Further, some anecdotal evidence of a shift in the position correction: in 77-78, the majority of the top 10 rebounders (8 of 10 by my count) palyed the majority of their minutes as centers while in 87-88 5 of the top 10 rebounders (including the top 3 overall) were forwards. There appeared to be a shift in the rebounding burden which would make the difference in correction between forwards and centers diminished.
This may not have a huge effect on estimating team wins, but could have a big impact on comparing contributions from one season to the next.
dberri
June 19, 2007
Jason,
When I first looked at the 77-78 season I thought it wouldn’t work at all for exactly the reasons you stated. There were more rebounds in the 1970s. Yet when I looked at the Blazers and 76ers I got “credible” results. Frankly I was pretty surpised that using results from the past 15 years would work 30 years ago. My hope is to go back and do all the teams from the 1970s and use the averages from those years. I suspect my answers will stay roughly the same. But maybe I’m wrong.
As for PAWSmin, when we look at Bill Walton we get an estimate of 19.8 wins when we calculate Wins Produced the traditional (long) way. When we take the PAWSmin approach we get 18.3 wins. Again, I am surprised that it would be less with the shortcut. It should be more.
Nevertheless, it does look like I can go back to teams from the late 1970s and 1980s and via this approach arrive at reasonable measures of player performance.
Chirstopher
June 20, 2007
Nice post. I’m been convinced for some time that Magic was the best player ever in the NBA (not Jordan, he’s 2nd) and this quantifies what I’ve thought for some time. A question for you: Have you ever thought of including some strength of opposition factor in your model? When Player X has a triple double against the Bobcats one night and then a few weeks later does the same thing against the Mavs you’d think that later would be worth more than the former.
Arizona Home Owner Insurance
June 26, 2007
Kareem is still under rated, not that anything you’ve said here is incorrect. He just does not get the kudos he deserves, partly because he was relatively quiet from what I remember.